Media Player .NET (MPDN) - D3D HQ GPU Video Renderer [v2.49.0/v1.31.0 27 Dec 2018]

[ryrynz]

Lav 0.68 has been released. I've made an inquiry with Alex about possibly releasing a new version of Sanear since it's been a while.

Belphemur, could you fix the updater so it always checks for the latest version before displaying the update window rather than relying on the cache? I believe you're already aware of this happening.

For example, I was prompted to update to 2.46 and 1.26.1 so I shutdown the player, afterwards
it refreshed the extensions to 1.26.2, ideally the notification should be up to date.

[Anima123]

I have found that video source with good quality and resolution, for example good encoded 720p source, tend to be overly-sharped, if using image-doubler, e.g., Super-xBR with default value together with SSIM. In this case, adjust the 'sharpness' down to a lower value, like 0.40 will give visually pleasant quality.

As contrary, if the source is low-resolution, like 720 x 404, Super-xBR with low 'sharpness' most likely will blur the image, in this case 'sharpness' as high as 80-100 will result in visually pleasant images.

I am not sure anyone has experienced the same. The reason behind it might be that lower resolution will need more "sharpness" to get closer to the 'grand truth' images, while images with higher-resolution and high quality just need to keep the sharpness before it has been overly sharpened, and hence brings other kind of artefacts.

What's your opinion on this?

[Zachs]

Kinda makes sense I suppose. You should try using the conditional script to make sure you have the sharpness you like for each resolution or a range of resolutions. Perhaps share it here too so others could use it.

[aufkrawall]

Is it by design that sometimes there are delayed frames reported with Fluid Motion?

[Zachs]

No. It only happens when the driver or GPU missed the presentation window. Happens more often if your display rate is higher because the window is smaller.

[YxP]

Any chance of an update for 3DLUT script behaviour? Preferably something like MVR's, where you can have more than one screen available and the player takes care of disabling OS's calibration. Well, you probably know what I mean. I have two calibrated screens and using 3DLUT right now is way too cumbersome.

[aufkrawall]

No. It only happens when the driver or GPU missed the presentation window. Happens more often if your display rate is higher because the window is smaller.
Do you think this is visible? :)

[huhn]

Do you think this is visible? :)

if the "wrong" frame is repeated is sure is.

[Zachs]

No. Not when your refresh rate is sufficiently high.

[YxP]

The player crashes every time I try to open a file if I have selected a script group to be used (using the premades by aufkrawall). I can, however, select one during playback and it works. Using latest LAV nightly -10. Win 8.1 x64.

[Blackfyre]

Hi Zacks,

Hope all is going well, I haven't been active for a long time here.

I'm trying to watch IP Man 3, and I have xySubFilter chosen and it's not loading the subtitles into the player. It's working fine with MPC but with MPDN even the tray-icon for xySubFilter is not showing. Maybe I'm missing something simple. Any help or guide regarding setting up xySubFilter with MPDN or just choosing it via options under subtitles should be enough?

Running the latest version + latest extensions.

I might send you a private message with a new guide/article for SVP 4 Full Version to add to the OP in a week or two. As the settings and their effects have changed significantly since SVP 3.

Regards, Blackfyre.

[Shiandow]

Hi all,

I've manage to improve the reconstructive bilateral scaler enough that I consider it (probably) ready for general use. I would really appreciate if you could try it out and see if there are indeed no more unacceptable artefacts.

Personally I'm pretty happy with the way it looks, in some cases it really manage to recover quite a bit of the chroma information, which is a lot more noticeable than you'd think (even at native resolution, but especially when upscaling).

[Blackfyre]

Hi all,

I've manage to improve the reconstructive bilateral scaler enough that I consider it (probably) ready for general use. I would really appreciate if you could try it out and see if there are indeed no more unacceptable artefacts.

Personally I'm pretty happy with the way it looks, in some cases it really manage to recover quite a bit of the chroma information, which is a lot more noticeable than you'd think (even at native resolution, but especially when upscaling).

We'll do some testing later today or tomorrow, thanks for the update Shiandow.

The player crashes every time I try to open a file if I have selected a script group to be used (using the premades by aufkrawall). I can, however, select one during playback and it works. Using latest LAV nightly -10. Win 8.1 x64.

I can confirm this, the script group examples linked in the OP are crashing MPDN when used.

Using Windows 10 latest built here.

MPDN x86

[ryrynz]

Being guided by luma is appaears to making the result appear to be more accurate, but it's super sharp at 0.5.. even at 0.1 it's sharper than madVR's Bilateral.

Using madshi's chroma test video from the movie Samsara (Two guys in red robes blowing horns) it really doesn't look that good even at 0.1 (at least with that example)
Although that is one of the worst examples you can use to show bilateral's faults.

Doesn't seem to be stable either, MPDN stops working when using it, no dump. I have to switch to it from another scaler, I can't load a file with it activated.

Still prefer super-xbr for chroma, you should add it into MPDN, perhaps it too could be improved by guiding by luma? If that's possible..
Would be interesting to see how Hyllian's recent changes affecting ringing affect it vs madVR's older version.

[Shiandow]

Being guided by luma is appaears to making the result appear to be more accurate, but it's super sharp at 0.5.. even at 0.1 it's sharper than madVR's Bilateral.

Using madshi's chroma test video from the movie Samsara (Two guys in red robes blowing horns) it really doesn't look that good even at 0.1 (at least with that example)
Although that is one of the worst examples you can use to show bilateral's faults.

Doesn't seem to be stable either, MPDN stops working when using it, no dump. I have to switch to it from another scaler, I can't load a file with it activated.

Still prefer super-xbr for chroma, you should add it into MPDN, perhaps it too could be improved by guiding by luma? If that's possible..
Would be interesting to see how Hyllian's recent changes affecting ringing affect it vs madVR's older version.

I think I know a way to make it softer, which would probably be better for more noisy sources as well. The 'strength' setting ideally shouldn't affect sharpness, it only affects how far the algorithm is willing to go to reconstruct the chroma. Not sure if the 'reasonable' values are spaced well though, could be that you'll get reasonable results at 0.01, or even 0.001 (but not 0). Do you have a link to the chroma test video? That would probably help.

And I think it's doable to make Super-Xbr scale chroma, but I don't see an obvious way to make it use the high resolution luma to guide the interpolation. Maybe Hyllian can find a way, but maybe not.

I'm also not sure what's happening with the stability issues, I'm not seeing any at the moment. Perhaps I inadvertently fixed it, but that seems unlikely.

[Anima123]

For natural videos, the bilateral chroma upscaling dose not seem right if 'strength' is set to 0.10. Since 0.10 is the step, it does not look good under 0.20 as far as I can figure.

[Shiandow]

In what sense doesn't it look 'right'?

[Anima123]

Some kind of 'vague' at the edge of human faces. I am not very sensitive to chroma than luma, that I cannot describe it more precisely.

[Shiandow]

Some kind of 'vague' at the edge of human faces. I am not very sensitive to chroma than luma, that I cannot describe it more precisely.

Haven't noticed that before, I'll see if I can find out what's happening. Some screenshots would help though.

[Anima123]

Here's 2 screenshots:

strength = 0.10
https://www.dropbox.com/s/qwfnijzour2jegv/Img01_bilateral_0.10.png?dl=0

strength = 0.20
https://www.dropbox.com/s/hztmghcidmmu12t/Img01_bilateral_0.20.png?dl=0

It's not as obvious as I thought, but 0.10 seems to have a little more blurry.

[Shiandow]

Hmm, seems the 'strength' parameter isn't too useful then. Unless anyone has found a situation where a value other than 0.5 looks better? If not then I'll simply remove that parameter, no use keeping an option that doesn't ever improve the image.

So if anyone finds a case where strength = 0.5 doesn't look good, or some other value of strength looks better, I'd love to hear about it.

[Anima123]

I would prefer strength = 0.2 ~ 0.3, if I have to choose.

[ryrynz]

Here's 2 screenshots:

It's not as obvious as I thought, but 0.10 seems to have a little more blurry.

Just like in madVR, it's the reds that really show the difference, that's not a good shot to use for comparison IMO..

Hmm, seems the 'strength' parameter isn't too useful then. Unless anyone has found a situation where a value other than 0.5 looks better? If not then I'll simply remove that parameter, no use keeping an option that doesn't ever improve the image.

So if anyone finds a case where strength = 0.5 doesn't look good, or some other value of strength looks better, I'd love to hear about it.

0.5 was incredibly sharp.. Too high IMO. Test on these videos 1 (https://mega.nz/#!DRZzyAZY!Oq-RxDtp9K0hRbmsa8CQZXqFwy_sZDoe5JdNLjaRM4M), 2 (https://mega.nz/#!jcAlzC5K!oFGoQyhiiHfFq1PPi3rfYbw7r7UlKIXOdhp-9igKJ18) taken from the madVR thread. I use the first frame from both. Focus on the robes of the first video and the Japanese lettering in the speech bubble on the second.

I personally find super-xbr 100 in MadVR to look the most accurate or natural to my expectations.

Comparison using the second video super-xbr 100 vs bilateral @ 0.1
http://screenshotcomparison.com/comparison/166076

Comparison using the first video super-xbr 100 vs bilateral @ 0.1
http://screenshotcomparison.com/comparison/166078

[Shiandow]

In both cases the chroma limiter seems to be having some issues, do they look better if you turn "Improve Chroma Reconstruction (Ctrl+Shift+C)" off ?

[Shiandow]

Okay, after some more testing I've concluded that the reconstructive method still has some more potential, however to push it to it's limit I need a better way to filter out artefacts. For now the current version is my best attempt at balancing accuracy vs stability.

[Anima123]

Keep up the good work Shiandow.

And I did found that "Improve Chroma Reconstruction" tend to destroy the whole image in some cases, so I usually set it as off.

[Shiandow]

Yeah, I might need to revise the 'improve chroma reconstruction' at some point. It doesn't handle some edge cases well. Fixing it is a bit tricky though (GPUs don't like situations with lots of cases that need to be handled differently).

[ryrynz]

In both cases the chroma limiter seems to be having some issues, do they look better if you turn "Improve Chroma Reconstruction (Ctrl+Shift+C)" off ?

Yes. When disabling ICR, the speech bubble looks better than even super-xbr ( got similar results using madVR)
It was also the option responsible for darkening various elements on the robes. My guess is that this could perhaps be toned down a little perhaps.
Of course the robes edges are still all blown out even with ICR disabled.

It will be interesting to see how madVR's Super-xbr handles chroma with two levels of anti-ringing. ATM I'm no fan of enabling it when using it on chroma, but that might change with the new release.
A luma guided ICR enabled double anti-ringing super-xbr might just end up being top tier chroma upscaling :D

[ryrynz]

Not sure if due to recent changes but using Fluid Motion is like watching 30+ FPS content from just 24 FPS.

There's an almost Frame Creation type of effect from it producing a pop effect almost 3D like in appearance.
Zach has talked about MPDN's benefits having been built from the ground up to attain the very best frame presentation but this is the first time I can say I've ever seen it in practice.

[Shiandow]

Hi,

I've made some improvements to the reconstructive bilateral method again, I've managed to detect when there's not enough luma information to reconstruct the chroma, and made it avoid those situations. It's not perfect, but it never will be, and I think any further (significant) improvements will require a different approach.

So, I'd be interested to know what you all think. Is it strong enough when it can be? Or is it still too strong where it shouldn't be? Does it work well with SuperChromaRes, or is SuperChromaRes not even necessary anymore?

Also, if you want a quick comparison with other chroma scalers, here (http://i.imgur.com/qFUivrc.png) is the result for a popular benchmark image, and here (http://screenshotcomparison.com/comparison/166640) is the same image when using MadVR's soft-reconstructive and superXbr chroma scaling algorithms. For this image at least the reconstructive bilateral method is clearly superior.

Cheers,
Shiandow

[Shiandow]

Here are some more comparisons of the reconstructive bilateral method vs simple bicubic scaling:

Netflix logo (http://screenshotcomparison.com/comparison/167142)
Samsara sample (http://screenshotcomparison.com/comparison/167140)
Anime sample (http://screenshotcomparison.com/comparison/167141)

[madshi]

That does look pretty promising!

[Anima123]

Here are some more comparisons of the reconstructive bilateral method vs simple bicubic scaling:

Netflix logo (http://screenshotcomparison.com/comparison/167142)
Samsara sample (http://screenshotcomparison.com/comparison/167140)
Anime sample (http://screenshotcomparison.com/comparison/167141)

Isn't Bilateral a little bit too strong for Samsara sample? I also noticed more alias on the red robe of the lama.

Does smaller strength work smoother here?

[Shiandow]

Lowering the strength to 0.2 makes it smoother, it pretty much reverts to bicubic on anything except edges, but whether this is 'better' is up for debate. Also not sure how well such a low strength performs on other images.

There's not much that can be done to fix the aliasing though, the aliasing is pretty clearly present in the luma channel, so the method assumes it must therefore be present in the chroma channel as well (which, for all I know, might have been the case).

[madshi]

Yes, Samsara has clearly visible aliasing in the luma channel. Which means any luma guided chroma upsampling algorithm will run into some trouble with that sample. But FWIW, the new Bilateral seems to have less trouble with that than most other luma guided algorithms I've seen so far, which is one reason why I have hopes for the new Bilateral algo to become the new "best" chroma upscaler. But we'll see, we'll need some wider testing from users with lots of different content.

[aufkrawall]

This sample does not just only show nasty banding, also the red arrows are quite critical for chroma upscaling:
http://www36.zippyshare.com/v/LZSWS3My/file.html
Shiandow, you might want to take a look at it again with new bileteral scaler? :)

[Shiandow]

It seems to perform quite well on that image (comparison (http://screenshotcomparison.com/comparison/167276), scaled to 1080p), did the previous version have some trouble with that image? Only difficult part is the gradient near at the tip of the arrow, it's hard to be sure what that was supposed to look like, but the result looks pretty decent.

The banding is horrible though, and possibly unrecoverable (that image not only has horrible banding all over it, but also one weak intentional band running over Turkey, it's nearly impossible to remove one but not the other).

If you really want to give the bilateral algorithm a run for its money you should try images with faint fine details in the luma channel, that can really trip it up quite badly (for obvious reasons).

[madshi]

The chroma position seems to be shifted to the right in that sample, for some reason.

[Shiandow]

I think it's supposed to look that way, the curve matches up nicely, and if the chroma offset were incorrect then bilateral should have fixed that to some extent (so you'd see the chroma jump sideways compared to bicubic).

[madshi]

The reason I think the chroma position is shifted is that the red chroma of the arrow-head bleeds into the green background on the right side of the arrow-head. If the chroma were positioned correctly, this bleed should not occur, or at least it should occur everywhere, e.g. on the top & bottom of the curve, not just on the right side of the arrow head.

[huhn]

the rest of the images doesn't show the same shift.

i think is was moving and maybe even intentional

or there was so little bit rate that the chroma couldn't be changed.

this would not be the first time i would see a part of chroma from an old frame in broadcast.

[madshi]

You're right, it's probably a compression problem. Although it's the same throughout the whole time the arrow moves. However, when the arrow is in its final position, chroma suddenly becomes nicely centered everywhere. So at least the chroma position is not generally shifted in the whole video. Well, anyway...

[Shiandow]

Ah, that might explain why you don't see a chroma shift in the curve, that part doesn't move. It also explains why the algorithm has some problems with the arrow head (since luma and chroma don't match up).

[huhn]

to be honest your algorithm shouldn't try to fix this.

i mean how should sub sampling and compress create such a huge gap there?

on a good source it shouldn't touch more than the neighbor pixel or something like that.

[Shiandow]

on a good source it shouldn't touch more than the neighbor pixel or something like that.

Agreed, but I think we can agree that that particular image wasn't a good source. Also when you scale subsampled 720p source to 1080p, the chroma pixels become huge.

[huhn]

you want to use bilateral for direct scaling not just for 4:2:0 to 4:4:4?

[Shiandow]

Haven't tried that yet, but I suppose it would work. Might even be better in some cases, although other than performance I don't think you'll notice much difference.

But even if you go 720p 4:2:0 -> 720p 4:4:4 -> 1080p, you make any problem the chroma signal has a lot bigger.

[Hrxn]

Hi guys,

I am new here, so please bear with me a bit :)

Okay, so a friend of mine recommended me MPDN as a great new media player, and now I'm curious and want to try it out.

I just read the project page (http://www.zachsaw.com/mpdn/), requirements are in line, but this is a bit strange to me:

Certificate
zachsaw.com Software Certificate - Install this first. You only need to do this once.

What is this about? I don't know, I think I've never seen this for an application before. Isn't this a security risk? Installing a certificate to the trusted root store??

Doesn't MPDN work without the cert?

[ryrynz]

What is this about? I don't know, I think I've never seen this for an application before. Isn't this a security risk? Installing a certificate to the trusted root store??

Doesn't MPDN work without the cert?

It works fine without the cert. Read below.


This prevents the annoying prompt that you're running an application from an unknown publisher for software from zachsaw.com.
It also gives you the added security of preventing you from getting desensitised by these prompts (many already are by simply pressing 'Yes'
whenever they're prompted). It also prevents other developers from claiming their software is from zachsaw.com - another security feature.

[Hrxn]

But why using a cert then? What is the benefit?

I'm just curious.

[huhn]

it stops the prompt as an example.

[Zachs]

But why using a cert then? What is the benefit?

I'm just curious.
In an ideal world, I'd be able to afford a proper code signing cert, but alas MPDN is just a pet project. It costs too much for me to get a cert, so the only alternative is to have a self signed one. If someone would like to sponsor the purchase of a code signing cert, feel free to contact me. Or if you have a better idea, feel free to share it.

[Hrxn]

it stops the prompt as an example.

That one UAC prompt when running the installer? That is all? Really?

[Zachs]

That one UAC prompt when running the installer? That is all? Really?
Did you read ryrynz's reply?

[Hrxn]

Yes, of course.

I also read http://www.zachsaw.com/mpdn/

From his quote:
This prevents the annoying prompt that you're running an application from an unknown publisher for software from zachsaw.com.

From the site:
This also prevents you from getting desensitised from the unknown publisher prompt each time you run MPDN and stops other apps from masquerading as software from zachsaw.com

His quote did not mention "each time", so I assumed the site would be outdated here or something.

So that still applies? UAC prompt for each time you start the application?

[Zachs]

Yes, of course.

I also read http://www.zachsaw.com/mpdn/

From his quote:


From the site:


His quote did not mention "each time", so I assumed the site would be outdated here or something.

So that still applies? UAC prompt for each time you start the application?

If you've installed it via the installer, no.

[Hrxn]

This means that you get two different binary application files, when using the installer in comparison to the standalone archive?

Okay, that is new to me..

[Zachs]

This means that you get two different binary application files, when using the installer in comparison to the standalone archive?

Okay, that is new to me..

It's the same binary. What are you trying to achieve by asking these questions? It's a bit of a waste of time. I've already explained everything that needed to be explained. If you can't accept it, move on.

[Hrxn]

Well, I'm sorry, it is definitely not my intention to waste your (or anyone else's) time.

I am not trying to achieve anything specific here, I am just curious and try to understand.

As I said, a friend recommended MPDN to me, but please imagine the following scenario: An interested user comes to your site, and encounters this:

Step 1: Install this certificate into your OS cert store.

Don't you think that might be a bit deterring to some users?

So I just wondered what might be the reasoning behind this, and here we are...


And now you are telling me that the application binary you get from either the installer or the zip archive is the same (as it should be, obviously), but if I installed it via the installer, I don't get these UAC prompts each time,
If you've installed it via the installer, no.hereby implying that when using the binary from the zip archive, I'll get these prompts. Sorry, but this doesn't make sense to me..

[Zachs]

Well, I'm sorry, it is definitely not my intention to waste your (or anyone else's) time.

I am not trying to achieve anything specific here, I am just curious and try to understand.

As I said, a friend recommended MPDN to me, but please imagine the following scenario: An interested user comes to your site, and encounters this:



Don't you think that might be a bit deterring to some users?

So I just wondered what might be the reasoning behind this, and here we are...


And now you are telling me that the application binary you get from either the installer or the zip archive is the same (as it should be, obviously), but if I installed it via the installer, I don't get these UAC prompts each time,
hereby implying that when using the binary from the zip archive, I'll get these prompts. Sorry, but this doesn't make sense to me..
That's just how it is. UAC works the same way regardless of whether you've installed the cert out not. You'd have to complain to Microsoft if you think it doesn't make sense.

[foxyshadis]

And now you are telling me that the application binary you get from either the installer or the zip archive is the same (as it should be, obviously), but if I installed it via the installer, I don't get these UAC prompts each time,
hereby implying that when using the binary from the zip archive, I'll get these prompts. Sorry, but this doesn't make sense to me..

Once the certificate is installed into the Trusted store, you'll never get another prompt again, zip, installer, or whatever.

If you've installed it via the installer, no.

Might want to check this. I've installed it via the installer many times, and the zachsaw.com cert isn't in my Local Computer Trusted Publishers. Or anywhere else, for that matter.

[Hrxn]

Well, I've tried it for myself in the meantime, maybe this helps a bit to clear things up finally.

I tried both the installer and the standalone archive, and this were the results:

Installer
One UAC prompt when executing the installer. The "yellow" warning prompt, you always receive when running a program from an "unknown publisher" (i.e. not signed) that tries to make changes to the system outside of your specific user dir.

No prompts or warning when running MPDN, not one.

Standalone
Extracted the archive file into a new MPDN dir, a subdir of a custom program folder directly on C:
Made a shortcut and ran the program: No prompts or warnings, not one.

No differences in video playback observable.
And I definitely did not install any cert whatsoever.

So, I still don't get the point..

;)

[Zachs]

Might want to check this. I've installed it via the installer many times, and the zachsaw.com cert isn't in my Local Computer Trusted Publishers. Or anywhere else, for that matter.

Ah I meant installing MPDN via the installer won't trigger UAC when you run MPDN. The cert has to be installed separately (optional of course).

FWIW, I've bitten the bullet and bought a code signing cert. The verification process is a pain though. It'll probably take a bit of time to sort it out.

[Zachs]

So, I still don't get the point..

;)

Read this from OP.

stops other apps from masquerading as software from zachsaw.com. It also stops people with malicious intents from altering MPDN and making it available online as an alternate download

e.g. Someone alters MPDN (via the same technique as a virus would) and posts it online for download. You, unknowingly, download the altered MPDN. There's no way to tell if the MPDN you just downloaded came from me if you don't have the cert installed. If you do, however, you'll get a prompt saying it's from unknown publisher, as opposed to the same prompt for both original and altered MPDN.

[Hrxn]

Yes, I'm aware. An additional security layer that protects the end user from running a manipulated file, if you will. Good for inexperienced users, on the other hand, asking an inexperienced user to install a self-signed cert, I don't know, I think it's a bit of a Catch-22, just like moving the 'question of trust', if I may call it like that, to another layer ;)

I think this is just one of the basic rules of Internet usage. Do not, under no circumstances, download your apps or software or programs from any shady third-party site (Or all third-party sites general). Always use the official source, or an official mirror at least.

Okay, now to some questions that are a lot more interesting ;) :

What are the best settings (or best render scripts, I guess you have to set up both) for upscaling some really old video clips, for example with dimensions of 640x480 pixels, or maybe even worse?

;)

[Zachs]

Hi guys,

I've made a simple MPDN remote for Android (https://play.google.com/store/apps/details?id=com.zachsaw.mpdn_remote&hl=en) (free with no ads). It's missing a lot of features at the moment compared to the Windows one made by DeadlyEmbrace but I hope to add more features when time permits.

Cheers.

[foxyshadis]

Hi guys,

I've made a simple MPDN remote for Android (https://play.google.com/store/apps/details?id=com.zachsaw.mpdn_remote&hl=en) (free with no ads). It's missing a lot of features at the moment compared to the Windows one made by DeadlyEmbrace but I hope to add more features when time permits.

Cheers.

Nice, it connects and works for me. Only Fullscreen, Mute, and Stop seem to work, the other buttons never activate, but it's a strong first showing. Looking forward to even more couch-potatoness in the future!

[Zachs]

Nice, it connects and works for me. Only Fullscreen, Mute, and Stop seem to work, the other buttons never activate, but it's a strong first showing. Looking forward to even more couch-potatoness in the future!

It seems the client never received a valid media length from the server in your case, which is the only reason why the other buttons aren't activated. What does the status text at the bottom of the screen say? e.g. "00:00:01.000/00:02:00.00" (position/duration). And when the media is playing, does the position text change (along with the seekbar even if it's not activated)?

EDIT: Just found out that the server uses the current culture to convert numbers to string. If your phone has a different culture setting to your PC, it'll most likely fail to work - I'll fix the server.

EDIT2: Can you try the latest version for both the app and MPDN extensions? Let me know if you still have the same problem.

[DeadlyEmbrace]

Hi guys,

I've made a simple MPDN remote for Android (https://play.google.com/store/apps/details?id=com.zachsaw.mpdn_remote&hl=en) (free with no ads). It's missing a lot of features at the moment compared to the Windows one made by DeadlyEmbrace but I hope to add more features when time permits.

Cheers.

Very nice :D It looks awesome. (Too bad I can't test it... :/ )
Hoping to get around to iOS one based on Xamarin sometime soon, just have a lot of other things on my hands atm.

[Zachs]

Ah I wrote it with Xamarin too. Works very well on Android. Too bad I haven't got an Apple device or a Mac for development.

[DeadlyEmbrace]

Cool :) Very glad MS bought Xamarin, going to make cross device dev a lot easier.
Yeah... I bought a Mac Mini ages ago, began to think it was a waste but now I can put it to use. Just need to update it.

[Zachs]

I think we should refactor all of that client server code out to a separate unit. I'd like to use the same controller code across all the different platforms. The view code obviously will be different. I'll try to do that when I have some time.

[DeadlyEmbrace]

Agreed, would be easiest if we had a cross platform class library to handle all of the communication logic.
Truth be told I have actually been thinking of rewriting the entire remote control module for a while now, there are quite a few things in there that doesn't work quite as well as it should.
Would really like to switch communication to be JSON based instead of the current custom solution and would also like to implement zeroconf for client setup (no more fiddling with ports)
Will have to keep it as a separate module though, as I don't want to break the existing logic until a full new solution is in place (so plugin, plus remotes)

[Zachs]

Absolutely. The current comms logic don't work too well at all when it comes to the need to escape the delimiters. I'm thinking of something like WCF that is cross platform. I used one that's mono compatible before but can't recall what it's called now...

[Zachs]

Hi guys,

Latest version of the new Android remote supports the following features:
* Remote file browser
* Save a link to the remote and play the link on MPDN

The last one had been on my todo list for a long time - I can finally use the YouTube app to find clips of interest and cast it to MPDN for actual playback.

[Zachs]

Agreed, would be easiest if we had a cross platform class library to handle all of the communication logic.
Truth be told I have actually been thinking of rewriting the entire remote control module for a while now, there are quite a few things in there that doesn't work quite as well as it should.
Would really like to switch communication to be JSON based instead of the current custom solution and would also like to implement zeroconf for client setup (no more fiddling with ports)
Will have to keep it as a separate module though, as I don't want to break the existing logic until a full new solution is in place (so plugin, plus remotes)

This is the RMI framework I used a few years ago for Mono/.NET comms.
https://github.com/hikalkan/scs

It is open source (MIT licensed, so compatible with MPDN Extensions) and has a much better performance compared to WCF.

I'll write the server plugin when I find some time.

[DeadlyEmbrace]

Cool. :) Will take a look at it once I have some time.
Edit: That actually looks like an awesome framework, should make the implementation a lot easier, especially when combined with Xamarin for cross platform implementation.

[sat4all]

hi zachs,

i would love to use mpdn as external player for kodi, but the lack of commad key editor ( like mpc-hc ) holding me back. is there any alternative for mpdn?
another thing, mpdn doesn't not respect lav splitter subtitles preferred language prio for external subs if embedded exist.:confused:

[laforcetrankil]

Hello everyone,

Is there a place where a person explains the configuration options of this software?

Thk's

[Anima123]

Zachs,

Would you please take a look at this avi file? MPDN seems cannot playback it with full fps, it stuck at around 20 fps, while it actually is 29.976.
https://www.dropbox.com/s/gjmuxpmbz3jmty4/oned-649.avi.001?dl=0

Regards,

[Zachs]

Hello everyone,

Is there a place where a person explains the configuration options of this software?

Thk's

There's some documentation available (that said, we could always use more if anyone's willing to contribute).

MPDN Extensions:
https://github.com/zachsaw/MPDN_Extensions/wiki

MPDN itself:
:search:

Zachs,

Would you please take a look at this avi file? MPDN seems cannot playback it with full fps, it stuck at around 20 fps, while it actually is 29.976.
https://www.dropbox.com/s/gjmuxpmbz3jmty4/oned-649.avi.001?dl=0

Regards,

Are you using an audio script? I commited a bug fix just last week that addresses exactly what you've described.

[Anima123]

Are you using an audio script? I commited a bug fix just last week that addresses exactly what you've described.

I have noticed the updates, it does not help with my case though.

PS., Just checked my configuration, no audio script applied.

[Zachs]

I'll take a look at the sample when I have some time

[sat4all]

Enabling 3d-lut script lead to a black screen while i still can hear audio, the same lut file work just fine with madvr.
any idea?

[Zachs]

Enabling 3d-lut script lead to a black screen while i still can hear audio, the same lut file work just fine with madvr.
any idea?

I just tested 3dlut script and it works fine - is your 3dlut file 96MB?

[Zachs]

I have noticed the updates, it does not help with my case though.

PS., Just checked my configuration, no audio script applied.

Hmm the sample timestamps seems to repeat, which isn't something MPDN expected (i.e. the decoder drops the repeats without ever sending it to its renderer). It's probably not a well encoded source, but I've made it work in the next version of MPDN anyway.

[sat4all]

I just tested 3dlut script and it works fine - is your 3dlut file 96MB?

indeed it's.

[Zachs]

Are you using it with any other scripts? I can't produce the problem.

[sat4all]

no other scripts, my pc is an intel nuc i5 haswell. except setting jinc-ar for luma and chroma upscaling & limited range output, i didn't touch anything else in mpdn settings.

[Zachs]

no other scripts, my pc is an intel nuc i5 haswell. except setting jinc-ar for luma and chroma upscaling & limited range output, i didn't touch anything else in mpdn settings.

I tested your exact settings on a lowly Intel HD Graphics 3000 (Sandy Bridge i5) and it still works fine. It works perfectly well on my Xeon Haswell with P4600 too.

[RyaNJ]

Does the "always on top while playing" function work in full screen mode only or also in windowed mode? Also is there some way to keep the play list hidden? It seems to always re-appear at startup even when it has been disabled.

[Garteal]

Also is there some way to keep the play list hidden? It seems to always re-appear at startup even when it has been disabled.
Yes, uncheck "Show playlist on startup" in the Playlist Configuration window. Either press the "O" key with the playlist selected to open this, or press the cogwheel (or go the roundabout way via the PlayerExtensions dialog).

[RyaNJ]

Yes, uncheck "Show playlist on startup" in the Playlist Configuration window. Either press the "O" key with the playlist selected to open this, or press the cogwheel (or go the roundabout way via the PlayerExtensions dialog).

Thanks for that, somehow I missed that option!

[DeadlyEmbrace]

Updated version of the Remote Control for Windows is now available.
It supports both the older RemoteControl extension (Extension 1.27.1 and older) as well as the new extension (1.27.2+)
You can grab a copy here (https://github.com/DeadlyEmbrace/MPDN_RemoteControl/releases/tag/v1.4.0.72)

[Zachs]

Thanks DeadlyEmbrace! I was going to update it at some point but haven't got to it yet.

I'm mucking about with Xamarin.iOS at the moment. iOS programming model isn't as professional as Android's though. Even with Apple's xcode, it's primitive. It's designer is super clunky as you're forced to use the mouse to design the UI.

[Shiandow]

Hi all,

I finally had some success combining the theory behind the SSIM downscaler and SuperRes. Which gave rise to a new experimental algorithm: Structural Simmilarity (SSIM) SuperRes, or SSSR for short. It's still a bit of a work in progress, and I don't want to give too much away yet but suffice it to say it's sharp.

Should be available with the next extensions release, or by grabbing the code from the git repository.

Cheers.

[Zachs]

Fantastic!!! I'll do a quick release in a bit!

EDIT: Looks like a new release will have to wait - I haven't got my signing certificate with me.

[foozoor]

Hi all,

I finally had some success combining the theory behind the SSIM downscaler and SuperRes. Which gave rise to a new experimental algorithm: Structural Simmilarity (SSIM) SuperRes, or SSSR for short. It's still a bit of a work in progress, and I don't want to give too much away yet but suffice it to say it's sharp.

Should be available with the next extensions release, or by grabbing the code from the git repository.

Cheers.

Any screenshot comparisons?

[ryrynz]

Any screenshot comparisons?
If there was he would've linked them. Why not update your extensions and screen shot them yourself? Would take you like five minutes.

[Zachs]

Alright guys I've just released MPDN Extensions v1.28.0. It has Shiandow's SSSR scaler. Go nuts! :)

[Zachs]

Any screenshot comparisons?

Personally I don't think still shots will do it justice - you'll have to see it in motion to appreciate the difference.

[castaway]

Hello Zachs,

I would like to know if MPDN would be able to use the CPU instead of the GPU for part or all of his rendering job. Is there a way to enable this?

Thanks for your player it's impressive!
Cast

[Zachs]

You can use avisynth or something equivalent but the GPU is much more suitable for rendering for good reasons.

[blindbox]

I've tried many other sharpening filters, and none has been able to satisfy me (msharpen, xsharpen, etc). SSSR on the other hand, is AMAZING. I thought it was pretty much impossible for a good sharpening filter to exists.

[Shiandow]

Thanks, although technically it's an upscaling algorithm. It could be made into a sharpener though, that could be an interesting future project.

In the meantime I'm making some adjustments to SSSR. When they're ready, I'd be interested to know what you think of them.

[blindbox]

I think the fact that you're focusing it on upscaling helps shape its capabilities as a sharpening filter. All of my videos that require sharpening always requires upscaling as well, and all other sharpening filters that I've tried dont take that into account. If someone needs it as a proper sharpening filter, there's always downscaling after upscaling.

[Braum]

What are the difference between madvr and mpdn implementation of SSIM downscaler ?

Because for the same configuration I got different render times.

Windows 10 64bits
AMD 7870XT
Driver AMD up to date (16.5.3)

Source : 1920x1080 Bluray
Monitor : 1680x1050

MPDN v2.47.0.3602 64bits :
Chroma scaling : Jinc 4 AR
Downscaling : SSIM 1.50
Deband
Render time : 11ms

Madvr v0.90.20 32bits (because of Reclock) :
Chroma scaling : Jinc AR
Downscaling : 2D SSIM 75% AR relaxed LL
Deband
Render time : 22ms

Why is there such a gap in rendering time ?
Oh I'm dumb it's because of the reduce ringing artifacts fitler.

The following question still stand :

Also I can't find an explanation concerning the 1D - 2D option of SSIM in Madvr, can someone explain it to me please ?

Thank you by advance for taking the time to answer my questions !

[huhn]

madVr jinc should be 6 tab in mpdn and not 4. and even that doesn't mean they are 100 % identical.

the anti ringing filter is nearly for sure not the same too.

you are using an extra AR filter in madVR.
the deband filter are fundamental different too.

so this is not the same configuration.

[ryrynz]

Anyone else losing audio sync when seeking? Happens fairly often for me.

[Anima123]

Anyone else losing audio sync when seeking? Happens fairly often for me.

Me too.

I am not sure if it's just after seeking though, since for quite some time I have been suffering audio sync issues that I have to adjust the audio delay manually to re-gain the sync.

Also I cannot be sure if this phenomenon starts from which version or which LAVFilter version.

[Zachs]

Can you try going back to a release that didn't have the problem? I'd like to find out if it's related to any recent changes.

[Anima123]

Problem is, with the latest versions the audio sync problem seems gone.

Can ryrynz confirms it? Or you're referring to a different issue?

[ryrynz]

And now I know why I've never noticed it until now, it only happens when using DirectSound (I would normally use Sanear)
The problem occurs from 2.44.0 and up, all I have to do is pause and play and sync is lost, FWIW I'm bitstreaming with latest LAV.

[Zachs]

So the problem started from v2.44.0?

Does it only happen when you're bitstreaming?

[madshi]

What are the difference between madvr and mpdn implementation of SSIM downscaler ?

Why is there such a gap in rendering time ?
Oh I'm dumb it's because of the reduce ringing artifacts fitler.

The following question still stand :

Also I can't find an explanation concerning the 1D - 2D option of SSIM in Madvr, can someone explain it to me please ?
Yes, the madVR SSIM AR filter consumes extra power (MPDN doesn't have an extra AR filter for SSIM). But it's not the only difference. huhn already explained some more differences (deband + Jinc taps).

"SSIM" in MPDN is comparable to "1D SSIM" in madVR. "2D SSIM" in madVR consumes much more GPU power. The difference in image quality is rather small, though. So if you want to compare performance, use "1D SSIM" in madVR.

[ryrynz]

So the problem started from v2.44.0?

Does it only happen when you're bitstreaming?

Correct, and yes it seem to only occur while bitstreaming.

Also, with Sanear and Directsound after seeking (sometimes requires multiple seeks) the picture can freeze for up to ten seconds or so while audio continues, unsure if this is any better with 2.43.1 but I could check.

Checked. 2.43.1/1.20.0 doesn't appear to have any issues.

[Braum]

madVr jinc should be 6 tab in mpdn and not 4. and even that doesn't mean they are 100 % identical.

the anti ringing filter is nearly for sure not the same too.

you are using an extra AR filter in madVR.
the deband filter are fundamental different too.

so this is not the same configuration.

Yes, the madVR SSIM AR filter consumes extra power (MPDN doesn't have an extra AR filter for SSIM). But it's not the only difference. huhn already explained some more differences (deband + Jinc taps).

"SSIM" in MPDN is comparable to "1D SSIM" in madVR. "2D SSIM" in madVR consumes much more GPU power. The difference in image quality is rather small, though. So if you want to compare performance, use "1D SSIM" in madVR.
Thank you a lot both of you, it's all clear now. I was also wondering the difference between 1d and 2d SSIM and I got an answer ! [emoji2]

[Zachs]

That's got nothing to do with MPDN. Wrong thread mate.

[Olivier C.]

That's got nothing to do with MPDN. Wrong thread mate.

I am sorry, i've just deleted and moved it to the appropriate thread

[Magik Mark]

Is it possible to include OSD in FSE similar to dsplayer? Something like stop & play button? This would be a great replacement as dsplayer is no longer being developed

[Heaud]

MPDN 2.47.0.3602 64-bit hangs for me when using the bilateral filter with 10-bit h264 content. There is no hanging when bilateral is used as a chroma prescaler in conjuction with nnedi3 prescaling for superres.

[Zachs]

You're right. I'll look into it.

[Zachs]

Shiandow, can you find out what's wrong? I've traced it to the render script framework causing stack overflow when Update() is called.

[Shiandow]

That doesn't narrow it down much, but I'll have a look.

Edit: Should be fixed now.

[Zachs]

Sorry guys I know Shiandow has fixed the issue but I just need to find some time to do a release. Been a really busy time for me.

[Zachs]

MPDN 2.47.0.3602 64-bit hangs for me when using the bilateral filter with 10-bit h264 content. There is no hanging when bilateral is used as a chroma prescaler in conjuction with nnedi3 prescaling for superres.

Just released v1.28.1. This issue should now be fixed.

[Zachs]

That doesn't narrow it down much, but I'll have a look.

Edit: Should be fixed now.

Hmm. I'm still facing this issue...

SubTags is causing stackoverflow by recursively calling itself non-stop.

[Shiandow]

Well I did fix something...

Does the SubTags issue go away when using Bilateral + NNEDI3? Or does it occur regardless?

[Zachs]

Well I did fix something...

Does the SubTags issue go away when using Bilateral + NNEDI3? Or does it occur regardless?

Problem goes away when using Bilateral + NNEDI3.

[Shiandow]

Well I can't reproduce it with the code in my repository, unfortunately I don't really have the time to look into it right now.

[Zachs]

It only happens if the video is 10bits.

[Shiandow]

I've finally managed to reproduce it. It might only happen with newer versions os MPDN (v2.46.0 seems to be fine), no clue why, this issue keeps getting weirder.

By the way it seems MPDN starts the video, resizes the window, and then continues playback. It would be more optimal if the window was resized beforehand, because then the render chain wouldn't need to be built twice.

[Zachs]

Thanks to Shiandow, the bilateral scaler hang issue should now be fixed with v1.28.2.

[Zachs]

By the way it seems MPDN starts the video, resizes the window, and then continues playback. It would be more optimal if the window was resized beforehand, because then the render chain wouldn't need to be built twice.

I'll try to find out what's causing that resize.

[david555]

I'll try to find out what's causing that resize.

hello

i instal mpdn latest version but i can't open it.
Win 10 64bits
4790k
gtx 1080

Do you know why?

thanks

[huhn]

what is happening when you start it?

[david555]

Mpdn close directly.

[amayra]

can you add assFilterMod (http://forum.doom9.org/showthread.php?p=1769770) Support please

[foozoor]

Are you still working on the youtube-dl extension?

[Zachs]

That's already working via std pipe

[foozoor]

That's already working via std pipe
What do you mean by std pipe? Any example?
So no gui way to handle youtube-dl options?

[suanm]

Why don't I download 'MPDN Extensions Installer'?How should I get the software?

[suanm]

Why is it difficult for me to download 'MPDN Extensions Binaries' or 'MPDN Extensions installer'?

[suanm]

Zachs had deleted the files of 'MPDN Extensions Binaries' and 'MPDN Extensions installer' ?

[pirlouy]

Links in the first post are working (github).
Working link as example: https://github.com/zachsaw/MPDN_Extensions/releases/download/1.28.2/Mpdn.Extensions.zip

[Zachs]

Zachs had deleted the files of 'MPDN Extensions Binaries' and 'MPDN Extensions installer' ?

It works fine here. It's hosted on Github.

[suanm]

It works fine here. It's hosted on Github.
Hi,cool dude.I've downloaded the 'MPDN Extensions Binaries' just now.When I play movies with MPDN, the MPDN player prompts a message that MPDN player is incompatible with MPDN Extensions.What should I do?

[suanm]

By the way,It has yet been difficult for me to download the software 'MPDN Extensions installer'?
Who could tell me why this is?

[pirlouy]

It's difficult surely because of your Internet provider.

[suanm]

It's difficult surely because of your Internet provider.
Could you send 'MPDN-Extensions_v1.28.2_Installer.exe' to me via e-mail? Thanks a lot in advance

[pirlouy]

I can't + I'm not interested to be honest.
Plus with my gmail address, I suppose you would not get the mail either + I would be surveyed.

Indeed, your government tends to block github: https://en.wikipedia.org/wiki/Censorship_of_GitHub (oh well, Wikipedia must be blocked too :/)

GL avoiding the Great Firewall.

[Zachs]

The extensions are available on mpdn.zachsaw.com too btw.

[suanm]

that's ok.pirlouy and Zachs.I'm gonna give it up,because I can't get hold of the player 'MPDN.dotnet and its extension version' .Thanks for the hints you both gave

[Sephiroth_FF]

Hi, I have been using MPDN for more than a year and it was really the best player IMHO so a big thanks to @Zachs and everyone involved. I do not know if you are accepting feature request at this time if not, I am very sorry.

First request, I do know LAV splits HDR metadata in version 0.68, so I wandered if you can could have a quick look at this page: https://developer.nvidia.com/high-dynamic-range-display-development you can download the HDR SDK at the bottom of the page (there is also links to blog posts on how NVIDIA implements HDR). It's the source code of NVIDIA's implementation of HDR display mode. Could you see if it is feasible to add support of HDR in MPDN.
Edit: Microsoft has released the Windows 10 Anniversary Update that include DXGI 1.5 with vendor agnostic HDR10 and WCG support at the bottom of this page: https://msdn.microsoft.com/en-us/library/windows/desktop/mt742103(v=vs.85).aspx

Second request, if you ever move the player to DirectX 12 could you also support DX12 Multiadapter there's even been a abstraction layer released recently to support that feature: https://github.com/Microsoft/DirectX-Graphics-Samples/tree/master/Samples/Desktop/D3D12LinkedGpus and of course, OpenCL is already multi-gpu capable, I know all the synchronization will make all this extremely complex and nearly impossible but, hopefully one day your player could the first to support a multi-gpu rig.

Thank you
Sephiroth_FF

Edit: Sorry just found out about Mantis (http://mpdn.zachsaw.com/bug/view_all_bug_page.php) for feature requests will post there later.

[amayra]

if you add image viewer mod with keyboard shortcuts this well be great

[qjb2016]

hi， I'm new user , I'm try the MPDN to the anamorphic lens ,but unsatisfy .

madvr Support the anamorphic lens is very good!

Who knows what's going on? Thanks!

[Shiandow]

So what exactly goes wrong, does the image look stretched or is there something else wrong?

[qjb2016]

So what exactly goes wrong, does the image look stretched or is there something else wrong?

Can not automatically cut the black edge to do stretching
Can not customize the deformation ratio 1.33X or 1.25X .eg

[Anima123]

Super-XBR on its own creates ugly halos but combined with SuperRes it works quite well except with geometric forms and straight lines.

This quote is from other thread, which explained why I didn't like SuperRes much if the video has hard-coded subtitles.

Is there's anything that can be improved on SuperRes to get better quality around lines like hard-coded subtitles?

[Zetti]

One problem with "Multichannel ASIO DirectShow Renderer"??

I have enabled it in Audio Render, How to set the settings for the render??

[Shiandow]

This quote is from other thread, which explained why I didn't like SuperRes much if the video has hard-coded subtitles.

Is there's anything that can be improved on SuperRes to get better quality around lines like hard-coded subtitles?

Well, for a start you could try to explain why you don't you like what SuperRes does to hard coded subtitles, and what you'd prefer. Might also be worth trying SSSR, I've made some small improvements to it, which should be online soon (be sure to reset to the default settings).

[Zachs]

Can not automatically cut the black edge to do stretching
Can not customize the deformation ratio 1.33X or 1.25X .eg
Customise deformation ratio? If it's the same as aspect ratio, you can definitely do that.

Letterboxing options allow you to get rid of black bars.

[Garteal]

Finally tried out SSSR (the latest) and it's generally very pleasing and a tad sharper than SuperRes by itself.
Upon further inspection though (and most noticable on hardcoded characters (http://screenshotcomparison.com/comparison/182600)), it seems to 'miss' some of lighter lines.
Turning the "locality" option down seems to help, but everything else gets even thicker, so not the ideal way to combat it.
There's also some curves that become a lil trapezoid.

Just some pixel-peeving observations here.

Compared to SuperRes, SSSR fortunately doesn't seem to be as resource intensive as I thought.

[Shiandow]

Upon further inspection though (and most noticable on hardcoded characters (http://screenshotcomparison.com/comparison/182600)), it seems to 'miss' some of lighter lines.

Do you mean the small features inside some of the characters? Because I fear those should be considered 'lost'. There's no way to distinguish between a small black feature on a white background and a bigger grey feature on the same background. It might be possible using character recognition and deep neural nets, but that's far beyond what you can reasonably do in real time.


Compared to SuperRes, SSSR fortunately doesn't seem to be as resource intensive as I thought.

Yeah, the algorithm is relatively simple, seems to need less iterations to converge, and doesn't need softening*. So for most use cases it should be faster than SuperRes.

*in fact it functions as a deringing algorithm, which is why by default the prescaler doesn't even have anti-ringing enabled, if you try to do this with SuperRes you'd get massive amounts of ringing.

[aufkrawall]

To the coding gods here:
Wouldn't features like expensive scaling and postprocessing be an ideal combination for GCN's Async Compute (shaders) capabilities?
We have a lot of graphics and compute load at the same time, e.g. NNEDI3 + SuperRes. One could also run postprocessing shaders in compute queue.

[Shiandow]

To the coding gods here:
Wouldn't features like expensive scaling and postprocessing be an ideal combination for GCN's Async Compute (shaders) capabilities?
We have a lot of graphics and compute load at the same time, e.g. NNEDI3 + SuperRes. One could also run postprocessing shaders in compute queue.

Well, performing tasks asynchronously is only advantageous if MPDN somehow couldn't manage to drive the GPU to maximum load by delivering tasks sequentially, however unlike with games MPDN can easily predict its tasks in advance and should be able to push tasks to the GPU as fast as the GPU can handle them. Now maybe Async compute allows the GPU to prepare the next shader while its busy computing the previous one, which could improve performance, but I haven't seen anything to support that.

[Garteal]

Do you mean the small features inside some of the characters? Because I fear those should be considered 'lost'. There's no way to distinguish between a small black feature on a white background and a bigger grey feature on the same background. It might be possible using character recognition and deep neural nets, but that's far beyond what you can reasonably do in real time.
Yep I can understand that, though to be fair SSSR does seem to do its job there as well.

I was more towards this
http://i.imgur.com/ySRphYr.png
Red gets better coverage than green, but I suppose red has a 'darker' color which fits the criteria?


Yeah, the algorithm is relatively simple, seems to need less iterations to converge, and doesn't need softening*. So for most use cases it should be faster than SuperRes.

*in fact it functions as a deringing algorithm, which is why by default the prescaler doesn't even have anti-ringing enabled, if you try to do this with SuperRes you'd get massive amounts of ringing.
Cool, cause I actually enabled anti-ringing but saw no effect. Guess I can leave that off.

[Zachs]

Well, performing tasks asynchronously is only advantageous if MPDN somehow couldn't manage to drive the GPU to maximum load by delivering tasks sequentially, however unlike with games MPDN can easily predict its tasks in advance and should be able to push tasks to the GPU as fast as the GPU can handle them. Now maybe Async compute allows the GPU to prepare the next shader while its busy computing the previous one, which could improve performance, but I haven't seen anything to support that.
To top it off, AsyncCompute is a DX12 feature. A lot of the image processing shaders are still in SM3. And the last time I checked, DX12's interop support isn't great because it's not meant to be used with other APIs.

The conclusion was to use AsyncCompute, the most logical choice would be to ditch the current engine and rewrite it in DX12. All shaders would have to be rewritten as well. Not sure how well DX12 does interop with OpenCL and still benefit from AsyncCompute.

[lanzorg]

The conclusion was to use AsyncCompute, the most logical choice would be to ditch the current engine and rewrite it in DX12. All shaders would have to be rewritten as well. Not sure how well DX12 does interop with OpenCL and still benefit from AsyncCompute.

And why not Vulkan?
It's still pretty verbose but it will not be limited to Windows!

[Zachs]

Vulkan is an AMD specific API... I could be wrong but from memory it also doesn't do AsyncCompute when running shaders designed for other APIs.

[aufkrawall]

Alright, thanks for your evaluation, guys. Appreciate it!

[lanzorg]

Vulkan is an AMD specific API... I could be wrong but from memory it also doesn't do AsyncCompute when running shaders designed for other APIs.
Vulkan is derived from and built upon components of AMD's Mantle API.
But it evolved a lot the last years with Khronos.

[Zachs]

Ah I was thinking of Mantle... But the problem is still the same. With a video player like MPDN where a lot of interops need to take place per frame, AsyncCompute really doesn't work very well.

[lanzorg]

Ah I was thinking of Mantle... But the problem is still the same. With a video player like MPDN where a lot of interops need to take place per frame, AsyncCompute really doesn't work very well.
Are you sure it will need to keep this OpenCL interop with Vulkan or DX12?

[Zachs]

Are you sure it will need to keep this OpenCL interop with Vulkan or DX12?
Definitely. And not just OpenCL, it's everything else from SM3 to DirectCompute. MPDN extensions expect interops to work transparently.

[Shiandow]

To top it off, AsyncCompute is a DX12 feature. A lot of the image processing shaders are still in SM3. And the last time I checked, DX12's interop support isn't great because it's not meant to be used with other APIs.

The conclusion was to use AsyncCompute, the most logical choice would be to ditch the current engine and rewrite it in DX12. All shaders would have to be rewritten as well. Not sure how well DX12 does interop with OpenCL and still benefit from AsyncCompute.

On a somewhat tangentially related note, I noticed it's possible to compile ps_3_0 shaders as ps_5_0 when you enable the backwards compatibility flag (/Gec when using fxc). Not sure if that's something useful for MPDN; I haven't checked if it affects performance (although in theory it could). In theory this would allow shaders to use more registers etc. although if you go past the official limits the shader will no longer compile for older GPUs (although the alternative is that the shader won't work at all).

[Shiandow]

I was more towards this
http://i.imgur.com/ySRphYr.png
Red gets better coverage than green, but I suppose red has a 'darker' color which fits the criteria?

Could just be a matter of getting lucky with the pixel boundaries, depending on where a thin black line ends up it could be a 2 pixel wide gray line or a 1 pixel wide black line, detecting the difference isn't particularly easy (sometimes possible, and SSSR tries it best, but it's impossible to always get it right).

[XRyche]

Still no Windows Media Remote Control support?

[Zachs]

Just a quick update - I'll release a new version MPDN extensions when I get the chance. Currently on a very limited data plan.

[Anima123]

Now I am quite satisfied with the following render chain settings, thanks Shiandow & Zachs bringing us flexible settings that make it possible:
<ListOfPresetOfITextureFilterIRenderScript xmlns:yaxlib="http://www.sinairv.com/yaxlib/">
<PresetOfITextureFilterIRenderScript Name="Bilateral" Guid="326be196-0220-4e3f-b1f8-ee9efa5fd033">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Shiandow.Bilateral.BilateralUi">
<Settings>
<Strength>0.5</Strength>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
<PresetOfITextureFilterIRenderScript Name="Deband" Guid="6554618d-475a-4478-b45b-4c992856e11f">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Shiandow.Deband.DebandUi">
<Settings>
<maxbitdepth>8</maxbitdepth>
<power>0.5</power>
<grain>False</grain>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
<PresetOfITextureFilterIRenderScript Name="SSSR" Guid="d8d36e98-df33-4a02-8934-258ec6948a56">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Shiandow.SuperRes.SSSRUi">
<Settings>
<PrescalerGroup>
<SelectedIndex>2</SelectedIndex>
<Hotkey></Hotkey>
<Options>
<PresetOfITextureFilterIRenderScript Name="EWA Sinc-Jinc" Guid="95efb650-f467-492a-b023-3bec706f790e">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Mpdn.EwaScaler.EwaScalerScaler">
<Settings>
<TapCount>Six</TapCount>
<AntiRingingEnabled>False</AntiRingingEnabled>
<AntiRingingStrength>0.85</AntiRingingStrength>
<Scaler yaxlib:realtype="Mpdn.Extensions.CustomLinearScalers.SincJinc" />
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
<PresetOfITextureFilterIRenderScript Name="Super-xBR (Fast)" Guid="0836bfbf-8d01-4ce0-92bd-761a9f71d65f">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Hylian.SuperXbr.SuperXbrUi">
<Settings>
<EdgeStrength>1</EdgeStrength>
<Sharpness>1</Sharpness>
<FastMethod>True</FastMethod>
<ThirdPass>False</ThirdPass>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
<PresetOfITextureFilterIRenderScript Name="Super-xBR" Guid="0f559010-216e-4aa4-a5d1-b43445e963d9">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Hylian.SuperXbr.SuperXbrUi">
<Settings>
<EdgeStrength>0.7</EdgeStrength>
<Sharpness>0.4</Sharpness>
<FastMethod>False</FastMethod>
<ThirdPass>False</ThirdPass>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
<PresetOfITextureFilterIRenderScript Name="NEDI" Guid="16455dd6-59b3-4a99-884c-d05fc81c7a72">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Shiandow.Nedi.NediScaler">
<Settings>
<AlwaysDoubleImage>False</AlwaysDoubleImage>
<ForceCentered>False</ForceCentered>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
<PresetOfITextureFilterIRenderScript Name="NNEDI3" Guid="590ede3c-69d1-4b74-a0aa-95363ea4e9cc">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Shiandow.NNedi3.NNedi3Scaler">
<Settings>
<Neurons1>Neurons32</Neurons1>
<Neurons2>Neurons16</Neurons2>
<CodePath>UnrolledVectorDot</CodePath>
<Structured>False</Structured>
<ChromaScalers>
<ChromaScalerPreset Name="None" Guid="08f594fe-5e3e-46e2-948a-de164536d741">
<Script yaxlib:realtype="Mpdn.Extensions.Framework.RenderChain.RenderChainUi+IdentityRenderChainUi">
<Settings />
</Script>
</ChromaScalerPreset>
<ChromaScalerPreset Name="Bilateral" Guid="4820a55f-f314-42e8-a042-c5e82a22c3e3">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Shiandow.Bilateral.BilateralUi">
<Settings>
<Strength>0.5</Strength>
</Settings>
</Script>
</ChromaScalerPreset>
<ChromaScalerPreset Name="ChromaScaler" Guid="4279ba6d-24a9-45f8-b447-b74445a17efd">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Shiandow.Chroma.ChromaScaler">
<Settings>
<B>0.3333333</B>
<C>0.3333333</C>
<Preset>MitchellNetravali</Preset>
</Settings>
</Script>
</ChromaScalerPreset>
<ChromaScalerPreset Name="EWA Chroma Scaler" Guid="764e9035-d67a-4fcb-8755-c5af2979189c">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Mpdn.EwaScaler.EwaScalerChromaScalerUi">
<Settings>
<TapCount>Four</TapCount>
<AntiRingingEnabled>False</AntiRingingEnabled>
<AntiRingingStrength>0.85</AntiRingingStrength>
<Scaler yaxlib:realtype="Mpdn.Extensions.RenderScripts.Mpdn.EwaScaler.EwaScaler+JincScaler" />
</Settings>
</Script>
</ChromaScalerPreset>
<ChromaScalerPreset Name="NNEDI3 Chroma Doubler" Guid="d7b65505-6951-474e-8d26-6fc7ef0dbc0c">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Shiandow.NNedi3.Chroma.NNedi3ChromaScaler">
<Settings>
<Neurons1>Neurons16</Neurons1>
<Neurons2>Neurons16</Neurons2>
<CodePath>ScalarMad</CodePath>
<Structured>False</Structured>
</Settings>
</Script>
</ChromaScalerPreset>
<ChromaScalerPreset Name="OpenCL NNEDI3 Chroma Doubler" Guid="5f70af35-dbf9-42d4-83f6-3512b36ecf4b">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Mpdn.OclNNedi3.Chroma.OclNNedi3ChromaScaler">
<Settings>
<Neurons1>Neurons16</Neurons1>
<Neurons2>Neurons16</Neurons2>
</Settings>
</Script>
</ChromaScalerPreset>
<ChromaScalerPreset Name="SuperChromaRes" Guid="680279f1-9c0c-4a50-8cfa-7e3553d39514">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Shiandow.SuperRes.SuperChromaResUi">
<Settings>
<Passes>1</Passes>
<Strength>1</Strength>
<Softness>0</Softness>
<Prescaler>True</Prescaler>
<LegacyDownscaling>False</LegacyDownscaling>
</Settings>
</Script>
</ChromaScalerPreset>
</ChromaScalers>
<ChromaScalerGuid>53534bbf-4749-4599-98c0-603302772b44</ChromaScalerGuid>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
<PresetOfITextureFilterIRenderScript Name="OpenCL NNEDI3" Guid="ccfae239-6c17-4c01-b9ee-6faba49d28eb">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Mpdn.OclNNedi3.OclNNedi3Scaler">
<Settings>
<Neurons1>Neurons16</Neurons1>
<Neurons2>Neurons16</Neurons2>
<ChromaScalers />
<ChromaScalerGuid>00000000-0000-0000-0000-000000000000</ChromaScalerGuid>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
</Options>
</PrescalerGroup>
<Passes>2</Passes>
<OverSharp>0</OverSharp>
<Locality>4</Locality>
<Mode>Hybrid</Mode>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
</ListOfPresetOfITextureFilterIRenderScript>

[SamuelMaki]

Hey! Figured to give MPDN an another chance, but have couple questions first. Is there a similar output as with madVR Overlay mode? Aka taking care of the .icc profile and bypassing Windows. Also getting an error message http://anony.ws/i/2016/09/05/Sieppaa.png when installed latest 64-bit MPDN with the installer and the latest extensions installer. Using Windows 10 64-bit updated to the Anniversary if it matters.

[Shiandow]

By the way, I've added some new options to the extensions that need testing. They're available in the latest version of the git repository.

Firstly, I've made the linear light processing of SSSR optional, this could improve quality in some cases, although it has a tendency to exaggerate highlights (especially when using the 'sharp' mode).

Secondly, I've added a new mode for the Bilateral chroma scaler, which seems to work better in 'difficult' cases. It uses a somewhat more general mathematical model, which means it's far more resilient to bad sources. The model I used for the previous version didn't always hold (although if it did then the quality was extremely good, sometimes even better than the new mode).

Please give those a try, and tell me what you think.

Cheers.

[Shiandow]

Hey! Figured to give MPDN an another chance, but have couple questions first. Is there a similar output as with madVR Overlay mode? Aka taking care of the .icc profile and bypassing Windows. Also getting an error message [...] when installed latest 64-bit MPDN with the installer and the latest extensions installer. Using Windows 10 64-bit updated to the Anniversary if it matters.

The exclusive mode should be similar to MadVR's overlay mode, although I'm not sure if it bypasses windows' processing completely.

Not sure why you're getting an error. As far as I can tell the latest version of MPDN combined with the latest extensions should just work. It's also unfortunate it either works or fails completely, the loading process really needs to be made more robust at some point. Anyway, can you double-check you've got the right versions, and maybe try using the standalone files if that doesn't work?

[SamuelMaki]

The exclusive mode should be similar to MadVR's overlay mode, although I'm not sure if it bypasses windows' processing completely.

Not sure why you're getting an error. As far as I can tell the latest version of MPDN combined with the latest extensions should just work. It's also unfortunate it either works or fails completely, the loading process really needs to be made more robust at some point. Anyway, can you double-check you've got the right versions, and maybe try using the standalone files if that doesn't work?

Thanks for the reply!
Yes, I have tried with couple versions now actually (just the two newest installers of both MPDN and the Extensions). There actually was a previous user who mentioned this, but their English was not very clear so might have got misunderstood. The same happens with standalone/portable files, it shows that screen on load and couple more error messages after a while: http://anony.ws/i/2016/09/05/Sieppaa2.png
http://anony.ws/i/2016/09/05/Sieppaa3.png
Sorry about the Finnish language, Google translate gives:
"The destination has caused an exception"
"A reference to an object can not determine the instance of the object"
That first (second actually, after clicking "ok" from the first in previous message) will kill the application if I click it away, that second appears when I go to the settings of MPDN and try to browse the Extensions. I have successfully used MPDN quite some while back before even the November update, so that probably not helping much. I ticked clean install and tried without it as well.


EDIT: Same happens with 2.45.5 and the 1.25 Extensions (API 8) so the new API probably not to blame.

[Shiandow]

Right, clearly something has gone pretty badly wrong.

Now I suspect ticking 'clean install' has already done this, but if you want to make really sure you could try uninstalling MPDN. Then removing everything in "%LocalAppData%\MediaPlayerDotNet" (make a copy if you want to use those settings later). Copy the MPDN standalone files in an empty directory. Try to see if it works. Then copy the Extension files. And see if it still fails.

You could also try to use DebugView (https://technet.microsoft.com/en-us/sysinternals/debugview.aspx) while running MPDN to see if that gives any information on what went wrong.

[SamuelMaki]

Alright, removed the folder manually, uninstalled MPDN and downloaded the latest portable version of both, installed to my Documents directory in an empty folder just to be safe, turned the debugging on inside MPDN and got the following error messages:
http://pastebin.com/xEr9pVr3
Again, sorry for the Finnish language, but the codes of the errors should be enough to understand? Basically it is complaining about mscorlib as far as I can tell.

[Shiandow]

Hmm, those error messages just look like the aftermath of whatever went wrong, not the actual problem itself. If you use DebugView it might be possible to get some more detailed information.

[Zachs]

I'll have a quick look to see if loader errors are dumped out onto debug stream. I think they are. Will make sure.

Wonder if it's possible to redirect via command line to a file...

[Shiandow]

A crashdump would probably help a lot.

As far as I can tell the errors SamuelMaki managed to recover are not the source of the problem, only weird thing I can find is that it's initializing the "Settings" field only to subsequently dispose of it. It might be better to initialize the field during the creation to prevent this kind of thing from happening (will probably add this to a pull request when I have some more important changes), but the main problem seems to be that somewhere it fails to create an object, which isn't supposed to happen.

It would also be nice if one error during the loading didn't take out all extensions. That would at least make it easier to tell which one failed. And eventually we're going to need some way to selectively enable/disable extensions, anyway.

[Zachs]

The trouble is those sort of errors that occur only on one PC usually mean the framework itself also fail to get loaded.

The only thing that would be helpful for that particular instance is the loader error message which doesn't come with the exception message itself.

[Shiandow]

In this case it seems the framework managed to load, at least partially. Something seems to have gone wrong when initializing the extensions.

So, what stages are there in loading the extensions? I suspect there will first be some kind of initialisation where you load the dll, after or during that step the static classes are created, then after that the extension classes are created, and initialized. It should at least be possible to figure out at what stage things go wrong.

[Zachs]

Yeah I agree there needs to be better error reporting when it comes to extensions loading...

I'll see what I can do.

[foozoor]

It seems I am not alone with the black screen issue on Intel/AMD hybrid laptops!
http://imgur.com/a/LJDnm

[sauma144]

It seems I am not alone with the black screen issue on Intel/AMD hybrid laptops!
http://imgur.com/a/LJDnm
I am the one who created this imgur album.
Is there a way to fix this weird issue?

[Zachs]

Is this something like Optimus?

[sauma144]

Is this something like Optimus?

Probably but I am not an expert.
Only the intel hd 3000 is working with MPDN.
If I set to use the AMD card I get blackscreen. :\

[Zachs]

It might only work if you have an external monitor. Otherwise it's dependent on the driver doing the right thing to copy the video frame buffer to Intel one for final display.

Hybrid GPU setup as is currently implemented is the worst technology ever!

[sauma144]

I know the technology is ugly and caused a lot of weird behaviors.
I mostly watch videos on external monitor and I get blackscreen too.

However it works with other players/renderers perfectly.
I don't know why MPDN doesn't identify all the GPUs in Graphics Adapter.
Is it working with SLI/CrossFire setups?

[Zachs]

I know the technology is ugly and caused a lot of weird behaviors.
I mostly watch videos on external monitor and I get blackscreen too.

However it works with other players/renderers perfectly.
I don't know why MPDN doesn't identify all the GPUs in Graphics Adapter.
Is it working with SLI/CrossFire setups?
MPDN simply presents all adapters as given by D3D. If it's not there, it means the driver hasn't reported it to the API.

Crossfire/SLI unfortunately doesn't work too well when it comes to having to present a frame at a required time window. The APIs don't allow individual control of the cards.

[sauma144]

So why does MPDN have a combobox to select the graphics adapter?
I have never seen this kind of thing with other players or renderers and it seems MPDN forces itself to use the selected adapter and that's why I get blackscreen when I use the dedicated GPU.

[Zachs]

So how would you tell MPDN to use a particular GPU when you have more than one without a combobox?

[sauma144]

I don't know, I am wondering how other players do.
Maybe they let the drivers do the things, I don't know if there is a way to set a primary GPU in SLI/CrossFire setups.
I really want to use MPDN with my dedicated GPU on this laptop, so if there are things to test, I am in.

[Zachs]

Other players automatically choose one. Most simply use the primary adapter.

[sauma144]

So why not adding an option in MPDN to let the driver deciding?
Maybe it will fix blackscreen with Intel/AMD hybrid setups.

[Zachs]

Driver doesn't decide. Player does.

[sauma144]

So why not an option to use the primary adapter like other players do by default? ;)

[nevcairiel]

So how would you tell MPDN to use a particular GPU when you have more than one without a combobox?

Use whichever is in control of the screen your window is on?
Thats how all other players work, really. You can't really use any other.

[Zachs]

Well personally I like the ability to choose. That's why I did it that way. If you have the window span across multiple screens, you'll have to pick just one of the adapters anyway.

[sauma144]

So you are not interested to add this "feature" to see if it fixes blackscreen issue with intel/amd hybrid laptops. :(

[Zachs]

It won't be magically fixed by making MPDN auto select from the combobox - MPDN would then simply pick one out of the list, which you could do on your own right now.

[sauma144]

Did you see the imgur album?
http://m.imgur.com/a/LJDnm
MPDN doesn't recognize/list the AMD card.
Now I understand it will need more time than just make an auto-selection in the combobox.

[Zachs]

Did you see the imgur album?
http://m.imgur.com/a/LJDnm
MPDN doesn't recognize/list the AMD card.
Now I understand it will need more time than just make an auto-selection in the combobox.

If there's only a single adapter listed, then obviously the problem is with AMD not reporting its GPU properly. If MPDN can only see the Intel GPU, then what makes you think it can magically use the AMD one?

[sauma144]

If MPDN can only see the Intel GPU, then what makes you think it can magically use the AMD one?
So why does it work with madvr, kodi, mpv and so on? It's really weird! No?
I think the way MPDN identifies the GPUs doesn't work or is too restrictive with hybrid laptops.

[Zachs]

MadVR works fine with AMD hybrid?

[sauma144]

Yes absolutely!

[Zachs]

Not sure whats going on then. But without an AMD hybrid it'll be hard to do anything...

[sauma144]

Can I do something to help you?
As already said, I really want to use MPDN with the dedicated GPU on this laptop!

[Zachs]

Unfortunately I'm very busy at the moment with a lot of personal things happening at the same time.
If you can do coding, one thing you could try is the enumeration routine for D3D9 to see if it returns the AMD adapter.

[sauma144]

If you can do coding, one thing you could try is the enumeration routine for D3D9 to see if it returns the AMD adapter.

Sorry, but no. :(

However I have to disable this option with mpv to make the AMD card working without crash.
I don't know if it can help you.


Allows DirectComposition when using the ANGLE backend (default: yes). DirectComposition implies flip-model presentation, which can improve rendering efficiency on Windows 8+ by avoiding a copy of the video frame. mpv uses it by default where possible, but it can cause poor behaviour with some drivers, such as a black screen or graphical corruption when leaving full-screen mode. Use "no" to disable it.


Furthermore MPDN crashes when entering exclusive fullscreen with the AMD card, so with blackscreen.

[Zachs]

I thought the problem was the AMD card isn't even listed? How do you enter full screen with the AMD card then?

[sauma144]

How do you enter full screen with the AMD card then?

I double click inside the window or CTRL+ALT but it crashes only when I tick exclusive fullscreen inside the MPDN settings.

I have blackscreen but it runs, I have the sound and also the OSD with Decoder Queue, Render Queue, ...

[mecedo]

Hi. I want to report bug in MPDN because I think "Feature/Bug Tracker" page is abandoned.

After pausing movie for longer then 20 seconds playing is stuttered. The only way to fix is close and reopen MPDN

[Nevilne]

Searched this thread/player website for this and couldn't find anything - are custom skins planned/available?

[Shiandow]

Hi. I want to report bug in MPDN because I think "Feature/Bug Tracker" page is abandoned.

After pausing movie for longer then 20 seconds playing is stuttered. The only way to fix is close and reopen MPDN

The feature tracker still works, but Zachs is rather busy so response time can be a bit slow. He'll get round to it eventually.

Anyway, some more details might help, is the video stuttering, or just the audio? Does the OSD (Ctrl+J) show anything? Are you using any non-default settings?

Searched this thread/player website for this and couldn't find anything - are custom skins planned/available?

Unfortunately the answer is no, to both your questions.

[Zachs]

Sorry guys I know this is long overdue but I've just released a new version of the extensions with Shiandow's latest changes.
Enjoy!

[Shiandow]

Some of the things available with the new extensions:


A new experimental chroma scaling method (not as new as it once was, but feedback is still appreciated).
Linear light processing for SSSR is now optional. Disabling it slightly exaggerates highlights, but seems to be sharper.

As always, the more feedback I get the more I can improve the algorithms, or come up with new ones.

[aufkrawall]

New bileteral is much more stable regarding "broken" sources like really low-quality cartoons. Sometimes it still "decides wrong" though, but it's a hell of an improvement. And it's also very sharp.
With an animation of a news broadcast (the one which I also posted in regard of banding) it looks a bit aliased, which the old bileteral doesn't do.

Having gamma light for SSSR is really good. But it doesn't look very different to me compared to SuperRes in madVR. Maybe even if it's a bit better, it's still much more important which prescaler is used (which imho only can be NNEDI3).

Some other observations:
-SSIM downscaling after SSSR doesn't work, it always uses MPDN's internal scaler.
-My old comprehensive profile-script sometimes produces just pages of error messages. :(
http://pastebin.com/UJBwP0A9

[Shiandow]

New bileteral is much more stable regarding "broken" sources like really low-quality cartoons. Sometimes it still "decides wrong" though, but it's a hell of an improvement. And it's also very sharp.
With an animation of a news broadcast (the one which I also posted in regard of banding) it looks a bit aliased, which the old bileteral doesn't do.

Thanks for the feedback, glad you're seeing some improvement. And yeah, there are some cases where things go wrong still, especially with low quality sources. To some extent this is inherent to reconstructive methods, so there will always be a trade off between sharpness and artefacts.

I used the sample you're talking about for tests, and I agree that the previous method gave a better result, but that only seemed to work well for that particular image.


Having gamma light for SSSR is really good. But it doesn't look very different to me compared to SuperRes in madVR. Maybe even if it's a bit better, it's still much more important which prescaler is used (which imho only can be NNEDI3).

Sharpness might not differ too much, but in my opinion SSSR is more 'balanced'; it suppresses ringing and even (to some extent) aliasing. It also converges a lot faster (especially using the 'hybrid' method), so you need fewer passes to get a good result.


Some other observations:
-SSIM downscaling after SSSR doesn't work, it always uses MPDN's internal scaler.
-My old comprehensive profile-script sometimes produces just pages of error messages. :(
http://pastebin.com/UJBwP0A9

I can't reproduce either of those, but I'm probably not using the exact same versions of MPDN and the extensions. The first one simply shouldn't happen, but it could be the OSD that's wrong, are the rendering times affected if you add / remove SSIM?

Not sure what's happening to your script (seems to be working fine, as far as I can tell) but one or two of those pages of error messages would probably clear things up.

[sauma144]

Sorry guys I know this is long overdue but I've just released a new version of the extensions with Shiandow's latest changes.
Enjoy!

Thank you for the update!

Could you or someone else update the presets or create new ones with the new algorithms and pretty results?

I still have blackscreen (with sound) issue when using the AMD GPU on a Intel/AMD hybrid laptop. Did you find what is causing this? Maybe should I add an issue in the bug tracker.

[aufkrawall]

I used the sample you're talking about for tests, and I agree that the previous method gave a better result, but that only seemed to work well for that particular image.

Well, if the aliasing could be reduced further.. :)
That sample was a good demonstration where the old bileteral outperformed every other algorithm. And it's true "real world" content, the frames originate untouched from the news site (and this pathetic quality is the best they offer).
Btw: This clip still screams for a more aggressive deband filter. :p


Sharpness might not differ too much, but in my opinion SSSR is more 'balanced'; it suppresses ringing and even (to some extent) aliasing. It also converges a lot faster (especially using the 'hybrid' method), so you need fewer passes to get a good result.

I will create some comparisons and then judge again. :)


I can't reproduce either of those, but I'm probably not using the exact same versions of MPDN and the extensions. The first one simply shouldn't happen, but it could be the OSD that's wrong, are the rendering times affected if you add / remove SSIM?

I think I made a mistake, just adding SSIM to each end of my script chain profiles seems to work.


Not sure what's happening to your script (seems to be working fine, as far as I can tell) but one or two of those pages of error messages would probably clear things up.
My fault, I forgot to copy the Adaptive Sharpen shaders to the correct path (it's been some time since I last used MPDN).
The script still seems to be working well with newest extensions.

[Shiandow]

My fault, I forgot to copy the Adaptive Sharpen shaders to the correct path (it's been some time since I last used MPDN).
The script still seems to be working well with newest extensions.

I did at some point allow Image Processor to use shaders in external folders as well. Putting the Adaptive Sharpen in some dedicated folder might make things easier in future.

Out of interest what does the OSD read if your script uses it's most complicated process? I'm planning to improve the OSD somewhat, because I'm pretty sure it's currently unreadable in any even vaguely complicated processing chain.

[Shiandow]

Btw: This clip still screams for a more aggressive deband filter. :p

If it had a voice it'd be screaming all right, not sure if it'd still be intelligible.

[aufkrawall]

It doesn't look too messy to me :) :
http://abload.de/img/mpdncrs58.png (http://abload.de/image.php?img=mpdncrs58.png)

What's messy is when I edit my script chains for various scaling factors and fps, but I guess that's simply the price to pay for having all the freedom one needs.
Somehow clicking through less dialogues would be nice, but I have no idea if less clicking would work without undesired side effects.

Edit: Any idea why SSIM isn't used?
http://abload.de/thumb/ssimw5s44.png (http://abload.de/image.php?img=ssimw5s44.png)
Adding Resizer before doesn't help.

[Shiandow]

It doesn't look too messy to me :) :
What's messy is when I edit my script chains for various scaling factors and fps, but I guess that's simply the price to pay for having all the freedom one needs.
Somehow clicking through less dialogues would be nice, but I have no idea if less clicking would work without undesired side effects.


Inside the script chain you can right click and use the "Configure..." item to go through the script chains, but I agree that it isn't ideal. Maybe adding this somewhere in the main context menu would help? *

Edit: Any idea why SSIM isn't used?
http://abload.de/thumb/ssimw5s44.png (http://abload.de/image.php?img=ssimw5s44.png)
Adding Resizer before doesn't help.

Looks like SSSR already resized the image to the target size, there's simply nothing left to do for SSIM. Maybe you meant to put SSIM after the NNEDI3 prescaler?

*EDIT: Ah I see what's happening, the "Conditionals" don't 'publish' their subfilters, that's unfortunate. I also noticed you pretty much always use a script chain inside the conditional, might make more sense to just make that the default and turn Conditionals into a kind of Script Chain.

[BetA13]

It doesn't look too messy to me :) :
http://abload.de/img/mpdncrs58.png

What's messy is when I edit my script chains for various scaling factors and fps, but I guess that's simply the price to pay for having all the freedom one needs.
Somehow clicking through less dialogues would be nice, but I have no idea if less clicking would work without undesired side effects.



may i ask, how do you get the Render Chain window to display? that is really helpfull, i need this :)

greetz

[Anima123]

may i ask, how do you get the Render Chain window to display? that is really helpfull, i need this :)

greetz

Ctrl + K, which is a well known shortcut from the start.

[Shiandow]

What's messy is when I edit my script chains for various scaling factors and fps, but I guess that's simply the price to pay for having all the freedom one needs.
Somehow clicking through less dialogues would be nice, but I have no idea if less clicking would work without undesired side effects.


The following feature might save a few clicks:

http://imgur.com/hxM0Gnt.jpg

It's even backwards compatible, although I'll likely deprecate the old method at some point (but not before I get around to improving the conditional filter's GUI).

I'll probably include this with the next extension update.

[Garteal]

I personally just use a Script Group with Deband, SuperChromaRes and SSSR and call it a day, but it looks like I should spend some time exploring the wonderful system you've created!

[Shiandow]

I personally just use a Script Group with Deband, SuperChromaRes and SSSR and call it a day, but it looks like I should spend some time exploring the wonderful system you've created!

To be fair, that's pretty much what I use for most videos as well (although I put in SSIM downscaling as well). You might want to try the new Bilateral on it's own some time.

[aufkrawall]

I think new biliteral could be very interesting for videos which have higher resolution, e.g. at least 720p. I suppose the higher the resolution, the less likely are crucial errors. Or the errors would be much smaller and hard to notice.

[BetA13]

Ctrl + K, which is a well known shortcut from the start.

Thanks a lot.

-------------------------------------

I also have an rather strange need/ question.

is there a way to manually set the RGB Output levels?
i really would dig an option like madvr has:
https://abload.de/img/4eos79.png

i have a, well, lets call it a strange TV and setting levels from 4 (5) - 255 is perfect for me. no details lost and colors are great..

with 0 - 255 i have lost details.
with 16 - 255, the darks get crushed.

im really happy with mpdn, but this is one option im missing..
Would this be possible? Or maybe, is there a other way i can handle this? (plugin, shader,etc)..

greetz :)

[Shiandow]

I also have an rather strange need/ question.

is there a way to manually set the RGB Output levels?
i really would dig an option like madvr has:
https://abload.de/img/4eos79.png

i have a, well, lets call it a strange TV and setting levels from 4 (5) - 255 is perfect for me. no details lost and colors are great..

with 0 - 255 i have lost details.
with 16 - 255, the darks get crushed.

im really happy with mpdn, but this is one option im missing..
Would this be possible? Or maybe, is there a other way i can handle this? (plugin, shader,etc)..

greetz :)

Something like that can easily be done using a custom shader. The following code should do the trick:


sampler s0 : register(s0);

#define BlackLevel 4
#define WhiteLevel 255

// -- Main Code --
float4 main(float2 tex : TEXCOORD0) : COLOR{
float4 c0 = tex2D(s0, tex);

c0.rgb = (c0.rgb*255.0 + BlackLevel) / (WhiteLevel + BlackLevel);

return c0;
}


Creating a new shader from MPDN itself is somewhat tedious, but you can just put the code in a file called "FixOutputLevels.hlsl" or something and load it in an Image Processing script. If you then want to change the code you can edit it from within MPDN.

[BetA13]

niceee.. thanks a lot Shiandow !1!!
works just like expected :)
im really a noob in this kind of thing, so dont mind me, haha...I do think i want to get into this a bit. i love fiddling with those shaders and reshade also..

Thanks..

[Shiandow]

Hi everyone,

Since recently I'm now also involved in the development of MPDN itself. I'm still getting familiar with the code, so any fundamental changes will be a ways off.

I did, however, manage to make some small improvements to MPDN's rendering. To quote the changelog:

Improved fluid motion by timing frames more accurately
Improved dithering methods (especially ordered dithering)
Fixed: Pixels were misaligned when there was an unequal number of blank lines on either side of the image.


These changes require some testing (especially the timing related changes), so they've been added into a new Test Build (https://mpdn.zachsaw.com/Test%20Builds/3603/). If things go well these changes will eventually be included in the next release. Please let us know if the quality improves noticeably / barely / not at all, or more importantly if this new version causes any problems.

Cheers,
Shiandow

[Anima123]

Playback is more smooth with the new test version, with or without fluid motion, which I'm guessing is partly due to the more accurate time frames changes.

MPDN used to play 59.97 fps files with some jagger, now it's all good with the latest test build. Shiandow, I'd say it's the right direction. BTW, recently MPDN became my main video-payer, so any improvements are appreciated.

Edit: As for the dithering, since I'm not sensitive to it, there's not much to mention. Where to look at exactly?

[Shiandow]

Edit: As for the dithering, since I'm not sensitive to it, there's not much to mention. Where to look at exactly?

Yeah, it's kind of hard to see, especially when everything is working correctly. The previous dithering method occasionally created some ugly patterns, which were especially visible in dark gradients. It also added more noise than the current method.

[Zachs]

Hi everyone,

Since recently I'm now also involved in the development of MPDN itself. I'm still getting familiar with the code, so any fundamental changes will be a ways off.

I did, however, manage to make some small improvements to MPDN's rendering. To quote the changelog:

Improved fluid motion by timing frames more accurately
Improved dithering methods (especially ordered dithering)
Fixed: Pixels were misaligned when there was an unequal number of blank lines on either side of the image.


These changes require some testing (especially the timing related changes), so they've been added into a new Test Build (https://mpdn.zachsaw.com/Test%20Builds/3603/). If things go well these changes will eventually be included in the next release. Please let us know if the quality improves noticeably / barely / not at all, or more importantly if this new version causes any problems.

Cheers,
Shiandow

Thanks Shiandow! You're a valuable dev to any project, as always!

[aufkrawall]

Could you guys please provide me with a hint how to use SSIM downscaling after using image doubling/quadrupling?
imho it would really be most intuitive if just adding SSIM after an image doubler in the script chain would do the trick.

[Shiandow]

Could you guys please provide me with a hint how to use SSIM downscaling after using image doubling/quadrupling?
imho it would really be most intuitive if just adding SSIM after an image doubler in the script chain would do the trick.

Well, that should do the trick. Is there some case where it's not working?

Yeah, it's kind of hard to see, especially when everything is working correctly. The previous dithering method occasionally created some ugly patterns, which were especially visible in dark gradients. It also added more noise than the current method.

To expand on that, here is a pretty pathological example: old dithering (http://i.imgur.com/SkJLAB3.png), new dithering (http://i.imgur.com/v01LLnO.png). In most cases the ordered dithering wasn't that bad, but even in the best of cases the new method should be better.

Thanks Shiandow! You're a valuable dev to any project, as always!

You're welcome :).

[aufkrawall]

Well, that should do the trick. Is there some case where it's not working?

Hm, what does the arrow before Triangle indicate then?
http://abload.de/img/ssr0du94.png (http://abload.de/image.php?img=ssr0du94.png)
If it was just about offset correction, shouldn't it point to the left and right?

[Anima123]

It means downscaling I believe.

[aufkrawall]

Yes, but it shouldn't use Triangle (bilinear) since SSIM is in the script chain after SSSR.

Edit: With new Fluid Motion there are still sometimes delayed frames reported:
http://abload.de/img/delayed8iuh8.png (http://abload.de/image.php?img=delayed8iuh8.png)
Happens mostly when the GPU load is very high, but queues are always full.

[Shiandow]

Yes, but it shouldn't use Triangle (bilinear) since SSIM is in the script chain after SSSR.

If you want that you should use SSIM inside the prescaler, not after SSSR. Right now it only tries to downscale after using SSSR. I'm working on a new Tag system that hopefully will make the distinction clearer. What you want would look something like this:
http://i.imgur.com/TTF4i7h.png


Edit: With new Fluid Motion there are still sometimes delayed frames reported:
[...]
Happens mostly when the GPU load is very high, but queues are always full.

Seems it's too busy with something to get the frame out in time. Probably not an easy fix, unless it turns out to be another timing issue.

[aufkrawall]

If you want that you should use SSIM inside the prescaler, not after SSSR. Right now it only tries to downscale after using SSSR. I'm working on a new Tag system that hopefully will make the distinction clearer. What you want would look something like this:
http://i.imgur.com/TTF4i7h.png

Could you please provide me with an example script? :o

[Shiandow]

Could you please provide me with an example script? :o

The following should do the trick:


<ListOfPresetOfITextureFilterIRenderScript xmlns:yaxlib="http://www.sinairv.com/yaxlib/">
<PresetOfITextureFilterIRenderScript Name="SSSR" Guid="de825cc3-332d-4848-9d5f-588774c4d8ae">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Shiandow.SuperRes.SSSRUi">
<Settings>
<PrescalerGroup>
<SelectedIndex>0</SelectedIndex>
<Hotkey></Hotkey>
<Options>
<PresetOfITextureFilterIRenderScript Name="OpenCL + SSIM" Guid="be40a0fc-efde-473d-9e5d-219af6aeef83">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Mpdn.ScriptChain.ScriptChainScript">
<Settings>
<Options>
<PresetOfITextureFilterIRenderScript Name="OpenCL NNEDI3" Guid="943b652d-977e-43db-a2fa-a97db3ec4dbd">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Mpdn.OclNNedi3.OclNNedi3Scaler">
<Settings>
<Neurons1>Neurons16</Neurons1>
<Neurons2>Neurons16</Neurons2>
<ChromaScalers />
<ChromaScalerGuid>00000000-0000-0000-0000-000000000000</ChromaScalerGuid>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
<PresetOfITextureFilterIRenderScript Name="SSIM downscaler" Guid="7e600065-baec-44bd-9394-4c376a96b0c3">
<Script yaxlib:realtype="Mpdn.Extensions.RenderScripts.Shiandow.SSimDownscaling.SSimDownscalerUi">
<Settings>
<SoftDownscaling>False</SoftDownscaling>
<Strength>0.5</Strength>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
</Options>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
</Options>
</PrescalerGroup>
<Passes>2</Passes>
<LinearLight>False</LinearLight>
<OverSharp>0</OverSharp>
<Locality>4</Locality>
<Mode>Hybrid</Mode>
</Settings>
</Script>
</PresetOfITextureFilterIRenderScript>
</ListOfPresetOfITextureFilterIRenderScript>

[aufkrawall]

Dam, I again forgot that one can add script chains as prescalers. *facepalm*
Working now. :o

[Garteal]

Hi everyone,
Since recently I'm now also involved in the development of MPDN itself. I'm still getting familiar with the code, so any fundamental changes will be a ways off.

I did, however, manage to make some small improvements to MPDN's rendering. To quote the changelog:

Niiice! Going to to be using this build for playback when I get to it.

I also tried your new deband method, and it definitely produces better results.
Old (https://abload.de/img/old-dtxunu.png) // New (https://abload.de/img/new-dzguwp.png)
ScreenshotComparison (http://screenshotcomparison.com/comparison/189490)

Edit: Btw enabling "add grain" results in a compile error. Somethings undefined.

[Shiandow]

I also tried your new deband method, and it definitely produces better results.
Old (https://abload.de/img/old-dtxunu.png) // New (https://abload.de/img/new-dzguwp.png)
ScreenshotComparison (http://screenshotcomparison.com/comparison/189490)

Edit: Btw enabling "add grain" results in a compile error. Somethings undefined.

You're early, I hadn't even pushed that to the official repository yet. Glad you like the results though. That compilation error should be fixed in the official version. I'm probably removing the 'add grain' option entirely (unless it turns out to be way more popular than anticipated, but then I'd still make some changes).

[aufkrawall]

New deband, is it christmas already?
Thx a lot, Shiandow. :)

[Anima123]

Is the new deband filter in render script or not? I haven't noticed it yet.

[Shiandow]

Is the new deband filter in render script or not? I haven't noticed it yet.

It's on my personal github, it hasn't really been officially released yet.

[ryrynz]

I also tried your new deband method, and it definitely produces better results.
Old (https://abload.de/img/old-dtxunu.png) // New (https://abload.de/img/new-dzguwp.png)
ScreenshotComparison (http://screenshotcomparison.com/comparison/189490)

Probably not a good test there as there's no real banding in that image to speak of. Most of the change occurs in the top left where I can see that the pixels are averaged out more but it's basically imperceptible.

[huhn]

both images use an deband filter so no banding.

[Garteal]

Yup what ^ said.

Speaking of which, Shiandow, is there anything that can be done against banding flickering? It's quite noticeable in motion, especially on sources with horrible banding.

[Shiandow]

Speaking of which, Shiandow, is there anything that can be done against banding flickering? It's quite noticeable in motion, especially on sources with horrible banding.

Not easily, usually the source itself is flickering as well. It might also be a consequence of the recursive debanding method I use. Or it may even be inherent to debanding large gradients, since those change quite a bit with small changes.

Anyway, if possible I'd like to wait with reviewing the new debanding method until there's actually an official release of it.

[aufkrawall]

Here someone created some juddertest videos:
http://www.oc-burner.de/ftp/Videos/juddertest/
I sometimes notice jumping even though MPDN doesn't report any anomalies. madVR looks smoother to me.

[ryrynz]

Yup what ^ said.


I forgot when I started posting that I was looking at two debanded images.. any way, what I said still stands that the resulting images are near identical.
I'm sure you could find something that shows more significantly the improvements in the new version :P

[Zachs]

Just released Shiandow's latest improvements v2.48.0. Enjoy!

[Shiandow]

Hi all,

Please let me know if this latest release fixes the remaining timing related issues with fluid motion.

Also, I've noticed that the difference between SSSR in gamma light, and SSSR in linear light is a lot smaller than I originally though, could anyone confirm or deny this?

[Hrxn]

FYI, the automatic update check of MPDN is currently not working: 500 - Internal Server Error

[Magik Mark]

Hi Shiandow!

Is there hope to add "Stop, Pause, Play & Exit" OSD buttons during FSE?

Thanks a lot for your work here. We really appreciate it

[Shiandow]

FYI, the automatic update check of MPDN is currently not working: 500 - Internal Server Error

Seems to be some trouble with the server, the links on the first page should still work.


Is there hope to add "Stop, Pause, Play & Exit" OSD buttons during FSE?


That should be possible, I'll have a look when I have some time. It could probably be done with an extension, if anyone else wants to give it a try.

[Magik Mark]

That should be possible, I'll have a look when I have some time. It could probably be done with an extension, if anyone else wants to give it a try.

Thanks a Lot! Looking forward to it

[aufkrawall]

Please let me know if this latest release fixes the remaining timing related issues with fluid motion.

It seems to, have to try a few more times to confirm.


Also, I've noticed that the difference between SSSR in gamma light, and SSSR in linear light is a lot smaller than I originally though, could anyone confirm or deny this?
In madVR with SuperRes, the difference is only very noticable with large scaling factors (doubling/quadrupling/...) and content with fine bright or dark structures.

[Anima123]

It seems the audio volume control didn't apply when starting a video file, maybe just before the buffer been filled.

Edit: Just noticed that it happened with the Sanear Audio renderer, not with the good ole directsound audio renderer.

[Zachs]

In that case it's most likely a Sanear bug.

[Shiandow]

Seems the new debanding algorithm may have to wait a bit. I just figured out what the technique I'd partially rediscovered is called (pyramid decomposition). Using it properly I managed to get an algorithm with even better quality in about a day. It would take too long to get the new version completely finished, so I will probably just release an experimental version so you guys can compare.

[kingpage]

There's new upscaling algorithm called RAISR (Rapid and Accurate Image Super Resolution).

Romano, et al ‎2016 Paper (https://arxiv.org/pdf/1606.01299v3.pdf)

Supplementary Results (https://drive.google.com/file/d/0BzCe024Ewz8ab2RKUFVFZGJ4OWc/view)

Google Research Blog Article (https://research.googleblog.com/2016/11/enhance-raisr-sharp-images-with-machine.html)

It seems more than fast enough for real-time video rendering and yet comparable to the very best upscaling techniques out there (such as A+ & SRCNN).

They report a runtime of about 20-30ms for 630K-pixel upscaled images on a Xeon 6-core CPU. They also implemented another version that used mobile GPU and ran at 200Mpix/s, which is like 10ms at 1080p or 40ms at 2160p.

Based on the above, it should be a lot faster on a modern desktop GPU (<5ms). Anyone with the know-how interested to write a script to see how it performs in MPDN?

Edit: temporal artifacts could be a problem as it's never been used for video rendering, perhaps some additional filter is needed to ensure consistent changes to each frame. The authors did say their new method not only doesn't introduce artefacts, it can reduce them via learning. What do you think?

In terms of getting the trained filters and/or code, the lead author can be reached here (http://www.cs.technion.ac.il/~yromano/#software).

[Shiandow]

It might be fast, but so far I've not seen any results from it that actually impress me. In most cases their image seems similar to what I can get using SSSR.

[kingpage]

What about in terms of artifacts? SSSR is very sharp, but it seems to exacerbate compression artifacts and film grain, etc. RAISR is supposed to be able to "suppress the aliasing in addition to alleviating moderate compression artifacts", by using moderate (95 quality JPEG) level of compression in training. I loved SSSR but for some movies, it seems to oversharpen everything. How do you use SSSR (for both upscaling e.g. from 720p to 1080p and native resolution e.g. 1080p)?

I used to use Jinc for both upscaling and downscaling, now I use Sinc-Jinc for that. I still don't know whether I like NEDDI3, SSSR, Super-xBR or SuperRes better. How would you describe their differences?

BTW, I've seen screenshots of render chain OSD, is there a hotkey for that?

Edit: I came across another one with great potential. FSRCNN (http://mmlab.ie.cuhk.edu.hk/projects/FSRCNN.html), even better (higher PSNR & SSIM) than SRCNN but much faster (seems to be quite a bit faster than RAISR from a quick glance). There's matlab code and training code and dataset included. I would love to be able to implement these myself, but I don't really know where to start. Is there somewhere I can learn how to write render scripts for MPDN or in general?

[BetA13]

what about in terms of artifacts? Sssr is very sharp, but it seems to exacerbate compression artifacts and film grain, etc. Raisr is supposed to be able to "suppress the aliasing in addition to alleviating moderate compression artifacts", by using moderate (95 quality jpeg) level of compression in training. I loved sssr but for some movies, it seems to oversharpen everything. How do you use sssr (for both upscaling e.g. From 720p to 1080p and native resolution e.g. 1080p)?

I used to use jinc for both upscaling and downscaling, now i use sinc-jinc for that. I still don't know whether i like neddi3, sssr, super-xbr or superres better. How would you describe their differences?

Btw, i've seen screenshots of render chain osd, is there a hotkey for that?



ctrl/strg - k

[Shiandow]

What about in terms of artifacts? SSSR is very sharp, but it seems to exacerbate compression artifacts and film grain, etc. RAISR is supposed to be able to "suppress the aliasing in addition to alleviating moderate compression artifacts", by using moderate (95 quality JPEG) level of compression in training. I loved SSSR but for some movies, it seems to oversharpen everything. How do you use SSSR (for both upscaling e.g. from 720p to 1080p and native resolution e.g. 1080p)?

Lowering compression artefacts would be nice, but is not really the job of a scaler if you ask me.

The default settings of SSSR are designed for scaling 720p to 1080p. Be sure to use the 'soft' or 'hybrid' modes, as the 'sharp' mode is indeed pretty sharp. In theory it can't oversharpen (it tries to ensure the contrast in the final image is the same as in the original) but you never know; using a softer prescaler (super-xbr, or even NEDI) could help.


I used to use Jinc for both upscaling and downscaling, now I use Sinc-Jinc for that. I still don't know whether I like NEDDI3, SSSR, Super-xBR or SuperRes better. How would you describe their differences?

They're sharper, without adding much aliasing or ringing (although SuperRes can have some problems with ringing). They each have their own advantages and disadvantages though, you really need to decide for yourself what you like best.


Edit: I came across another one with great potential. FSRCNN (http://mmlab.ie.cuhk.edu.hk/projects/FSRCNN.html), even better (higher PSNR & SSIM) than SRCNN but much faster (seems to be quite a bit faster than RAISR from a quick glance).

So far I'm more partial to methods based on this paper (https://pdfs.semanticscholar.org/7df0/39049948d54fd1f4d75526743f315cf4865f.pdf).


There's matlab code and training code and dataset included. I would love to be able to implement these myself, but I don't really know where to start. Is there somewhere I can learn how to write render scripts for MPDN or in general?

Unfortunately, no. That said most of the work is basically knowing how to write and use shaders, the easiest way to start with that is probably modifying existing shaders (like the ones in the MPDN extensions folder) to see what effect this has, then move on to writing your own shaders and finally learn how to connect everything to MPDN. Now I can help with that last step, but learning how to write shaders is going to take most of the time. Also, implementing something like FSRCNN is going to take some effort, you need to somehow fit the calculations into shaders (I suppose you could use OpenCL instead, but I'm not too sure you'll find it easier).

[kingpage]

So far I'm more partial to methods based on this paper (https://pdfs.semanticscholar.org/7df0/39049948d54fd1f4d75526743f315cf4865f.pdf).


https://github.com/huangzehao/Super-Resolution.Benckmark

Benchmark results tend to suggest external example-based methods are superior to internal example-based methods. I can sort of understand why it is harder to implement something like FSRCNN than a method that doesn't need training. Literature seems to favour convolutional networks.

I know SSSR stands for Structural Similarity (SSIM) SuperRes. Could you tell me which papers are SSSR, Super-xBR and SuperRes based on?

Edit: Found your post about SuperRes (http://forum.doom9.org/showthread.php?p=1685124#post1685124) which is based on Lukin et al. 2006 (http://www.graphicon.ru/html/2006/proceedings/papers/we08_94_LukinKrylovNasonov.pdf). xBR stands for scale by rules, Super-xBR is developed by Hylian 2015 (https://drive.google.com/file/d/0B_yrhrCRtu8GYkxreElSaktxS3M/view?pref=2&pli=1).

[BetA13]

hey there,

just a quick question:

is there any differnces in using dx11 api over dx9e?

do i loose any features? or is it more optimized?

the thing why i ask, is because fullscreen exclusive mode only works with dx9e here..when i try it with dx11 my screen goes fullscreen But i only see 2/3 of the picture, its kinda too big, dunno..

And i was just wondering, if i stay with dx9 since it works for me, is there anything thats different now besides the api?

THX..

[kingpage]

Is the forced image doubling option only available for NEDI? Is there a way to force it to use SSSR, SuperRes, etc for 1080p material rendering at native 1080p? Nothing seems to show up on the render chain OSD except chroma upscaling (or forced NEDI) unless the source material is <1080p.

[Shiandow]

just a quick question:

is there any differnces in using dx11 api over dx9e?

do i loose any features? or is it more optimized?

The only difference is how it tells windows to display the images, you shouldn't lose/gain any features by changing the renderer. Performance also shouldn't differ much.

[BetA13]

very good..

Thanks a lot, like allways, greetz Beta

[Shiandow]

Is the forced image doubling option only available for NEDI? Is there a way to force it to use SSSR, SuperRes, etc for 1080p material rendering at native 1080p? Nothing seems to show up on the render chain OSD except chroma upscaling (or forced NEDI) unless the source material is <1080p.

Currently there isn't a way to do that. Although it's fairly easy to change NNEDI3's code to make it always double if you want it to. All you'd have to do is remove the line that checks if the target size is bigger than the input.

[RyaNJ]

Hi Guys.

Since yesterday I've been getting an issue that I have absolutely no idea how to resolve. The issue prevents MDPN from even loading correctly - let alone playing anything. Every time when I attempt to load MPDN the following screen is displayed:

http://i.imgur.com/EBCS6c5.png

Followed by a bit of other information, none of which helps me figure out exactly what has apparently happened here. This box appears right after the optimising extensions message is shown. Upon closing MPDN the following message then appears:

http://i.imgur.com/Qdtyaec.png

However the program doesn't unload and has to be forcibly terminated.

What on earth has happened here and does anyone have any ideas how I can fix this?

[dipje]

All I know is I had the same thing, tried searching for it but couldn't find anything (it doesn't help that doom9's search-in-thread seems a bit quick to reject certain words these days).

Then I took the 32bit version, which worked fine.


To be honest - and it might be my old-skool coder thing and not getting all the dotNet stuff - but the 'extensions' are supplied as a bunch of files + support DLLs (Mpdn.Extensions.dll + Mpdn.Extensions.Framework.dll + a whole lot of stuff in the Libs\ folder). If I download the 'mpdn extensions' there are only single versions of these files, and I don't think it's possible to make a single DLL for both x86 and x64.. so the whole extensions seem to be compiled for x86, and of course they are not going to work with x64. Just me thinking out loud, no clue if this is true or not.

MPDN x64 just the bare installer (so, MediaPlayerDotNet_x64_2_48_0_3603_Installer.exe) works fine. The moment I try to install extensions (or some of the extensions) I get the same as you. Using the x86 version seems to 'fix' it.

[RyaNJ]

All I know is I had the same thing, tried searching for it but couldn't find anything (it doesn't help that doom9's search-in-thread seems a bit quick to reject certain words these days).

Then I took the 32bit version, which worked fine.


To be honest - and it might be my old-skool coder thing and not getting all the dotNet stuff - but the 'extensions' are supplied as a bunch of files + support DLLs (Mpdn.Extensions.dll + Mpdn.Extensions.Framework.dll + a whole lot of stuff in the Libs\ folder). If I download the 'mpdn extensions' there are only single versions of these files, and I don't think it's possible to make a single DLL for both x86 and x64.. so the whole extensions seem to be compiled for x86, and of course they are not going to work with x64. Just me thinking out loud, no clue if this is true or not.

MPDN x64 just the bare installer (so, MediaPlayerDotNet_x64_2_48_0_3603_Installer.exe) works fine. The moment I try to install extensions (or some of the extensions) I get the same as you. Using the x86 version seems to 'fix' it.

Using the full 32-bit install does seem to fix it, the issue seems to only exist on the 62-bit version. Well spotted.

Another bug to report here. If I set MDPN to auto-load a video from last played position on startup it continues playing from there - even if I attempted to launch another media player file externally.

I would have expected it to do something different, like open the file that was expected to be launched instead. Is this a bug or a feature?

[Shiandow]

Hi Guys.

Since yesterday I've been getting an issue that I have absolutely no idea how to resolve. The issue prevents MDPN from even loading correctly - let alone playing anything. Every time when I attempt to load MPDN the following screen is displayed:

[...]

Followed by a bit of other information, none of which helps me figure out exactly what has apparently happened here. This box appears right after the optimising extensions message is shown. Upon closing MPDN the following message then appears:

[...]

However the program doesn't unload and has to be forcibly terminated.

What on earth has happened here and does anyone have any ideas how I can fix this?

All I know is I had the same thing, tried searching for it but couldn't find anything (it doesn't help that doom9's search-in-thread seems a bit quick to reject certain words these days).

Then I took the 32bit version, which worked fine.


Using the full 32-bit install does seem to fix it, the issue seems to only exist on the 62-bit version. Well spotted.

My best guess is that the settings are somehow causing a problem. Which shouldn't happen obviously, but that might explain why it behaves differently for the 32- and 64-bit versions.

You could check this by removing the settings for the 64 bit version and see if that fixes things.


Another bug to report here. If I set MDPN to auto-load a video from last played position on startup it continues playing from there - even if I attempted to launch another media player file externally.

I would have expected it to do something different, like open the file that was expected to be launched instead. Is this a bug or a feature?

That's a feature of the playlist right? You should probably post this as an issue on the MPDN Extensions github, hopefully Garteal can find some time to fix it, if it's not working as intended.

[dipje]

My best guess is that the settings are somehow causing a problem. Which shouldn't happen obviously, but that might explain why it behaves differently for the 32- and 64-bit versions.

You could check this by removing the settings for the 64 bit version and see if that fixes things.

I had this issue while _trying out_ MPDN for the very first time. It was a clean install, never installed it before. 64bit was my first attempt with the extensions-installer. Had the same issue, unworkable. Tried searching around, found no solution besides cleaning stuff from the usual setting-places (looking for mpdn-related folders in %appdata%, %localappdata%, %programdata% and registry). Nothing fixed it besides going 32bit.

[Zachs]

It sounds to me like a corrupt windows installation. Installing it on a cleanly installed windows should fix it.
To test the theory, just try it out on a VM.

[EncodedMango]

I had this issue while _trying out_ MPDN for the very first time. It was a clean install, never installed it before. 64bit was my first attempt with the extensions-installer. Had the same issue, unworkable. Tried searching around, found no solution besides cleaning stuff from the usual setting-places (looking for mpdn-related folders in %appdata%, %localappdata%, %programdata% and registry).
Same here, though I haven't tried MPDN yet because of that 64 bit failure, never thought it was an x64 issue.

[dipje]

I see my post was lost again.

@Zachs you were right that the 64bit version works fine with extensions in a clean installed win10-anni version.

But my win10 install is far from corrupt . The fact that the 32bit version works fine proves that it can't be a disabled service or something.

Is there a way to get more debugging info ? See which extension fails to load and why ?

I'm guessing it's a DLL file in my system32 or my devenvironment is causing a conflict.

I'm apparently not the only one with the issue.

[trandoanhung1991]

I have a question.

I distinctly remember it being said that when a script chain is activated, it won't go through the included upscalers/downscalers. Is this not the case with script groups?

Because what I'm seeing is the downscaling script group, and on top of it the internal downscaler, whatever it's set to.

EDIT: Nevermind, brainfarted. I had only chroma scaling in the script groups, so luma scaling is still being done internally. D'oh!

[Hrxn]

[..]

I distinctly remember it being said that when a script chain is activated, it won't go through the included upscalers/downscalers. Is this not the case with script groups?

Because what I'm seeing is the downscaling script group, and on top of it the internal downscaler, whatever it's set to.

[..]

What do you mean by included scalers? Why would the script not use the scalers?

[Zachs]

I see my post was lost again.

@Zachs you were right that the 64bit version works fine with extensions in a clean installed win10-anni version.

But my win10 install is far from corrupt . The fact that the 32bit version works fine proves that it can't be a disabled service or something.

Is there a way to get more debugging info ? See which extension fails to load and why ?

I'm guessing it's a DLL file in my system32 or my devenvironment is causing a conflict.

I'm apparently not the only one with the issue.

Try running dbgview and see what error messages MPDN dump out when it encounters the error. Yes I agree it would be helpful if the loader can isolate failed extensions.

My feeling is something it requires in 64 bit is missing but present on 32. Gut feeling says it's the JavaScript engine failing to resolve dependencies.

[Sephiroth_FF]

Hi @Zachs, I am really sorry for bothering you again about this but unfortunately, NVIDIA has made thier HDR SDK exclusive to DirectX 11 and as far as I know MPDN is the only DirectX 11 video renderer. I would like to know if you can have a look at this SDK https://developer.nvidia.com/high-dynamic-range-display-development and see if it is feasible to integrate it in MPDN.

Once again sorry for my insistence
Thank you very much
Sephiroth_FF

[Shiandow]

Not sure exactly how difficult it will be to add support for HDR displays, but either way adding support for HDR footage should probably be done first.

[anta777]

How to use multiple render scripts ?

[Shiandow]

Use the 'Script Chain' renderscript. That one can chain multiple renderscripts.

[Hrxn]

By the way, while we're at it:

I've been using Script Chains for quite a while now, but not really satisfied with the results so far. That is I manage to get a "different" result all the time, but not "better" :)

Could anyone give some examples (or point to a place that does) for script chains that are suitable for improving 720p material, or 480p, or maybe even worse?

[anta777]

Use the 'Script Chain' renderscript. That one can chain multiple renderscripts.

Thank you very much, Shiandow!
What format supports 3dlut mpdn?
My 3dlut from madvr - I receive an error.
HRESULT 0x80004005
http://saveimg.ru/pictures/05-12-16/7376a25b1287028b94174a71abbf4a30.png
http://saveimg.ru/show-image.php?id=04cd2481536f6109468506a6288d1f64

[Shiandow]

Thank you very much, Shiandow!
What format supports 3dlut mpdn?
My 3dlut from madvr - I receive an error.
HRESULT 0x80004005

http://saveimg.ru/show-image.php?id=04cd2481536f6109468506a6288d1f64

The error message isn't particularly helpful, but it doesn't seem to be coming from MPDN. It looks like DirectX is having some trouble allocating a 3D texture. At least, that's what I make of it.

What GPU are you using?

[anta777]

hd 5770 windows 7 x64 non aero theme
3dlut - http://rgho.st/82zrhKrLV

If I use bilateral (render script) it replaces chroma upscaler in video render?

There is something like showrendersteps (madvr) in mpdn ?

[Shiandow]

Hmm, well it doesn't look like the GPU is the problem then. If it was the 3DLUT I would expect it to crash earlier in the process, but maybe it somehow manages to do something silly (e.g. allocating an impossibly large/small texture). Any chance you could send a copy of the 3DLUT?

[anta777]

No problem
http://rgho.st/82zrhKrLV

[anta777]

No problem.
http://rgho.st/82zrhKrLV

[Shiandow]

hd 5770 windows 7 x64 non aero theme
3dlut - http://rgho.st/82zrhKrLV

Well, the good news is that it's working on my version of MPDN. The bad news is that I don't know what went wrong.


If I use bilateral (render script) it replaces chroma upscaler in video render?

Bilateral is itself a chroma scaler, so that's supposed to happen.


There is something like showrendersteps (madvr) in mpdn ?

With Ctrl+K you get a short summary. More detailed information isn't available yet, although I might add it at some point since it could be useful for debugging.

[Zachs]

Hmm try reducing render queue and try not to use maximum image quality. Those will easy up your GPU memory. 3dlut needs a lot of memory on the GPU itself. It would sound like you've run out of it.

[anta777]

Ctrl+K or Ctrl+J ?
all scripts in render script replace analog scripts in video render (luma upscaler, chroma upscaler, luma downscaler, chroma downscaler) ?

[anta777]

Hmm try reducing render queue and try not to use maximum image quality. Those will easy up your GPU memory. 3dlut needs a lot of memory on the GPU itself. It would sound like you've run out of it.

did not help

[Shiandow]

Ctrl+K or Ctrl+J ?
all scripts in render script replace analog scripts in video render (luma upscaler, chroma upscaler, luma downscaler, chroma downscaler) ?

Both Ctrl+K and Ctrl+J should show some information. And yeah quite a few of the renderscripts replace some of the default functionality, usually with better results.

Do you happen to know if your GPU has 512MB memory or 1GB? Or to be sure you could try using something like GPU-Z (https://www.techpowerup.com/gpuz/) to see if your GPU is running out of memory.

[anta777]

512 MB
my 3dlut work in mavdr

Ctrl+K did not work

[Shiandow]

512 MB
my 3dlut work in mavdr

Ctrl+K did not work

If you go to Options > General > Player Extensions, do you see something called 'Scale Chain OSD painter'? If you doubleclick that you should be able to enable the renderscript OSD manually. If it's not there then something went wrong when installing the extensions.

Still don't know what's happening with the 3dlut. Lack of memory makes the most sense, but it works in MadVR so obviously you should have enough total memory. Could you try using GPU-Z to check memory usage anyway? Then we can either rule out memory, or confirm that that is the problem.

[anta777]

http://gpuz.techpowerup.com/16/12/05/nq4.png

with 3dlut
http://gpuz.techpowerup.com/16/12/05/bay.png

Why minimal core and memory clock?
Options > General > Player Extensions, i see something called 'Scale Chain OSD painter
and doubleclick that you should be able to enable the renderscript OSD manually?
but no effect/
Ctrl+K not work

[Zachs]

Sorry i may have misinterpreted the error message. It would seem that the driver is faulting in the most unlikeliest place. I have a 4350 which i have tested 3dlut. I think it's the driver you're using. It would be legacy now for the 5000 series as it would for mine wouldn't it?

[anta777]

Radeon Software Crimson Edition Beta Crimson Edition 16.2.1 Beta 3/1/2016
and AMD OCL

[dipje]

That crimson beta driver breaks opencl for older cards. You need the older catalyst drivers for working opencl.

[anta777]

Opencl working, i use AMD OCL.

[Zachs]

Try going to non beta drivers. It's pretty obvious from the error message it's due to a driver problem.

[BetA13]

is there a way to get Superxbr as chroma scaler/doubler?
or are the other chroma scaler superior to sxbr?

greetz

[Shiandow]

Well, the jury is still out on whether it's superior, but you could try out the new Bilateral chroma scaling method. Just select the Bilateral chroma scaler, and in its options set the mode to 'Krig'.

[RyaNJ]

While digging into some issues I also seem to have found something interesting, hopefully you guys will be able to help me make heads or tails of it!

I noticed that my computer seemed to be preforming poorly when MPDN was open. Initially I thought it may have been some strange interaction between my Anti Virus software and the program so I disabled it and rested. Same result. MPDN, even when paused, seems to be maxing out the disk IO for my primary drive and I have absolutely no clue why this would be the case.

To check it is definitely MPDN and not something else I closed down everything I could (Anti Virus software included) and tested. Same result. I've attached a copy of the screen capture from Windows Task Manager along with this post. The red line is the point at which MPDN was loaded. The high usage basically continued forever after the program is loaded, sometimes fluctuating but mostly hovering at the max usage. This seems a little odd to me so hopefully you guys will be able to help me figure out what is causing this strangeness.

Cheers.

Ryan.

[Shiandow]

Attachments take forever to approve here, for images it generally a lot easier to just use something like imgur.

Windows 7 has a resource monitor that can show you which file it's accessing, I'm assuming later versions of windows have something similar. That information could help a lot in narrowing down the problem.

[dipje]

Opencl working, i use AMD OCL.

then it's not opencl but some other gpgpu language.

But all gpu-assisted plugins / programs (knlmeanscl, svp, madvr) stopped working on those 16.2.1 drivers. There were attempts to 'merge' some files from the catalyst drivers to get 'best of both worlds' but it never really worked.

The moment I restored the latest 15.x catalyst drivers for my legacy card all gpu-assisted stuff works again.

Make of it what you will, but if the auther says "it is clearly a driver problem" and you're using a driver that is known to break gpu-assisted stuff on 'legacy' AMD cards, I'd think twice :)

[RyaNJ]

Attachments take forever to approve here, for images it generally a lot easier to just use something like imgur.

Windows 7 has a resource monitor that can show you which file it's accessing, I'm assuming later versions of windows have something similar. That information could help a lot in narrowing down the problem.


Fair point. http://imgur.com/zKgKIpp

I'll try some other things first, a clean reinstall of the software and whatnot. You know how temperamental Windows can be. It's just utterly bizarre and I've never seen anything quite like this before.

[Shiandow]

Fair point. http://imgur.com/a/LdrLS

I'll try some other things first, a clean reinstall of the software and whatnot. You know how temperamental Windows can be. It's just utterly bizarre and I've never seen anything quite like this before.

Looking at the performance tab doesn't give much information, but it looks like it isn't actually transferring much data, just checking the disk a lot. If you click on 'open resource monitor' you might be able to figure out more detailed information on which files are being accessed, and by what program.

By the way, which version of the extensions are you using?

[RyaNJ]

Looking at the performance tab doesn't give much information, but it looks like it isn't actually transferring much data, just checking the disk a lot. If you click on 'open resource monitor' you might be able to figure out more detailed information on which files are being accessed, and by what program.

By the way, which version of the extensions are you using?

The latest version of the extensions according to the software.

I've done some other things, disabled pre and super fetching and that seems to have resolved it. I have no idea why, it hasn't ever been a problem before but I have seen similar elsewhere and figured it would be worth a shot.

[Zachs]

Hmm that's really strange indeed... Have you tried MPDN without extensions just to see if the problem still occurs?

[RyaNJ]

Hmm that's really strange indeed... Have you tried MPDN without extensions just to see if the problem still occurs?

Same effect with or without. Though as I have said, turning odd the pre and superfetch have fixed it. I have no idea why this suddenly became a thing... but as long as it's fixed I don't really mind.

I've got no idea what the underlying cause is, it could have been a Windows update that did something to cause it as I hadn't updated MDPN between the time it worked and stopped working.

[Lollerke]

Hi!

I'm using the latest version of MPDN, which is 2.48.0.3603. LAV Filters are at version 0.68.1. Everything works great, but I have on problem. One particular video doesn't get deinterlaced no matter which Deinterlacing mode and method I choose. Other interlaced videos are working fine and this problematic video is working fine with the MPC-HC player. Here's the MediaInfo of the video:

General
ID : 1 (0x1)
Complete name : G:\London.New.Year's.Fireworks.2016.H264.422.1080i.30Mbps-aB.ts
Format : MPEG-TS
File size : 3.52 GiB
Duration : 15mn 58s
Overall bit rate mode : Variable
Overall bit rate : 31.6 Mbps

Video
ID : 308 (0x134)
Menu ID : 1 (0x1)
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High 4:2:2@L4.1
Format settings, CABAC : Yes
Format settings, ReFrames : 4 frames
Codec ID : 27
Duration : 15mn 58s
Bit rate : 29.3 Mbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate : 25.000 fps
Color space : YUV
Chroma subsampling : 4:2:2
Bit depth : 10 bits
Scan type : Interlaced
Scan order : Top Field First
Bits/(Pixel*Frame) : 0.564
Stream size : 3.26 GiB (93%)
Color range : Limited
Color primaries : BT.709
Transfer characteristics : BT.709
Matrix coefficients : BT.709

Audio #1
ID : 256 (0x100)
Menu ID : 1 (0x1)
Format : MPEG Audio
Format version : Version 1
Format profile : Layer 2
Codec ID : 3
Duration : 15mn 57s
Bit rate mode : Constant
Bit rate : 384 Kbps
Channel(s) : 2 channels
Sampling rate : 48.0 KHz
Compression mode : Lossy
Delay relative to video : -370ms
Stream size : 43.9 MiB (1%)
Language : English

Audio #2
ID : 257 (0x101)
Menu ID : 1 (0x1)
Format : MPEG Audio
Format version : Version 1
Format profile : Layer 2
Codec ID : 3
Duration : 15mn 58s
Bit rate mode : Constant
Bit rate : 384 Kbps
Channel(s) : 2 channels
Sampling rate : 48.0 KHz
Compression mode : Lossy
Delay relative to video : -450ms
Stream size : 43.9 MiB (1%)
Language : French

Menu
ID : 32 (0x20)
Menu ID : 1 (0x1)
Duration : 15mn 58s
List : 308 (0x134) (AVC) / 256 (0x100) (MPEG Audio, English) / 257 (0x101) (MPEG Audio, French)
Language : / English / French
Maximum bit rate : 200000000

[Zachs]

MPDN doesn't actually do any deinterlacing. It leaves this to the decoder. Perhaps the LAV filters forum would be the best place to ask?

[nevcairiel]

High-bit-depth content like this movie (ie. its 10-bit) is not supported for software deinterlacing in LAV, its left up to the video renderer to do this, which MPC-HC presumably does.

[edison]

Is there any way to solve this problem now?

[Asmodian]

High-bit-depth content like this movie (ie. its 10-bit) is not supported for software deinterlacing in LAV, its left up to the video renderer to do this, which MPC-HC presumably does.

Interlaced 10-bit! How strange. :p

[huhn]

the 64 bit version of mpdn doesn't run properly on my PC too.

i tried a simply script chain of super XBR -> EWA scaler -> 3D LUT.
and the 3D LUT result were terrible bad in bright scenes.

[Lollerke]

Thanks for the explanation :) Guess I use MPC-HC then.

Interlaced 10-bit! How strange. :p

Not that stange. BBC has 10 bit 4:2:2 1920x1080 as broadcasting requirement.

[nevcairiel]

Not that stange. BBC has 10 bit 4:2:2 1920x1080 as broadcasting requirement.

Its strange for consumers to have that content, because it doesn't usually broadcast to consumers, but is only used internally as a transmission format.

[Shiandow]

the 64 bit version of mpdn doesn't run properly on my PC too.

i tried a simply script chain of super XBR -> EWA scaler -> 3D LUT.
and the 3D LUT result were terrible bad in bright scenes.

It kind of sounds like it's clipping the RGB values, which is kind of expected (probably best to use 3D LUT before scaling, if only because it's faster). The reason this isn't a problem in 32 bit could be because you may have set the image quality in 32 bit to either 'performance' or 'max quality'. Perhaps that option should be renamed (since it only has to do with the type of textures used for intermediate results, and calling something 'max quality' while it really shouldn't be used is often a bad idea).

Anyway I'd recommend using the 3D Lut before upscaling, rather than afterwards.

[Shiandow]

High-bit-depth content like this movie (ie. its 10-bit) is not supported for software deinterlacing in LAV, its left up to the video renderer to do this, which MPC-HC presumably does.

Any plans to add support in the future?

[huhn]

no the player doesn't work at all in 64 bit like other had reported before the extensions can't be loaded.

the results of the 3d LUT are always a disaster no matter where it is applied and the 3D LUT should be used last no matter what it is changing the gamma and changing the outcome of a scaler.

i don't know how the script chain works but a 3D LUT needs limited range input and it is outputting limited range i couldn't find a range script to play around with it to see if that helps.

the quality setting wasn't changed
edit: it is set to "image quality".

[Shiandow]

I see, I thought your problems with the 3D LUT were a description of why the 64 bit version didn't work.

Oddly enough the stand alone x64 version seems to work fine, I'll try to see if the installer has different results.

[huhn]

the stand alone 64 bit version doesn't work for me too.

same error "one ore more extension failed to load!..."

[Shiandow]

Installers also seem to be working fine, odd.

After removing all settings it still works.

[Shiandow]

the stand alone 64 bit version doesn't work for me too.

same error "one ore more extension failed to load!..."

And you're using the latest extension installer?

Edit: From scratch?

[huhn]

i used the Binaries for the stand alone version and for the installed version i'm using the MPDN-Extensions_v1.29.0_Installer.exe from the first page.

the installed version was ibstalled using clean install or what ever it was called for both extensions and player.

the stand alone version is a new folder under "C:\test mpdn" with the content from "MediaPlayerDotNet_x64_2_48_0_3603.zip".
i guess this is fresh.

[Shiandow]

Okay, this is beginning to get (really) weird. What version of windows are you using?

[huhn]

well you don't gonna like that it is windows 10 14393.

the windows version is going to change in the next weeks.
if you think this is an windows issue we can just wait until i find the time to do that.
this is going to be a new install from scratch but other people reported issues like this too and if i remember this correctly it was on windows 10 too.

[Shiandow]

I really hope it isn't some weird difference between new and old versions of windows, but it's the only difference I've found so far. Not sure why such a problem would show up now though, you'd think older versions of the extensions would have the same problem, unless it depends on some windows update (please don't).

We might be able to narrow it down some more if you try older versions of the extensions, and perhaps try putting the uncompiled .cs files from the git repository in the extension folder. Can't guarantee that either will lead to a solution though.

[realghostvids]

I just started using MPDN and I've got a couple of questions.

1. Where can I find a list of hotkeys and can I edit them?
2. I have a GTX 1080 and a 1440p monitor. What settings/render chain stuff should I set to make my videos look their best? Even if it's all up to personal preference I'd appreciate some advice because I'm really not too knowledgeable when it comes to this stuff.

Thanks!

[huhn]

I really hope it isn't some weird difference between new and old versions of windows, but it's the only difference I've found so far. Not sure why such a problem would show up now though, you'd think older versions of the extensions would have the same problem, unless it depends on some windows update (please don't).

We might be able to narrow it down some more if you try older versions of the extensions, and perhaps try putting the uncompiled .cs files from the git repository in the extension folder. Can't guarantee that either will lead to a solution though.

ok i will do some of these test.

is it possible to remove some extensions from the script folder so they are not loaded and the needed dependency is not loaded ?

[Shiandow]

ok i will do some of these test.

is it possible to remove some extensions from the script folder so they are not loaded and the needed dependency is not loaded ?

If you download the individual .cs files you can remove any extension you want. I wouldn't recommend removing the ones in the 'Framework' folder though. And in some cases extensions do depend on others (e.g. SuperRes loads some other scripts as prescalers).

[Shiandow]

I just started using MPDN and I've got a couple of questions.

1. Where can I find a list of hotkeys and can I edit them?
2. I have a GTX 1080 and a 1440p monitor. What settings/render chain stuff should I set to make my videos look their best? Even if it's all up to personal preference I'd appreciate some advice because I'm really not too knowledgeable when it comes to this stuff.

Thanks!

To start off I'd recommend filling a script chain with the SSIM downscaling script, followed by a debanding script and SSSR. You could also try out the new 'krig' mode of the Bilateral chroma scaler if you're feeling adventurous. After that you can start fiddling with the settings, changing SSSR's prescaler etc.

[huhn]

i narrowed it down at least kind of.
the problem should be a renderscript.

i used the stand alone version with a zip from here: https://github.com/Shiandow/MPDN_Extensions

with the extension of this the player crashes as usual but i started to remove some extension.

by removing the renderscript the player starts (playback is not possible). it says the the extension is out dated and wants me to update.

the about pages says now the extension version is 0.0.0.0

[Shiandow]

The version information not working is normal I think, so basically you can get it to work by removing the renderscript folder?

[huhn]

yes the player does say i couldn't load an extension.
playback doesn't work nothing happens i wouldn't say it is working.

i will try older version next but i need sometime for that.

[Shiandow]

Ah, okay, so it's only partially working. Which suggests something fails in the framework that the renderscripts need, taking down everything when it tries to load those. Unfortunately that's still pretty vague.

[realghostvids]

To start off I'd recommend filling a script chain with the SSIM downscaling script, followed by a debanding script and SSSR. You could also try out the new 'krig' mode of the Bilateral chroma scaler if you're feeling adventurous. After that you can start fiddling with the settings, changing SSSR's prescaler etc.

Thanks, I immediately noticed a difference in sharpness once I enabled those scripts. I also saw some weird stuff with the debanding script and certain colours, here's an example.

Deband off
http://i.imgur.com/hVuHe2b.jpg

Deband on
http://i.imgur.com/nRMGBYq.jpg

Adjusting the deband settings don't seem to make any difference so I've just turned it off for now.

[Zachs]

Ah, okay, so it's only partially working. Which suggests something fails in the framework that the renderscripts need, taking down everything when it tries to load those. Unfortunately that's still pretty vague.

I think it's the JavaScript engine. It might be depending on something the fast ring version of Windows doesn't have. Strangely enough the JavaScript engine is Microsoft's too!

[Zachs]

Thanks, I immediately noticed a difference in sharpness once I enabled those scripts. I also saw some weird stuff with the debanding script and certain colours, here's an example.

Deband off
http://i.imgur.com/hVuHe2b.jpg

Deband on
http://i.imgur.com/nRMGBYq.jpg

Adjusting the deband settings don't seem to make any difference so I've just turned it off for now.

A short sample would be very useful. Whatt GPU btw?

[realghostvids]

A short sample would be very useful. Whatt GPU btw?

No problem, here's a 30 second clip demonstrating the issue. (https://my.mixtape.moe/puujzn.mp4) Or did you mean a short part of the video I was watching? Because I can get that for you as well.

I'm using a Nvidia GTX 1080 GPU.

[Shiandow]

You can try toggling the 'improve chroma' setting in Video Renderer -> General (I think it can be toggled by pressing Ctrl+Shift+C). If that doesn't do anything a short part of the video you were watching would be helpful.

[realghostvids]

You can try toggling the 'improve chroma' setting in Video Renderer -> General (I think it can be toggled by pressing Ctrl+Shift+C). If that doesn't do anything a short part of the video you were watching would be helpful.

Turning off improve chroma fixed it.

Here's a sample of the video I was watching anyway in case you want to take a look at it. (https://my.mixtape.moe/lgxfgc.mp4) It only happens when the video is scaled to fullscreen, if I'm viewing it in a window at 100% size the issue doesn't show up.

[Shiandow]

Okay, looks like improve chroma needs some fixing, I had some doubts about it for a while now, but now I know it can actually cause issues, thanks. Not sure why that exact combination of settings causes problems, but I fear 'improve chroma' has some numerical issues, so just about anything can happen. It's just incredibly hard to find those edge cases where things go wrong.

[aufkrawall]

I once found a case where it looked comic too, it was hair in front of a red wall.

[huhn]

maybe this help for the crash:


Application: MediaPlayerDotNet.exe Framework Version: v4.0.30319 Description: The process was terminated due to an unhandled exception. Exception Info: System.Runtime.InteropServices.COMException at System.RuntimeTypeHandle.CreateInstance(System.RuntimeType, Boolean, Boolean, Boolean ByRef, System.RuntimeMethodHandleInternal ByRef, Boolean ByRef) at System.RuntimeType.CreateInstanceSlow(Boolean, Boolean, Boolean, System.Threading.StackCrawlMark ByRef) at System.Activator.CreateInstance(System.Type, Boolean) at System.Activator.CreateInstance(System.Type) at Microsoft.ClearScript.Util.MiscHelpers.TryCreateCOMObject[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]](System.String, System.String, System.__Canon ByRef) at Microsoft.ClearScript.Windows.ProcessDebugManagerWrapper64.TryCreate(Microsoft.ClearScript.Windows.ProcessDebugManagerWrapper ByRef) at Microsoft.ClearScript.Windows.WindowsScriptEngine+<>c__DisplayClass2.<.ctor>b__0() at Microsoft.ClearScript.ScriptEngine.ScriptInvoke[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]](System.Func`1<System.__Canon>) at Microsoft.ClearScript.Windows.WindowsScriptEngine..ctor(System.String, System.String, Microsoft.ClearScript.Windows.WindowsScriptEngineFlags) at Microsoft.ClearScript.Windows.JScriptEngine..ctor(System.String, System.String, Microsoft.ClearScript.Windows.WindowsScriptEngineFlags) at Mpdn.Extensions.Framework.Scripting.ScriptEngine`3[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]]..ctor() at Mpdn.Extensions.Framework.Scripting.RenderScriptEngine.get_Instance() at Mpdn.Extensions.RenderScripts.Mpdn.ScriptedRenderChain.ScriptedRenderChain..ctor() Exception Info: System.Reflection.TargetInvocationException at System.RuntimeTypeHandle.CreateInstance(System.RuntimeType, Boolean, Boolean, Boolean ByRef, System.RuntimeMethodHandleInternal ByRef, Boolean ByRef) at System.RuntimeType.CreateInstanceSlow(Boolean, Boolean, Boolean, System.Threading.StackCrawlMark ByRef) at System.Activator.CreateInstance[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]]() at Mpdn.Extensions.Framework.Config.MemConfig`1[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]]..ctor() at Mpdn.Extensions.Framework.Config.ExtensionUi`3[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]].get_Settings() at Mpdn.Extensions.Framework.Chain.ChainUi`4[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]].Dispose(Boolean) at Mpdn.Extensions.Framework.Chain.ChainUi`4[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]].Finalize()

MediaPlayerDotNet.exe
2.48.0.3603
581d8fb5
KERNELBASE.dll
10.0.14393.479
582588e6
e0434352
0000000000017788
684
01d25d11e0608cf2
C:\Program Files\MPDN\MediaPlayerDotNet.exe
C:\WINDOWS\System32\KERNELBASE.dll
0f3bee0f-b2c8-4f6a-86d1-24d13d08472a

very wired is that the report is saying C:\Program Files\MPDN\MediaPlayerDotNet.exe
the used stand alone exe is at C:\test mpdn\MediaPlayerDotNet.exe

[Shiandow]

Okay, looks like Zach's suspicion is correct and it's failing to load the JS engine. Hard to say exactly why. In theory it should then work if you remove 'Scripting.cs' from the framework and all renderscripts that depend on it, but that removes a lot of functionality (you'd lose the Conditional scripts, for one).

This would have been a lot easier to diagnose if MPDN actually managed to say which renderscript failed to load. That's a feature that should probably be considered 'high priority' for the next version of MPDN.

Also a bad sign is that it's failing to load a script's settings while destroying the script, that probably shouldn't happen. So, there's a chance this is only a 'secondary' error, triggered by some other.

Edit: can't even figure out why ChainUIs had to have their own destructor anyway, either way it's not implemented well so it's probably better to just remove it.

[Zachs]

From memory, ChainUI used it need it because it was the owner of textures. Can you check if that's no longer the case? Textures don't automatically add memory pressure to GC. In fact it's GPU memory which the GC doesn't know anything about.

[Zachs]

Btw we really should report the bug to Microsoft. Either way it's Microsoft's problem. Clearscript and windows, both are Microsoft.

[Shiandow]

From memory, ChainUI used it need it because it was the owner of textures. Can you check if that's no longer the case? Textures don't automatically add memory pressure to GC. In fact it's GPU memory which the GC doesn't know anything about.

I think that used to be the case (although considering that pretty much nothing ever disposed ChainUIs, I'm surprised it didn't cause issues). Anyway IIRC we've now moved it to the Filters themselves. There's still a small possibility that someone messes up and removes all references to a Filter that manages some resource, without disposing the Filter, but I don't see how someone could remove all references to such a Filter without noticing the resource isn't used in the final result.

I suppose it's a bit tricky to check if this can't ever happen, since I haven't been too careful with references to filters that I know don't carry any additional resources. But you'd have to do something really weird to completely remove all references to a filter, you'd either have to completely ignore the texture in the Chain that allocates it, or completely ignore the input in a certain RenderChain (which I suppose might make sense for a script that generates content, but no such script exists yet).

[Hrxn]

To start off I'd recommend filling a script chain with the SSIM downscaling script, followed by a debanding script and SSSR. You could also try out the new 'krig' mode of the Bilateral chroma scaler if you're feeling adventurous. After that you can start fiddling with the settings, changing SSSR's prescaler etc.

SSIM downscaling a good start? That is new to me, what is the benefit in using this with, for example, 720p material?

And I've tried Bilateral chroma with 'Krig' now for a while, no issues so far, seems to work fine!

[Shiandow]

SSIM downscaling a good start? That is new to me, what is the benefit in using this with, for example, 720p material?

And I've tried Bilateral chroma with 'Krig' now for a while, no issues so far, seems to work fine!

Adding SSIM downscaling only has an effect if you're actually downscaling, which doesn't often happen with 720p material. If you have to downscale it's useful to do so first (or just after the chroma scaling) since it means the rest of the processing can be done on a smaller image.

[Pat357]

I installed MPDN v2.48 and extensions v1.29 (all latest).

I have a serious problem as soon as I installed the extensions.
Upon opening, I get a message like :

"One or more extensions failed to load!
The extensions you have installed may be incompatible with the current version of MPDN.
You are strongly advised to correct this problem before using MPDN."

If I press OK and try to do something (playing a media file), the I get an error and the program seems to crash.

If I activate debugging and restart I get the following error :

An unexpected error 'System.Reflection.TargetInvocationException'has occurred.
Additional information:
Exception has been thrown by the target of an invocation. (mscorlib)
Either a required impersonation level was not provided, or the provided impersonation level is invalid.
(Exception from HRESULT: 0x80070542) (ClearScript)

Any ideas how to get this working ?
Is there a requirement related to CUDA ? (min CUDA version ?)
Without the extensions, I could play any media with DirectX 3D v9, v10.1 and v12. All worked perfect

Some info from windows log :
Log Name: Application
Source: Application Error
Date: 27/12/2016 19:53:15
Event ID: 1000
Task Category: (100)
Level: Error
Keywords: Classic
User: N/A
Computer: Win7
Description:
Faulting application name: MediaPlayerDotNet.exe, version: 2.48.0.3603, time stamp: 0x581d8f93
Faulting module name: KERNELBASE.dll, version: 6.1.7601.23572, time stamp: 0x57fd0379
Exception code: 0xe0434352
Fault offset: 0x0000c54f
Faulting process id: 0x15bc
Faulting application start time: 0x01d260727222203f
Faulting application path: C:\Program Files (x86)\MPDN\MediaPlayerDotNet.exe
Faulting module path: C:\Windows\syswow64\KERNELBASE.dll
Report Id: b8ed151c-cc65-11e6-a215-5cd998a744f1
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
<System>
<Provider Name="Application Error" />
<EventID Qualifiers="0">1000</EventID>
<Level>2</Level>
<Task>100</Task>
<Keywords>0x80000000000000</Keywords>
<TimeCreated SystemTime="2016-12-27T18:53:15.000000000Z" />
<EventRecordID>307367</EventRecordID>
<Channel>Application</Channel>
<Computer>Win7</Computer>
<Security />
</System>
<EventData>
<Data>MediaPlayerDotNet.exe</Data>
<Data>2.48.0.3603</Data>
<Data>581d8f93</Data>
<Data>KERNELBASE.dll</Data>
<Data>6.1.7601.23572</Data>
<Data>57fd0379</Data>
<Data>e0434352</Data>
<Data>0000c54f</Data>
<Data>15bc</Data>
<Data>01d260727222203f</Data>
<Data>C:\Program Files (x86)\MPDN\MediaPlayerDotNet.exe</Data>
<Data>C:\Windows\syswow64\KERNELBASE.dll</Data>
<Data>b8ed151c-cc65-11e6-a215-5cd998a744f1</Data>
</EventData>
</Event>

Application: MediaPlayerDotNet.exe
Framework Version: v4.0.30319
Description: The process was terminated due to an unhandled exception.
Exception Info: System.Runtime.InteropServices.COMException
at Microsoft.ClearScript.Windows.IProcessDebugManager32.AddApplication(Microsoft.ClearScript.Windows.IDebugApplication32, UInt32 ByRef)
at Microsoft.ClearScript.Windows.ProcessDebugManagerWrapper32.AddApplication(Microsoft.ClearScript.Windows.DebugApplicationWrapper, UInt32 ByRef)
at Microsoft.ClearScript.Windows.WindowsScriptEngine+<>c__DisplayClass2.<.ctor>b__0()
at Microsoft.ClearScript.ScriptEngine.ScriptInvoke[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]](System.Func`1<System.__Canon>)
at Microsoft.ClearScript.Windows.WindowsScriptEngine..ctor(System.String, System.String, Microsoft.ClearScript.Windows.WindowsScriptEngineFlags)
at Microsoft.ClearScript.Windows.JScriptEngine..ctor(System.String, System.String, Microsoft.ClearScript.Windows.WindowsScriptEngineFlags)
at Microsoft.ClearScript.Windows.JScriptEngine..ctor(System.String, Microsoft.ClearScript.Windows.WindowsScriptEngineFlags)
at Mpdn.Extensions.Framework.Scripting.ScriptEngine`3[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]]..ctor()
at Mpdn.Extensions.Framework.Scripting.RenderScriptEngine.get_Instance()
at Mpdn.Extensions.RenderScripts.Mpdn.ScriptedRenderChain.ScriptedRenderChain..ctor()

Exception Info: System.Reflection.TargetInvocationException
at System.RuntimeTypeHandle.CreateInstance(System.RuntimeType, Boolean, Boolean, Boolean ByRef, System.RuntimeMethodHandleInternal ByRef, Boolean ByRef)
at System.RuntimeType.CreateInstanceSlow(Boolean, Boolean, Boolean, System.Threading.StackCrawlMark ByRef)
at System.RuntimeType.CreateInstanceDefaultCtor(Boolean, Boolean, Boolean, System.Threading.StackCrawlMark ByRef)
at System.Activator.CreateInstance[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]]()
at Mpdn.Extensions.Framework.Config.MemConfig`1[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]]..ctor()
at Mpdn.Extensions.Framework.Config.ExtensionUi`3[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]].get_Settings()
at Mpdn.Extensions.Framework.Chain.ChainUi`4[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]].Dispose(Boolean)
at Mpdn.Extensions.Framework.Chain.ChainUi`4[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089],[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]].Finalize()

---
MPDN 2.48 x86
MPDN extensions v1.29 x86
Windows 7 64bit SP1 all updates
Intel CPU only Opencl_runtime v16.1.1_x64
GPU : GTX-680 with drivers 347

[Zachs]

Bloody Microsoft Clearscript failing to load again!

Edit: I'll have a look later to see if they have an updated version.

[Zachs]

I just pushed an update to workaround the problem. Could those who've encountered the COMException with ClearScript give it a try? Grab the source from GitHub - I want to make sure it fixes the problem before doing a proper release.

[huhn]

it complains about the extension version but works with some settings changed.

damned are these AMD driver.
it crashes with a back buffer bigger than 7 with the current AMD driver.

and for some reasons subtitles are aliased with and without reduce xy subfilter subtitles aliasing.
is xy subfilter nad xy vsfilter hard code or why is there no proper way to use assfilter mod?

[Pat357]

Thanks a lot Zachs !
No more crashes or complaining !
On starting it says something like "optimizing extensions" and nothing more.
Seems to work nicely now, although I didn't test it all yet.

Thanks again !

Edit:
The 64-bit version probably can't work with the Master-extensions (libs are 32bit), can it ?

[Zachs]

The libs can work with both 32 and 64 bit mpdn. Did you get an error?

[Pat357]

The libs can work with both 32 and 64 bit mpdn. Did you get an error?

Indeed, for the 64bit I get an error saying that :
"One or more extensions could not be loaded...
The extensions you currently have installed may be incompatible with this MPDN (or something along that..)
It is advised to solve this first before using MPDN..."

Strange that it only happens with 64bit player...

This is what I get after activating debug in MPDN :

An unexpected error 'csscript.CompilerException' has occurred.
Additional information:
c:\Program Files\MPDN\Extensions\Properties\AssemblyInfo.cs(9,12): error CS0579: Duplicate 'AssemblyDescription' attribute
(CSScriptLibrary)


Again, without the extensions, the 64 bit player is also ok, but I had to manually copy my installed LAV-filters to the "filter" sub-directory.
Without this, the player failed to play any file.

Now the extensions for 64bit and then we have everything we could have dreamed of ! :D

[Shiandow]

This is what I get after activating debug in MPDN :

An unexpected error 'csscript.CompilerException' has occurred.
Additional information:
c:\Program Files\MPDN\Extensions\Properties\AssemblyInfo.cs(9,12): error CS0579: Duplicate 'AssemblyDescription' attribute
(CSScriptLibrary)


That's not the error I expected. Looks like you've got mutliple copies of the AssemblyInfo.cs file, which isn't even necessary as far as I know. You should probably only use the 'Extensions' folder from the git repo. Replace the MPDN\Extensions folder with the Extensions folder from the repository. Just to be safe you should probably also remove any 'Properties' folders you find.

[Zachs]

Try removing the extensions folder and start again. Make sure it's exactly the same as your 32-bit one.

[Pat357]

You should probably only use the 'Extensions' folder from the git repo. Replace the MPDN\Extensions folder with the Extensions folder from the repository. Just to be safe you should probably also remove any 'Properties' folders you find.

This worked like a charm, no more errors or other things !! Thanks.

With the 32-bit I'm quit sure I have extracted the whole "MPDN_Extensions-master.zip" inside the extensions folder and everything was OK.

I think it should be safe now to make a new release.

[Shiandow]

With the 32-bit I'm quit sure I have extracted the whole "MPDN_Extensions-master.zip" inside the extensions folder and everything was OK.

It might work, but there's a lot of irrelevant stuff in there. Also, I think you somehow ended up with several copies of some of the files.

[Pat357]

It might work, but there's a lot of irrelevant stuff in there. Also, I think you somehow ended up with several copies of some of the files.

Yeah, there maybe more snakes hiding in this long grass.

BTW : How do you detect the location from pre-installed LAV filters ?
Do you use the LAV registry key for this ?
Same for XYSubFilter ?

[Zachs]

It's via COM registration.

[realghostvids]

Hello again, I've got a couple more questions about MPDN.

Is there a way for the taskbar icon to always be more detailed? If I open a video in a window the taskbar icon looks like this:
http://i.imgur.com/VdkKseW.jpg

But as soon as I alt+enter and make it fullscreen it's just plain:
http://i.imgur.com/Br3jwpw.jpg

Also, are there keys for jumping backwards and forwards by just one frame?

[Shiandow]

Also, are there keys for jumping backwards and forwards by just one frame?

Ctrl+Left and Ctrl+Right should work.

Not sure about the taskbar icon thing, seems like windows is being difficult because technically MPDN doesn't have a windows when it's in full screen. Not sure if it's worth the trouble to fix it.

[mlauzon]

Is there anyway to change the behaviour of MPDN?

In most players double clicking makes a player full screen or back to normal size if it's already in full screen.

Clicking on MPDN pauses the video, for me, I think that is bad, it's why pause is typically done when pressing the space bar.

[huhn]

no the player doesn't work at all in 64 bit like other had reported before the extensions can't be loaded.

the results of the 3d LUT are always a disaster no matter where it is applied and the 3D LUT should be used last no matter what it is changing the gamma and changing the outcome of a scaler.

i don't know how the script chain works but a 3D LUT needs limited range input and it is outputting limited range i couldn't find a range script to play around with it to see if that helps.

the quality setting wasn't changed
edit: it is set to "image quality".

anything about the incorrect handling of 3D LUT or do i need a different 3D LUT?

[Shiandow]

The way the code currently works you'll need a LUT from (PC range) RGB to (whatever range your screen expects) RGB. Would it make more sense if it used a different colour space as input?

[huhn]

the "madVR" 3d lut from DisplayCAL are 16-235 input and 16-235 output.

so the input has to be limited range always and the output depends on the setting used while calibration which is usually full range. or simply the option selected in the player.

if the video range option is "full" under "video renderer -> composition" it needs to be expended to full range again.
a waste but that's how these 3D LUT work.

[madshi]

In madVR the 3dluts are always TV levels input and output, so the user doesn't have to redo the 3dlut if he switches between different input/output levels. Also, using TV levels means the 3dlut has control over BTB and WTW in case the calibration software knows some secret sauce to process BTB/WTW, too. If the 3dlut is PC levels in either input or output, it's hard to say what should happen with BTB/WTW content, when applying the 3dlut.

[EncodedMango]

I can't seem to find Back/Next buttons on the player, what am I missing?

[mrcorbo]

I can't seem to find Back/Next buttons on the player, what am I missing?

I don't believe there are buttons on the UI. The keyboard shortcuts are listed under Play->Navigation in the player window's menu or if you have a mouse with forward/back buttons for web navigation they work as as back/next as well.

[EncodedMango]

I don't believe there are buttons on the UI. The keyboard shortcuts are listed under Play->Navigation in the player window's menu or if you have a mouse with forward/back buttons for web navigation they work as as back/next as well.
It would be nice if the UI had those buttons or an option to add them if it takes away from UI idea but this works too, thanks.

[Zachs]

New extensions coming up soon...

Edit: released.

[Zachs]

It would be nice if the UI had those buttons or an option to add them if it takes away from UI idea but this works too, thanks.

Back and next buttons are playlist related buttons so you'll find them on the playlist window itself.

[Shiandow]

Hi all,

The new extensions v1.30.0 seem to be up now. These include some small tweaks to the Krig mode Bilateral script, and to SSSR which seems to have reached a somewhat more mature state, and can probably replace SuperRes in nearly all cases.

The new extensions also include some pretty large changes to the 'Deband experimental' renderscript (when I get round to it I'll give it a proper name). It can now handle some pretty difficult situations (e.g. banding + grain) really well, leaving intact small details while removing the large ugly bands. If you want it to remove details you can also untick 'preserve details' in which case it won't try to preserve smaller details, but use this with care. You can also raise the strength, in which case it will be more aggressive in removing bands (a word of warning: strength = 1 removes everything).

For those interested in knowing how it works: (the following may contain some technical terms) basically it tries to detect really low intensity edges and remove them. If you turn on 'preserve detail' it will also try to preserve local minima and maxima. Meaning that it, for example, won't lower the brightness of a pixel if it's brighter than it's surroundings. The method also uses a so called 'image pyramid' to look at the images across different scales (this is what allows it to blur large bands while keeping small details). Basically it creates a sequence of downscaled images and then starting from the end of the sequence it reconstructs the original image step by step, using the sequence of downscaled images to recover lost detail while throwing away the details it doesn't want.

Personally I think this results in what could be the best debanding method yet, I've yet to see any other method where you don't have to choose between removing bands and throwing away detail. I'm really interested to hear what you all think of it.

[sauma144]

The SSSR algorithm is really awesome and is really fast on my laptop. Thanks!
I am very happy to see some activities in this thread especially since the announcement of the mpc-hc developers.

[Heaud]

First time trying out SSSR and I must say, it looks great with Super-xBR set as a prescaler. There is no perceivable aliasing with my set up while using SSSR and Super-xBR.

Have yet to try out the experimental deband setting. Are the strength values identical to the regular deband?

[EncodedMango]

After updating to Extensions v1.30 the View-->OpenSubtitles menu doesn't work, it just pauses the video and seem to try to load the dialog but nothing happens.

EDIT: After doing that a few times, it's working fine again. My guess is maybe it was some issue with domain resolution.

[Shiandow]

Have yet to try out the experimental deband setting. Are the strength values identical to the regular deband?

Well to some extent, strength = 0.5 is somewhat conservative but high enough for most cases, and then the strength rises rapidly as you get closer to strength = 1. I'm actually thinking of changing this behaviour a bit, since strength = 1 isn't really a sensible setting, but I'd like some feedback on which values make sense for bad sources.

[Heaud]

I set deband strength to 0.2-0.3 with random dithering (1.6 noise strength) for clean anime videos as it removes enough bands without smearing subtle details.

After testing about for 10 minutes using ordered dithering, the default settings does look like a visual balance for heavily banded sources. The highest I would recommend for a strongest setting is 0.8 as anything higher destroys way too much detail even with the Preserve Detail option enabled.

[Shiandow]

Thanks for the feedback. I agree that raising it above 0.8 is probably excessive, when you do that you're essentially destroying all information in the lowest 3 bits, which is quite a lot.

Does 0.2-0.3 still remove banding sufficiently well? It seems lower than I expected.

[Hrxn]

[..] for clean anime videos [..]

Here's the thing. In my experience, settings that work well for anime/animated film are often useless for live/real film, and vice versa of course.
One should keep this in my mind when trying different settings and properties..

[Heaud]

Does 0.2-0.3 still remove banding sufficiently well? It seems lower than I expected.

0.3 is quite soft at debanding, I would consider it as a "lite" or "low strength" setting. 0.2 would be for the cleanest of anime blu-ray sources.



Here's the thing. In my experience, settings that work well for anime/animated film are often useless for live/real film, and vice versa of course.
One should keep this in my mind when trying different settings and properties..
I agree with this statement. The strength that would work with anime/cgi/cartoons does not work the same with live action.

[Shiandow]

Here's the thing. In my experience, settings that work well for anime/animated film are often useless for live/real film, and vice versa of course.
One should keep this in my mind when trying different settings and properties..

True, but the new debanding method seems to work well across a broad range of images. If someone has counterexamples they are most welcome.

I suppose it does end up softening the image in some cases, but I actually think this is necessary if you want to recover what the image looked like originally. In fact I'd even be interested if someone could find a case where the defaults soften the image unnecessarily (although I reserve the right to point out banding not visible on most screens / to the human eye).

[Garteal]

The new extensions also include some pretty large changes to the 'Deband experimental' renderscript (when I get round to it I'll give it a proper name). It can now handle some pretty difficult situations (e.g. banding + grain) really well, leaving intact small details while removing the large ugly bands. If you want it to remove details you can also untick 'preserve details' in which case it won't try to preserve smaller details, but use this with care. You can also raise the strength, in which case it will be more aggressive in removing bands (a word of warning: strength = 1 removes everything).

Been testing it and so far I like the old deband method a smidge better for my use-cases and from what I've seen so far. As for the range of the experimental debanding, can you make it more of a linear change rather than an exponential one?
Also Strength 1 is kinda useless, perhaps it'd make more sense to clamp it to a more usable range?

I've also noticed that increasing the value for the experimental debanding also affects the image colors. Not sure if it's intended or not reading your technical description.

Base: https://abload.de/img/basepoz2r.png
Deband old: https://abload.de/img/deband_oldxqblz.png
Deband new: https://abload.de/img/deband_newr8x3n.png
Deband new 0.7str: https://abload.de/img/deband_new_0.7strohlos.png
(Note: These are 2-3mb pngs!)

[Shiandow]

Thanks for the detailed analysis. I think changing the way strength works is probably a good idea, I guess most people wouldn't expect 'strength=1' to that extreme.

I think I have a good idea of what's happening in the images you sent, but I'd like to confirm a couple of things. First of all, you're not using the defaults for the old debanding method are you? Also did you disable 'preserve detail' when setting the strength to 0.7?

If possible it would also help if I could get a sample of the source.

[Heaud]

I decided to give another look at the debanding filter but this time comparing the original with both available filters. This is tested with Preserve Detail set to On and Ordered Dithering.

Classic Deband:
Strength 0.3 http://screenshotcomparison.com/comparison/203099
Strength 0.5 http://screenshotcomparison.com/comparison/203100
Strength 0.7 http://screenshotcomparison.com/comparison/203101

Experimental Deband:
Strength 0.3 http://screenshotcomparison.com/comparison/203103
Strength 0.5 http://screenshotcomparison.com/comparison/203104
Strength 0.7 http://screenshotcomparison.com/comparison/203105

And an extra comparison to show what Preserve Detail is achieving with strength set to 0.7:
Classic Deband http://screenshotcomparison.com/comparison/203107
Experimental Deband http://screenshotcomparison.com/comparison/203108

[Garteal]

^ Nice going. I can see slight blurring happening but I would take that personally over banding. 0.4 or 0.5 is probably my preference with preserve details checked.


First of all, you're not using the defaults for the old debanding method are you?

Oh sorry. I was using the default settings (0.5str & preserve details checked) where I didn't mention any specific settings.

Also did you disable 'preserve detail' when setting the strength to 0.7?
Just changed the strength. Disabling preserve details produces adverse effects similar to what Heauds image (Experimental Deband http://screenshotcomparison.com/comparison/203108) shows.

I'll cut you a sample and will edit this post with a link later today.

Edit: Grab the sample here (http://www109.zippyshare.com/v/j0oeSVB1/file.html).

[BetA13]

hey there,

i dont really know whats wrong here, i tried reinstalling several times now but still nothing changes..

My problem is that, when upgrading to the new version of render scripts, i now, cant use the "RenderScript Presets " anymore..they are in the list but when i go into detail and lets say change teh settings for "Native" (1080p) or upscaling etc, the settings ill enter there dont stick..There wherent even settings there in teh first place, they are just gone..They worked like charm before, i could change and modify settings, now they arent even there..

some screens:

This is how it looks now, It doesnt save anything i put in the BLANK field like its supposed to be:

https://i.imgsafe.org/5f21f0250a.png
https://i.imgsafe.org/5f21fcb40c.png


an here is how it was before when it worked fine, you can see theres settings in there where the "Blank" space was in the other pics:

https://i.imgsafe.org/5f220df5fb.png
https://i.imgsafe.org/5f221bb00f.png


im really confused here..It was working really good before i could change and make my own settings based on aufkrawalls render scrip preset, but now its broken somehow..

-i tried reinstalling several times.
-I deleted the Settings folder in appdata/Local
-i deleted everything in teh regestry related to mpdn
- tried portable VS installable

nothing worked, for me it looks like the render script chain is broken..

Please halp :)
Edit, sorry for big pics, cant use spoiler here :(

[Shiandow]

Thanks for the bug report, I've verified and fixed the problem. The good news is that it is actually saving your configuration, but it has trouble loading it into the setting dialog, so you if you try to modify it you have to start from scratch. Hopefully it'll be fixed soon, in the meantime:

Don't change your Conditional scripts' settings

You can actually probably change the condition, but you might want to backup your settings just in case. At any rate, if you press cancel then no changes will be made, no matter what it is displaying.

[BetA13]

great news...
And here i thought it was me, haha..

damnit, i allready deleted my old script settings, when hunting for the error..
i guess ill whait till its fixed and i can edit the settings again..

Thanks for the verification :)

[Zachs]

Shiandow, mate is the fix you did ready to be released?

[Shiandow]

Yeah, it should probably work. The one thing I'm slightly concerned about is that I thought I'd tested the previous method, and I can't figure out why that ever worked.

[Garteal]

Been playing a lil again with some ideas for a PlayerExtension for what I needed, but perhaps it's also useful for others. And maybe it can be expanded with more functionality.

Nothing special, just a simple OSD. But I found myself often looking for what episode I was playing which required me to close down fullscreen to look at the playlist.
That and a quick close button was also added for convenience and those that only have a pointer device handy. I was thinking of adding prev&next buttons for playlist control aswell.

http://i.imgur.com/ZX4eZHt.jpg

[Zachs]

Nice one! That's been requested a number of times, the quick close button in full screen mode.

[Zachs]

Btw, would be nice if you could find a way to make the title appear for a few seconds only when mouse was moved.

[Garteal]

It already does that, it'd be quite an annoyance if it didn't haha.
I'll probably move this to its own extension and have to move some logic before I can push it.

Also another idea I have for the Playlist (atleast interesting for me) is to maybe add an option to try and automatically skip openings and/or endings. Only usable for properly chaptered anime (or other media with proper tagging).
I don't really need to see them every time when I'm watching multiple episodes in one sitting. =p

It's so nice to be able to easily add functionality to the player. Small things like this make a big change for me.

[toniash]

Is it possible to move automatically the image to the top of the screen? (as in PotPlayer or Madvr)

[trandoanhung1991]

MPDN is crashing in Full Screen (exclusive or not) every time if I have DX10/11 rendering and new windowed rendering path enabled.
Running Windows 10, GTX 1070 with latest NVIDIA drivers (378.92). DSR and Share both are disabled.
Reinstalled MPDN and extensions, NVIDIA drivers, doesn't help.

So right now for MPDN to work in FS I have to either disable new windowed path or run DX9.

[Zachs]

Is it crashing only with the latest drivers?

[trandoanhung1991]

Is it crashing only with the latest drivers?

It's been crashing for the last few drivers. Using 381.65 now and it's still showing the same behavior.

[turab]

I'm having an issue playing video files since today, and it seems to be specific to the 64-bit version. Here's the log:

===================================

An unexpected error 'System.ArgumentOutOfRangeException' has occurred.

------------------------------
Error Type = System.ArgumentOutOfRangeException
Error Message = Specified argument was out of the range of valid values.
Error Source = MediaPlayerDotNet
Error Site = Mpdn.IScaler C(Mpdn.IScaler)
Error occurred = at a.C.C(IScaler )
at a.C.A(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.o(Object , EventArgs )
at System.Windows.Forms.ToolStripItem.RaiseEvent(Object key, EventArgs e)
at System.Windows.Forms.ToolStripMenuItem.OnClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleMouseUp(MouseEventArgs e)
at System.Windows.Forms.ToolStrip.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.ToolStripDropDown.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.Control.WmMouseUp(Message& m, MouseButtons button, Int32 clicks)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.ToolStrip.WndProc(Message& m)
at System.Windows.Forms.ToolStripDropDown.WndProc(Message& m)
===================================

An unexpected error 'System.ArgumentOutOfRangeException' has occurred.

------------------------------
Error Type = System.ArgumentOutOfRangeException
Error Message = Specified argument was out of the range of valid values.
Error Source = MediaPlayerDotNet
Error Site = Mpdn.IScaler C(Mpdn.IScaler)
Error occurred = at a.C.C(IScaler )
at a.C.A(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.o(Object , EventArgs )
at System.Windows.Forms.ToolStripItem.RaiseEvent(Object key, EventArgs e)
at System.Windows.Forms.ToolStripMenuItem.OnClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleMouseUp(MouseEventArgs e)
at System.Windows.Forms.ToolStrip.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.ToolStripDropDown.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.Control.WmMouseUp(Message& m, MouseButtons button, Int32 clicks)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.ToolStrip.WndProc(Message& m)
at System.Windows.Forms.ToolStripDropDown.WndProc(Message& m)
at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)

===================================

Specified argument was out of the range of valid values. (MediaPlayerDotNet)

------------------------------
Program Location:

at a.C.C(IScaler )
at a.C.A(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.o(Object , EventArgs )
at System.Windows.Forms.ToolStripItem.RaiseEvent(Object key, EventArgs e)
at System.Windows.Forms.ToolStripMenuItem.OnClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleMouseUp(MouseEventArgs e)
at System.Windows.Forms.ToolStrip.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.ToolStripDropDown.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.Control.WmMouseUp(Message& m, MouseButtons button, Int32 clicks)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.ToolStrip.WndProc(Message& m)
at System.Windows.Forms.ToolStripDropDown.WndProc(Message& m)
at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)

at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)

===================================

Specified argument was out of the range of valid values. (MediaPlayerDotNet)

------------------------------
Program Location:

at a.C.C(IScaler )
at a.C.A(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.o(Object , EventArgs )
at System.Windows.Forms.ToolStripItem.RaiseEvent(Object key, EventArgs e)
at System.Windows.Forms.ToolStripMenuItem.OnClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleMouseUp(MouseEventArgs e)
at System.Windows.Forms.ToolStrip.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.ToolStripDropDown.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.Control.WmMouseUp(Message& m, MouseButtons button, Int32 clicks)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.ToolStrip.WndProc(Message& m)
at System.Windows.Forms.ToolStripDropDown.WndProc(Message& m)
at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)


OS Version = Microsoft Windows NT 6.3.9600.0
Total Physical Memory = 8 GB
Available Physical Memory = 3.9 GB
Ran As Admin = False
HasShutdownStarted = False
ProcessorCount = 2
RuntimeVersion = 4.0.30319.42000
Thread Count = 58
Handle Count = 1128
VM Size = 21.4 GB
Peak VM Size = 21.5 GB
Working Set Size = 275.5 MB
Max Working Set Size = 1.3 MB
Min Working Set Size = 200 KB

Despite what the log says, I'm using Windows 10. There was a big update recently (Creators Update) that was installed on my system a couple of days ago. It didn't affect video playback yesterday, but it might somehow be related. I tried a clean install of MPDN, but the problem persists. The 32-bit version works fine, however.

[Shiandow]

If the 32 bit version is working fine then it might be an issue with the settings. Could you try deleting MPDN's settings (should be at %LOCALAPPDATA%\MediaPlayerDotNet)? Make a copy somewhere if you want to save any settings. This is just to test if that is indeed the problem.

Also ignore whatever version number windows claims it is, I don't think those have ever made any sense.

[turab]

If the 32 bit version is working fine then it might be an issue with the settings. Could you try deleting MPDN's settings (should be at %LOCALAPPDATA%\MediaPlayerDotNet)? Make a copy somewhere if you want to save any settings. This is just to test if that is indeed the problem.

Also ignore whatever version number windows claims it is, I don't think those have ever made any sense.
Thanks, that did the trick. I guess it was to do with MPDN extensions.

[Shiandow]

Thanks, that did the trick. I guess it was to do with MPDN extensions.

Not quite sure what was wrong but at least it works now. Any clue what triggered it? Did you update MPDN or the extensions recently?

[XMonarchY]

I wonder if this Player is a good choice for me. I am currently with good old MPC-HC x86 + LAV Filters x86 + ReClock x86 + madVR x86 and its all x86 32bit because ReClock does not support x64.

Does this player have integrated ReClock-like feature that uses x64??? I know JRiver Media Player has its own thing like ReClock and it runs in 64bit mode, but its not as snappy as MPC-HC.

If this player indeed can do 64bit ReClock-like timing, then I may very much switch to it from MPC-HC, provided I can disable most of its internal filters and renderers (aside from ReClock-like feature) to be purely powered by stand-alone LAV Filters and madVR!

[ryrynz]

@XMonarchY This player has it's own renderer and it cannot be changed.

[turab]

Not quite sure what was wrong but at least it works now. Any clue what triggered it? Did you update MPDN or the extensions recently?
The last time MPDN extensions was updated was at least a few weeks ago, and I'm pretty sure I didn't touch anything. Between the last time the player worked and the first time I noticed the error, my computer did crash. Although I wasn't using the player when it happened, maybe it had an inadvertent effect.

[Shiandow]

The last time MPDN extensions was updated was at least a few weeks ago, and I'm pretty sure I didn't touch anything. Between the last time the player worked and the first time I noticed the error, my computer did crash. Although I wasn't using the player when it happened, maybe it had an inadvertent effect.

Odd. Seems like something messed up the setting during that crash, but I can't be sure. It's going to be hard to protect against that though.

[XMonarchY]

@XMonarchY This player has it's own renderer and it cannot be changed.

Shoot... Is there a raw player like MPC HC that has 64bit ReClock-like feature and fully supports external LAV Filters and madVR?

[ryrynz]

AFAIK MPDN is the only player to have implemented this feature.

[foozoor]

AFAIK MPDN is the only player to have implemented this feature.
We are in 2017 and people are still ignoring mpv. :rolleyes:

[Olivier C.]

Odd. Seems like something messed up the setting during that crash, but I can't be sure. It's going to be hard to protect against that though.

Maybe something like this could do the trick ? :
https://www.codeproject.com/Articles/30216/Handling-Corrupt-user-config-Settings

(an alternative solution is given in comments)

[EncodedMango]

I have come across a very odd issue. Anytime I select any Upscaler/Downscaler to be Lanczos, MPDN gives an error when trying to render video.

"Specified argument was out of range of valid values."

Even if it's just luma and the video is being played at its resolution(no scaling)

Win10 x64 1703, MPDN x64

This seems to have started after Creator's Update. I think turab's issue (https://forum.doom9.org/showthread.php?p=1804039#post1804039)above may be related.

[Shiandow]

I have come across a very odd issue. Anytime I select any Upscaler/Downscaler to be Lanczos, MPDN gives an error when trying to render video.

"Specified argument was out of range of valid values."

Even if it's just luma and the video is being played at its resolution(no scaling)

Win10 x64 1703, MPDN x64

This seems to have started after Creator's Update. I think turab's issue (https://forum.doom9.org/showthread.php?p=1804039#post1804039)above may be related.

Turab's issue could be fixed by resetting the settings, is this the case with your issue as well?

[EncodedMango]

Turab's issue could be fixed by resetting the settings, is this the case with your issue as well?
Yes, that's how I initially fixed it after updating to CU. Then began checking what was causing it - Lanczos, infinitely repeatable.

All others work fine.

[Shiandow]

Any chance you can send me a copy of the settings that cause a crash and the ones that don't? I think I just need the Application.(64|32|AnyCpu).config files (as it doesn't seem to be related to a player extension).

[EncodedMango]

Any chance you can send me a copy of the settings that cause a crash and the ones that don't? I think I just need the Application.(64|32|AnyCpu).config files (as it doesn't seem to be related to a player extension).
Having issues with upload so I'll just give you contents.

The only difference between the two is these lines

Crashy:

<ChromaUpscaler yaxlib:realtype="MediaPlayerDotNet.Config.ChromaUpscalerSettings">
<Scaler yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.LanczosSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Scaler>
</ChromaUpscaler>


Working:

<ChromaUpscaler yaxlib:realtype="MediaPlayerDotNet.Config.ChromaUpscalerSettings">
<Scaler yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.JincSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Scaler>
</ChromaUpscaler>


In the above case I changed only Chroma Up to Lanczos for reproduction, but it happens for any of the 4 scalers.

Here's the entire(crashy) file's contents, Application.64.config :
<Configuration xmlns:yaxlib="http://www.sinairv.com/yaxlib/">
<ConfigVersion>
<Major>2</Major>
<Minor>0</Minor>
</ConfigVersion>
<GeneralSettings>
<RecentMediaFiles>
<String>F:\TD\TLS1.mkv</String>
<String>F:\TD\TLS2.mkv</String>
</RecentMediaFiles>
<RecentPlaybackPositions>
<FilePosition>

[Removed for Privacy]

</FilePosition>
</RecentPlaybackPositions>
<OptionsDialogSelectedNodePath>Video Renderer\Chroma Upscaler</OptionsDialogSelectedNodePath>
<LumaUpscalerUiSettings>
<Bicubic yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.BicubicSettings">
<Sharpness>0.75</Sharpness>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Bicubic>
<Softcubic yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.SoftcubicSettings">
<Softness>0.75</Softness>
</Softcubic>
<Lanczos yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.LanczosSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Lanczos>
<Spline yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.SplineSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Spline>
<Jinc yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.JincSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Jinc>
<Custom yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.CustomScalerSettings">
<ScalerGuid>647351ff-7fec-4eab-86c7-ce1bef43efd4</ScalerGuid>
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Custom>
</LumaUpscalerUiSettings>
<ChromaUpscalerUiSettings>
<Bicubic yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.BicubicSettings">
<Sharpness>0.75</Sharpness>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Bicubic>
<Softcubic yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.SoftcubicSettings">
<Softness>0.75</Softness>
</Softcubic>
<Lanczos yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.LanczosSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Lanczos>
<Spline yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.SplineSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Spline>
<Jinc yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.JincSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Jinc>
<Custom yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.CustomScalerSettings">
<ScalerGuid>647351ff-7fec-4eab-86c7-ce1bef43efd4</ScalerGuid>
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Custom>
</ChromaUpscalerUiSettings>
<LumaDownscalerUiSettings>
<Bicubic yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.BicubicSettings">
<Sharpness>0.75</Sharpness>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Bicubic>
<Softcubic yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.SoftcubicSettings">
<Softness>0.75</Softness>
</Softcubic>
<Lanczos yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.LanczosSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Lanczos>
<Spline yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.SplineSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Spline>
<Jinc yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.JincSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Jinc>
<Custom yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.CustomScalerSettings">
<ScalerGuid>647351ff-7fec-4eab-86c7-ce1bef43efd4</ScalerGuid>
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Custom>
</LumaDownscalerUiSettings>
<ChromaDownscalerUiSettings>
<Bicubic yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.BicubicSettings">
<Sharpness>0.75</Sharpness>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Bicubic>
<Softcubic yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.SoftcubicSettings">
<Softness>0.75</Softness>
</Softcubic>
<Lanczos yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.LanczosSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Lanczos>
<Spline yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.SplineSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Spline>
<Jinc yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.JincSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Jinc>
<Custom yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.CustomScalerSettings">
<ScalerGuid>647351ff-7fec-4eab-86c7-ce1bef43efd4</ScalerGuid>
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Custom>
</ChromaDownscalerUiSettings>
<PlayerStartupMode>Normal</PlayerStartupMode>
<ShowPlayerStats>False</ShowPlayerStats>
<ResumePlaybackPosition>True</ResumePlaybackPosition>
<ShowMenuBar>True</ShowMenuBar>
<ShowStatusBar>False</ShowStatusBar>
<ShowCaptionBar>True</ShowCaptionBar>
<AutoHideControlBar>False</AutoHideControlBar>
<LockToAspectRatio>False</LockToAspectRatio>
<SnapToScreenEdges>False</SnapToScreenEdges>
<AlwaysOnTop>True</AlwaysOnTop>
<LockWindowSize>False</LockWindowSize>
<LastSelectedMediaFileName>F:\TD\TLS2.mkv</LastSelectedMediaFileName>
<Volume>85</Volume>
<Mute>False</Mute>
<ResizeWindowToFitVideo>True</ResizeWindowToFitVideo>
<ResetWindowSize>True</ResetWindowSize>
<RememberWindowSizePos>False</RememberWindowSizePos>
<LaunchMediaInFullScreen>False</LaunchMediaInFullScreen>
<WindowBounds>
<Left>1038</Left>
<Top>555</Top>
<Width>408</Width>
<Height>330</Height>
</WindowBounds>
<WindowState>Normal</WindowState>
<UseDebugDialog>False</UseDebugDialog>
<ForceCrashOnAppHang>False</ForceCrashOnAppHang>
<AppHangTimeoutSeconds>15</AppHangTimeoutSeconds>
<SeekBarSizeInFullScreen>Medium</SeekBarSizeInFullScreen>
<AlwaysOnTopMode>Always</AlwaysOnTopMode>
<UseDarkThemeControlBar>True</UseDarkThemeControlBar>
<DoNotShowGetExtensionsDialog>False</DoNotShowGetExtensionsDialog>
<AudioScript>
<Guid>00000000-0000-0000-0000-000000000000</Guid>
</AudioScript>
</GeneralSettings>
<DirectShowSettings>
<AudioRenderer>df557071-c9fd-433a-9627-81e0d3640ed9</AudioRenderer>
<AudioRendererDevicePath>@device:sw:{E0F158E1-CB04-11D0-BD4E-00A0C911CE86}\{DF557071-C9FD-433A-9627-81E0D3640ED9}</AudioRendererDevicePath>
<ExtraAudioFilters />
<ExtraVideoFilters />
<LoadSubtitles>True</LoadSubtitles>
<UseXySubFilter>True</UseXySubFilter>
<ReduceXySubAliasing>False</ReduceXySubAliasing>
</DirectShowSettings>
<VideoRendererSettings>
<AdapterId>0</AdapterId>
<FrameComposerType>Direct3D11</FrameComposerType>
<OutputLevels>Auto</OutputLevels>
<OutputLimitedRange>False</OutputLimitedRange>
<UseFullScreenExclusiveMode>False</UseFullScreenExclusiveMode>
<UseFlipEx>True</UseFlipEx>
<DebugDx11>False</DebugDx11>
<OutputBitDepth>EightBits</OutputBitDepth>
<UseCustomAspectRatio>False</UseCustomAspectRatio>
<AspectRatioX>4</AspectRatioX>
<AspectRatioY>3</AspectRatioY>
<LetterBoxing>ApplyLetterBoxing</LetterBoxing>
<CustomFillRectangle />
<DecoderQueueSize>16</DecoderQueueSize>
<RenderQueueSize>12</RenderQueueSize>
<WindowedBackBufferSize>12</WindowedBackBufferSize>
<ExclusiveBackBufferSize>12</ExclusiveBackBufferSize>
<FseReenterDelay>500</FseReenterDelay>
<RenderQuality>MaxQuality</RenderQuality>
<ImproveChromaReconstruction>True</ImproveChromaReconstruction>
<ChromaOffset420>
<X>-0.25</X>
<Y>0</Y>
</ChromaOffset420>
<ChromaOffset422>
<X>-0.25</X>
<Y>0</Y>
</ChromaOffset422>
<NoWindowedVSync>False</NoWindowedVSync>
<ForceAltOclInterop>True</ForceAltOclInterop>
<LumaUpscaler yaxlib:realtype="MediaPlayerDotNet.Config.LumaUpscalerSettings">
<Scaler yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.JincSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Scaler>
</LumaUpscaler>
<LumaDownscaler yaxlib:realtype="MediaPlayerDotNet.Config.LumaDownscalerSettings">
<Scaler yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.JincSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Scaler>
</LumaDownscaler>
<ChromaUpscaler yaxlib:realtype="MediaPlayerDotNet.Config.ChromaUpscalerSettings">
<Scaler yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.LanczosSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Scaler>
</ChromaUpscaler>
<ChromaDownscaler yaxlib:realtype="MediaPlayerDotNet.Config.ChromaDownscalerSettings">
<Scaler yaxlib:realtype="MediaPlayerDotNet.Config.Scaler.JincSettings">
<Taps>Four</Taps>
<AntiRingingEnabled>False</AntiRingingEnabled>
</Scaler>
</ChromaDownscaler>
<Ditherer yaxlib:realtype="MediaPlayerDotNet.Config.DitheringSettings">
<Ditherer yaxlib:realtype="MediaPlayerDotNet.Config.Ditherer.DitherRandomSettings">
<Coloured>True</Coloured>
<PerFrameDithering>True</PerFrameDithering>
<Strength>1</Strength>
</Ditherer>
</Ditherer>
<FluidMotionSettings>
<Active>False</Active>
<OnlyWhenRateLessThanHalfDisplay>False</OnlyWhenRateLessThanHalfDisplay>
<ExceptMultiples>False</ExceptMultiples>
</FluidMotionSettings>
<RenderScript>
<Guid>00000000-0000-0000-0000-000000000000</Guid>
</RenderScript>
</VideoRendererSettings>
</Configuration>

[Shiandow]

Okay.. no obvious clues there. You said that resetting the config files fixed the issues, any idea what is different in the 'fresh' config files?

[Cy4n1d3]

Since the W10 Creators Update MPDN doesn't even start for me.
This applies to x86 and x64 versions.
Also tried deleting my settings, to no avail.

Any hints on how to debug this the easy way?

[Garteal]

MPDN works fine for me with the Creators Update with both a clean install and an upgrade.

I do however had an issue getting MPDN to run after a clean CU install (but no .NET app would probably run either) after the clean install because I had no keyboard in my my Region&Language settings in Windows for some reason. Everything was flawless after that.
(Just laying it out for people that might have the same issue)

Cy4n, is this a clean install or an upgrade and what happens exactly? Does your mouse show that something is happening but nothing pops-up?
Do you see any MPDN processes running in your task manager? Or have you tried to open it after a reboot?

[Cy4n1d3]

I do have a keyboard setting in Region & Language available and selected.

I installed the CU using the Upgrade Assistant (so, no clean install).
My mouse shows the spinning circle busy indicator thingy for a second, that's all.
There are no MPDN processes running according to the task manager.
Problem persists through full reboots.

Any way to get a debug log of what's happening?

[EncodedMango]

Okay.. no obvious clues there. You said that resetting the config files fixed the issues, any idea what is different in the 'fresh' config files?
Default doesn't have Lanczos selected on any scalers is my best guess, that's the only change that does anything.

My player config isn't complex, the only things I change are up/down scalers, playback resuming, Presentation API(Dx11) and disabling Fluid Motion because of occasional artifacts.

I have tried changing all of them, Lanczos is still the only culprit.

Ah, right so it still happens if you reset the settings and then select Lanczos again?
Yes.

[Shiandow]

Default doesn't have Lanczos selected on any scalers is my best guess, that's the only change that does anything.

My player config isn't complex, the only things I change are up/down scalers, playback resuming, Presentation API(Dx11) and disabling Fluid Motion because of occasional artifacts.

I have tried changing all of them, Lanczos is still the only culprit.

Ah, right so it still happens if you reset the settings and then select Lanczos again?

[Garteal]

Any way to get a debug log of what's happening?
Grab DebugView (https://technet.microsoft.com/en-us/sysinternals/debugview.aspx) and report back what is getting logged from it.

[turab]

I can also confirm that the problem I had before occurs when setting one of the scaler settings to Lanczos. This only happens with the 64-bit version of MPDN.

Also, after trying to set one of the scalers to Lanczos and failing, the following error eventually gets thrown (and the program then crashes):

Object reference not set to an instance of an object.

This appears to be due to a null pointer exception, as diagnosed using the Visual Studio debugger.

System.NullReferenceException occurred
HResult=0x80004003
Message=Object reference not set to an instance of an object.
Source=MediaPlayerDotNet
StackTrace:
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(Int64 , IList`1 , IList`1 )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(Int64 , IntPtr , IList`1 , IList`1 , IList`1 , Boolean )
at B.M.b(K )
at B.M.W()
at B.M.X()
at B.b.A.A()
at System.Threading.ExecutionContext.RunInternal(ExecutionContext executionContext, ContextCallback callback, Object state, Boolean preserveSyncCtx)
at System.Threading.ExecutionContext.Run(ExecutionContext executionContext, ContextCallback callback, Object state, Boolean preserveSyncCtx)
at System.Threading.ExecutionContext.Run(ExecutionContext executionContext, ContextCallback callback, Object state)
at System.Threading.ThreadHelper.ThreadStart()

Possibly a separate issue caused because the previous one wasn't handled properly.

[donger]

How am i supposed to play a DVD? I have extensions installed, i drag IFO file and the program just hangs.
How can i access the normal DVD menu with subtitles, audio etc

oh and also, any way i can use this magnificent software with rate tuner on my android box pc?

[Shiandow]

Unless I somehow missed it MPDN doesn't support DVD menus as of yet.

[Hrxn]

By the way, is a new release of MPDN in the works?

[sauma144]

By the way, is a new release of MPDN in the works?

Maybe something big like dx12/vulkan new renderer. :rolleyes:

[donger]

Can someone check if my settings are right? I don't know if my rate tuner is working properly
http://i.imgur.com/WsaAMY2.jpg
also i am getting gtx 1080, what setting should i use to crank up the quality?

[foozoor]

Maybe something big like dx12/vulkan new renderer. :rolleyes:
Mpv has already (not in master) a working vulkan renderer with +95% of its current opengl renderer and massive performance improvements. So it's not impossible for the MPDN team too!

[huhn]

oh i would like to see that improvement.
the main benefit of vulkan is the better use of the CPU something that is not important for video rendering at all...

not saying it is completely useless but never except something massive form these APIs for video rendering it's made for games in mind.

[foozoor]

oh i would like to see that improvement.
the main benefit of vulkan is the better use of the CPU something that is not important for video rendering at all...

not saying it is completely useless but never except something massive form these APIs for video rendering it's made for games in mind.

OpenGL has some nasty limitations... for video rendering too.
However the massive improvements are mostly for Linux, I didn't try on Android or Windows and I don't know if it's currently tested on these platforms. For Windows there is Angle that already "converts" OpenGL instructions to D3D9/D3D11.

[madshi]

I can imagine Vulkan being a noticeable improvement over OpenGL, but I don't think using Vulkan would help either MPDN or madVR. AMD/Nvidia spend most of their time optimizing their drivers for Direct3D, so using Direct3D should offer the best performance - on Windows at least. Of course for a multi platform project using Vulkan sounds like the best solution.

[aufkrawall]

Mpv has already (not in master) a working vulkan renderer with +95% of its current opengl renderer and massive performance improvements. So it's not impossible for the MPDN team too!
Is there some evidence for the performance improvements? I really doubt this since the OpenGL renderer is doing absolutely fine with low CPU and GPU usage on both Windows and Linux (who knows about that crippled Mac environment which doesn't get Vulkan anyway).

[stax76]

What could be the reason if the player don't start? I just hear some error beep.

[Shiandow]

If it doesn't start at all that sounds like the issue is in MPDN itself, possibly some problems with .Net. Are there any messages in the windows event logs? You could also try running Debugview (https://technet.microsoft.com/en-us/sysinternals/debugview.aspx).

[stax76]

I've tried both but didn't find anything.

[Shiandow]

Odd, I'd expect at least something to show up in the event log. What version of MPDN and Windows are you using?

[stax76]

W10 Pro 1703 x64 and MPDN just downloaded.

[Shiandow]

Why is it always windows 10 with the weird .net issues? Anyway, reinstalling the .net framework might help, but maybe MPDN itself needs to be recompiled before it works properly on windows 10. In theory neither of these should be necessary, but then again in theory it shouldn't just crash on one PC and run fine on another.

[madshi]

It's also always Windows 10 with the weird GPU driver issues... :(

[stax76]

It's not that Windows 7 don't has issues... batch file processing is broken for instance if you use unicode and Umlaute, I need to deal with batch files for several reasons like x265 requiring piping or piping with eac3to.

Note there are serious implementation bugs in Windows's code page 65001 support which will break many applications that rely on the C standard library IO methods, so this is very fragile. (Batch files also just stop working in 65001.) Unfortunately UTF-8 is a second-class citizen in Windows.

https://stackoverflow.com/questions/388490/unicode-characters-in-windows-command-line-how

Win 10 has it fixed along with 1000 other issues.

[YxP]

Maybe we have reached the maximum amount of code for humans to handle with windows 10 or something. Everyone seems to have at least one problem with it, but it rarely is the same. Did I just hear "minimal MadOS for HTPC"?

[Hrxn]

What could be the reason if the player don't start? I just hear some error beep.

This definitely shouldn't happen. I don't know if it is related to .NET at all, never heard of something specific here, to be honest.

I don't know, anything unusual about your Windows configuration, programs running in the background that are not considered as "common"?

[stax76]

I use win defender and autohotkey with a simple hotkey script.

ApplicationFrameHost, AutoHotkey, backgroundTaskHost, Calculator, conhost, csrss, dasHost, dllhost, dwm, explorer, fontdrvhost, GoogleCrashHandler, GoogleCrashHandler64, googledrivesync, HxCalendarAppImm, HxTsr, IAStorDataMgrSvc, Idle, igfxCUIService, igfxEM, IpOverUsbSvc, IPROSetMonitor, jusched, lsass, Memory Compression, Microsoft.Msn.Weather, MSASCuiL, MsMpEng, NisSrv, nvcontainer, NVDisplay.Container, NVIDIA Web Helper, NvTelemetryContainer, nvtray, OneDrive, powershell, PresentationFontCache, RemindersServer, RtkNGUI64, RuntimeBroker, SearchFilterHost, SearchIndexer, SearchProtocolHost, SearchUI, SecurityHealthService, services, SettingSyncHost, ShellExperienceHost, sihost, SkypeHost, smartscreen, smss, SpeechRuntime, spoolsv, sqlwriter, StaxLauncher, svchost, System, SystemSettings, taskhostw, thunderbird, vmcompute, vmms, vmnat, vmnetdhcp, vmware-authd, vmware-usbarbitrator64, wininit, winlogon, WinStore.App, WmiPrvSE, WUDFHost

[Hrxn]

Windows Defender itself is definitely not the problem, and Autohotkey shouldn't be the culprit either.

But it seems that you have VMware running? Maybe that's locking the GPU driver in some way, I don't know..

[donger]

Is there any reason that after correctly using rate tuner i still have to delay sound 150ms to avoid lipsync issues?

[Shiandow]

Well, the most obvious reason would be that the audio on the video you're trying to watch is early by about 150ms.

[pose]

Shiandow, hello! Can you tell me your preferences in scalers and shaders with MPDN, please? :) What things you use in different scenarios with various sources? Thank you!

[Shiandow]

Shiandow, hello! Can you tell me your preferences in scalers and shaders with MPDN, please? :) What things you use in different scenarios with various sources? Thank you!

With the latest extensions my default was Bilateral > SSIM > Deband Experimental > SSSR. All using the default settings (which shouldn't be too surprising, since I defined the defaults). I use this for pretty much all sources (although I heavily use script groups with shortcuts to switch between different settings or turn parts of the processing on or off).

However keep in mind that my setup is basically designed to test and improve the renderscripts I made. And that my version of the extensions might differ slightly from the current available version.

For example, one of the more recent changes I made allowed debanding to be done before chroma scaling (which is nice since you can work with the original 8 bit values). I've also added a shader at the end to convert from a 2.2 gamma to 2.4, since the current standards seems to suggest that SDR media should be displayed on a 100 nit screen with a 2.4 gamma.

[huhn]

I've also added a shader at the end to convert from a 2.2 gamma to 2.4, since the current standards seems to suggest that SDR media should be displayed on a 100 nit screen with a 2.4 gamma.

well you missed something important here.

this only makes sense when your screen is perfectly calibrated to 2.2 but... just calibrate it to 2.4 if that is what you want.
and gamma should be chosen on room ambient light.

the current "standard" is BT 1886 BTW.

[Hrxn]

With the latest extensions my default was Bilateral > SSIM > Deband Experimental > SSSR. [...]

SSIM? As in SSIM Downscaler?

Well, I've seen/heard this before, but never found an convincing explanation so far. What is the point of the downscaling step? How does it really improve anything?

[Shiandow]

well you missed something important here.

this only makes sense when your screen is perfectly calibrated to 2.2 but... just calibrate it to 2.4 if that is what you want.
and gamma should be chosen on room ambient light.

the current "standard" is BT 1886 BTW.

My screen happens to be pretty close to a 2.2 power curve, which is good for webpages and text, but not so good for video. I agree that ambient light means you need a slightly different gamma, which is why I've added a hotkey to enable/disable the gamma conversion.

SSIM? As in SSIM Downscaler?

Well, I've seen/heard this before, but never found an convincing explanation so far. What is the point of the downscaling step? How does it really improve anything?

Most downscaling methods end up softening the image. The SSIM downscaler basically sharpens the image just enough to make it as sharp as the original. It's subtle, but it is noticeable.

[pose]

Shiandow, what is the best Luma scaler to use in combo with SSSR (hybrid mode) in your opinion? :) I do 720p -> 1080p mostly.

[donger]

aaand frames tearing out of nowhere
this on top of ZERO help in here on the forum = uninstalled

[huhn]

so zero help means ignoring the dev that tried to help?

[stax76]

Problem is there is only one core dev and he is offline since a couple of months, I was once offline 1,5 years, I hope he is fine.

[Zachs]

Problem is there is only one core dev and he is offline since a couple of months, I was once offline 1,5 years, I hope he is fine.

There's actually two Devs. And I'm not really offline. Lol.

Just been busy with work. There's not much I wanted to do to improve the player anyway that is beyond what needs to be done within the core app. It's so extensible that anyone could add features to it.

[Zachs]

What could be the reason if the player don't start? I just hear some error beep.

Did you ever figure out what was wrong? The only thing I could think of is that NGEN generated incorrect code on your machine, which could only happen if you have had hardware...

[stax76]

I was just trying to say you haven't been posting since a few months. Who is the other core dev then? I didn't notice anybody seriously trying to help. I've not figured out what is wrong, I was hoping a dev will tell what to do. Please note that when I first tried the player when it was new the exact same problem was happening and it had something to to with the subtitle filter so obviously there was no properly working error handling.

[Zachs]

There's Shiandow too.

When you said it beeped when you tried starting the player, did you mean the player window itself wouldn't even appear? If so that's way before anything to do with MPDN. The first thing MPDN does is it shows the player window, before it loads Configs etc.

[stax76]

I was thinking Shiandow makes plugins and don't work on the core.

The window is only visible a few milliseconds.

[Zachs]

Hmm did you try running MPDN without extensions? If you have a clean VM (download one from MS if you don't), you could try it in there. MPDN runs fine with the default settings even in a VM.

[stax76]

I tried the portable and the installer both x64 and x86, it works in my vmware client Win 7 but not on my system Win 10, in Visual Studio it shows:

Exception thrown at 0x7498B802 in MediaPlayerDotNet.exe: Microsoft C++ exception: std::exception at memory location 0x00EFE914.

There is nothing logged in the Windows Event Log.

Suggestion: Upload a build with debug symbols.

[stax76]

Is Windows 10 officially supported and have you tested it on Windows 10 1703? Has anybody got it working on Windows 10 1703?

[mrcorbo]

Is Windows 10 officially supported and have you tested it on Windows 10 1703? Has anybody got it working on Windows 10 1703?

It works here on the latest Fast-Ring Insider build, FWIW.

[Garteal]

I've been using MPDN since release with zero issues on Windows 10 (including build 1703). Whatever problems you're having is something interfering on your end. The least you can do is provide some detail of what you have tried, your setup, what processes are running, etc. I mean you're a dev and all, you know what you can do to help us help you.

[stax76]

I'm just starting as player dev ;), it was the 4K display and/or 288 DPI setting, half of the players don't support it I forgot, one of the many reasons I get into player development.

[Zachs]

I'm just starting as player dev ;), it was the 4K display and/or 288 DPI setting, half of the players don't support it I forgot, one of the many reasons I get into player development.

What? So you knew what the problem was all along?

[stax76]

I did not know it along, I didn't spend much thought on it until being reminded being a dev and that if I don't help myself probably nobody will.

[Garteal]

You need to work on your reading comprehension.

Anyway, 4k works fine. The issue is that 288DPI you have set. I've tried a few other odd custom DPI values thinking there's an issue with odd custom values, but what I tested worked fine.

[Shiandow]

I was thinking Shiandow makes plugins and don't work on the core.

The window is only visible a few milliseconds.

For future reference I am now also a dev for the core code. But my changes so far has been mostly limited to some new timing code for the renderer (and also some dithering code, but I recently found out the actual rendering uses a different code path which still used the old ditherers).


Suggestion: Upload a build with debug symbols.

Unfortunately right now not all parts of the core code can be made open source (and therefore need to be obfuscated, so no debug symbols).

[vinnytx]

I hope 10-bit support will be improved.
I mean that it should be available even in windowed mode.
Also, 8-bit videos should not have 10-bit output

[madshi]

Why should 8bit videos not have 10bit output? Videos are YCbCr. Output is usually RGB. The conversion is floating point math, so you have more than 8bit after YCbCr -> RGB conversion.

[vinnytx]

I thought that conversion was lossy.

[huhn]

yes float to integer is lossy.
but float to 8 bit is more lossy than float to 10 bit.

[vinnytx]

I have problems when I play 4k 10-bit videos in fullscreen. I have this error.
https://i.imgur.com/vDWtR2G.jpg

I tried many settings in "Video Renderer" > General/Presentation/Composition and I couldn't solve the problem.


Example
http://www.mediafire.com/file/a3knqi0uaq5ssng/10_bit.mp4

[Shiandow]

You don't have enough memory on your GPU apparently. Are you using any renderscipts? Those could make things worse. I think SSIMDownscaler in particular could end up using quite a lot of memory.

[vinnytx]

Thanks
My video card is 960 gtx with 4GB. I don't use any render script. Have you tested the video I posted?

[ryrynz]

I have problems when I play 4k 10-bit videos in fullscreen. I have this error.


Usually these sorts of problems are related to your graphics card driver installation, uninstall and then reinstall the latest ones.

[vinnytx]

I uninstalled and reinstalled Nvidia driver and nothing has changed.

I found the problem: when I enable DXVA2 copy-back in Lav Filters x86 the problem occurs.
No problem with MPDN x64 and Lav Filters x64

[Shiandow]

Are you sure MPDN is actually part of the problem? Does Lav Filter x86 work when you use a different video player?

[vinnytx]

Yes, I'm sure. Zoom Player and PotPlayer don't have this problem

[stax76]

Any news on the 2880 DPI issue? I've my own player now but still follow development of other players.

[Shiandow]

Any news on the 2880 DPI issue? I've my own player now but still follow development of other players.

I've had a quick look and it looks like MPDN just didn't like it when it had to resize before it had a chance to initialise everything. I've managed to fix it.

Surprisingly most things actually worked. Apart from some small cosmetic problems.

Yes, I'm sure. Zoom Player and PotPlayer don't have this problem

Still not sure why the x64 vs x86 MPDN would behave differently. One possible explanation is that there is a problem in the settings of the x64 version. Also, the error dialog should have a button to copy the detailed information, if you could get the stacktrace from there I might be able to figure out where the problem occurs.

Also, there's still the possbility that you're actually just running out of GPU memory, and that minor changes in settings affect whether you run out completely or not.

[vinnytx]

I have problems with MPDN x86.

I uninstalled x86 version and I installed x64. At the moment I can't do other tests. I tried MPDN x86 portable version and it gives me an error regarding filters that are not properly configured (video graph failed to render). That's not a problem since I configured my pc for x64 video players

[Anime Viewer]

I'm encountering a problem during playback of:
http://tracker.minglong.org/torrents/%5BMezashite%5D%20Aikatsu%20Stars!%20-%20073%20%5BBAE1A1F9%5D.mkv.torrent

At 20:06 (twenty minutes and six seconds) into the video MPDN stops responding. I tried running in debug mode, but it didn't generate a crash dump. My settings are as follows:

MPDN 64-bit edition
version 2.48.0.3603
XySubFilter with reduce aliasing checked
Prefer: Image Quality
Always use new OpenCL Interop API (I tried disabling this and restarted, but it still crashed at the same part, so I have ruled this out)
Presentation API: Driect3D 11
Graphic Adapter: Nvidia GeForce GTX 680M
Video Renderer:
Colormetric: Automatic
Output Range: Full
Video Output bit depth: 10-bit
Script Chain Deband -> SuperChromaRes -> Super-xBR
Ordered dithering
I tried in full screen and in a window, and both crashed but anyway here is what I'm using for scalers:
Luma up: Jinc with anti-ring
Chroma up: Jinc 4 tap with anti-ring
Luma Down: Softcubic 75
Chroma down: Bicubic 75

If you need any more info, or testing with another setting let me know.

BTW: I ran the video with MPC-HC (with MadVR) and didn't encounter the crash/freeze.

[carpedee]

I really like this video player, I feel like it puts less strain on my cpu than MPC used with the LAV filters (DX11 helps a lot I guess).

I have a few questions/suggestions regarding some features I didn't find:

- Is it possible to edit the hotkeys like the one to enter in fullscreen mode for example?
- Is it possible to run several instances of MPDN at the same time?
- Is there a hotkey to take a screenshot when playing a video?

[kingpage]

I've just been reading up on Madshi's NGU (NG1). It's very impressive. Is there any plan on bring it to MPDN? Does anyone know if it's easier to run than SSSR (SuperxBR, NNEDI3, etc)?

[aufkrawall]

NGU will remain madVR exclusive and questions regarding it should be discussed in madVR thread. It got different quality settings and the performance costs range goes from relatively close to Jinc to ~NNEDI3 64 neurons.

[Heaud]

Is there a way for MPDN to gain window focus when opening a video? Everytime I open a file, the video loads but does not gain focus. Operating system is Windows 10 64-bit with Creators Update.

[Shiandow]

I've just been reading up on Madshi's NGU (NG1). It's very impressive. Is there any plan on bring it to MPDN? Does anyone know if it's easier to run than SSSR (SuperxBR, NNEDI3, etc)?

Whether NGU will be brought to MPDN, is up to Madshi, and really there's no reason for him to help improve a video player which uses a different renderer. Despite that Madshi has already helped MPDN quite a bit in various ways, but expecting him to implement all of MadVR's functionality in MPDN is just not realistic.

Is there a way for MPDN to gain window focus when opening a video? Everytime I open a file, the video loads but does not gain focus. Operating system is Windows 10 64-bit with Creators Update.

As of now that's the expected behaviour when MPDN is already running when the file opens. It would be fairly easy to change that with a player extension.

You can make MPDN stay on top while playing, which might help. You can enable this in the context menu > Window > Always on Top. And you can limit it to only stay on top while playing in Options > General > Window.

[Zachs]

I have come across a very odd issue. Anytime I select any Upscaler/Downscaler to be Lanczos, MPDN gives an error when trying to render video.

"Specified argument was out of range of valid values."

Even if it's just luma and the video is being played at its resolution(no scaling)

Win10 x64 1703, MPDN x64

This seems to have started after Creator's Update. I think turab's issue (https://forum.doom9.org/showthread.php?p=1804039#post1804039)above may be related.

Ok. Bit of an update regarding this particular bug. It's 100% on Microsoft. If you are up-to-date with your Windows Update, this bug will magically go away.

[Heaud]

As of now that's the expected behaviour when MPDN is already running when the file opens. It would be fairly easy to change that with a player extension.

You can make MPDN stay on top while playing, which might help. You can enable this in the context menu > Window > Always on Top. And you can limit it to only stay on top while playing in Options > General > Window.
Having the Always on Top is not the best option for my use as it will still not gain focus so keyboard shortcuts will not be active until I click on the video frame. Another issue is that clicking on the video frame also pauses the video so to gain focus so that keyboard shortcuts can be used would cause the video to be paused.

You mentioned that it would be easy to remedy this inconvenience with a player extension. Is it possible for someone with basic, but inexperienced understanding of scripting such as myself be able to create an extension that enables automatic window focus upon loading a video file?

[kingpage]

Whether NGU will be brought to MPDN, is up to Madshi, and really there's no reason for him to help improve a video player which uses a different renderer. Despite that Madshi has already helped MPDN quite a bit in various ways, but expecting him to implement all of MadVR's functionality in MPDN is just not realistic.


That's true. Can MadVR be used with MPDN? I can't seem to get it to work.

[ryrynz]

No. It cannot.

[nevcairiel]

That's true. Can MadVR be used with MPDN? I can't seem to get it to work.

No, MPDN can only use its own renderer.

[Zachs]

I've just been reading up on Madshi's NGU (NG1). It's very impressive. Is there any plan on bring it to MPDN? Does anyone know if it's easier to run than SSSR (SuperxBR, NNEDI3, etc)?

MPDN extension is an open source project - madshi's NGU is a closed source / proprietary algo, much like its madVR. It's inherently incompatible with MPDN extension's core concepts - our devs believe in sharing knowledge with the public with appropriate licensing - something the modern world agrees with (.net core, almost a hundred thousand nuget packages etc.).

The only thing I haven't been able to open source is the core player itself, mainly due to the strict copyright laws here where I live.

That said, it's unlikely madshi would ever make NGU compatible with open source licensing, even if we somehow know how it's done.

[aufkrawall]

I really like MPDN and especially the open render scripts, so please don't stone me to death. :o
However, there's also mpv, which is 100% open source, cross platform and constantly developing at a fast pace.
Sure MPDN still has its unique features, but I think it would be more helpful for more users to directly offer shaders, knowledge etc. to the mpv project in one or the other way (even though open source scene atmosphere might be a bit difficult, but dunno).

[Zachs]

I believe there's no reason why contributing to MPV wouldn't directly benefited MPDN or vice versa. That's the whole point of the open source initiative.

[Shiandow]

You mentioned that it would be easy to remedy this inconvenience with a player extension. Is it possible for someone with basic, but inexperienced understanding of scripting such as myself be able to create an extension that enables automatic window focus upon loading a video file?

Should be pretty doable. If you're up for it I would strongly recommend trying it yourself. Some things will be easier if you download Visual Studio (the community edition is free for personal use) but it's not necessary (it'll just be slightly harder to figure out what went wrong). You might want to go to Options > General > Debugging and enable the debugging dialog, because otherwise you're not going to get any useful information if it fails.

Anyway, here's a rough guid to what you'd need to do. If you're unsure about anything you can probably google the answer, or if not come back and ask in this thread.

To start I'd recommend basically copying the GoTo (https://github.com/zachsaw/MPDN_Extensions/blob/master/Extensions/PlayerExtensions/GoTo.cs) script. Just copy that file somewhere and open it up in your favorite text editor.

That extension already checks when a new file opens in order to enable / disable its item in the menu (try to figure out how it does this), so we can just reuse that to do something else (like bring the window to the front). The rest we don't need, you can remove everything but the "Descriptor" and the "Initialize", "Destroy" and "PlayerStateChanged" methods.

Then you need to fill out the descriptor with new information (most of this is optional, but you do need to change the GUID to another one, otherwise MPDN can't tell it apart from the GoTo extension). And be sure to rename the GoTO class and anything else that says "GoTo" to something more sensible.

After that bit of bookkeeping is over you can change the PlayerStateChanged method to actually bring the window into focus. You'll probably need to start with checking if the player started playing, using something like "if (e.NewState == PlayerState.Playing)". And then bring the window to the front. Some quick googling suggests calling "Activate()" on the window form (which is "Player.ActiveForm"), but not everyone got this to work apparently. There are a lot of other suggestions on google or stackoverflow on how to get window to the front in C# so you should probably try one of those if the simple method doesn't work.

After that you can copy the file containing your extension to MPDN / Extensions and it should work (although don't be too disheartened if it doesn't, try the debugging dialog and see if you can figure out what's wrong). Best of luck!

I really like MPDN and especially the open render scripts, so please don't stone me to death. :o
However, there's also mpv, which is 100% open source, cross platform and constantly developing at a fast pace.
Sure MPDN still has its unique features, but I think it would be more helpful for more users to directly offer shaders, knowledge etc. to the mpv project in one or the other way (even though open source scene atmosphere might be a bit difficult, but dunno).

I might be a tad biased, but in MPV their equivalent to "renderscripts" is some weird combination of GLSL with some 'special comments'. In MPDN you can use the whole of C# as your 'special comments' and use that to manipulate HLSL / OpenCL / DirectCompute shaders. Sure their API does allow people to port scripts to MPV, but it doesn't give you a lot of support for actually developing them. It also makes it very hard to tell what's happening, which isn't great for sharing ideas.

It's possible there's a more advanced API somewhere that I'm not aware of.

[pirlouy]

That's the whole point of the open source initiative.
Exactly: split/fork as soon as possible, in order to lose potential users, instead of trying to have one unique project, easier for newcomers. Total Linux Fork users < 1%. Sorry for this grievance.

I don't know why you're still mad at Madshi. Sometimes you don't want someone to tell you what to do, you want to do it your way, so closed source is an option. Or another example, imagine some company wanna buy his NGU algorithm. You have billion reasons to want closed-source project, the same as open-source. As good as your player is, whatever the reasons, it is still closed-source, and nobody will blame you for that.

Ahah it's fun, it was a long time since we had a discussion about Open/Closed Source. :)
I hope I didn't piss you off. I let you have the final word in the end.

@Shiandow: maybe you could ask a request on MPV tracker. I'm sure they would be interested in a "shader coder" (sorry if it's an approximate description).

[madshi]

I'm not sure if Zachs is mad at me. FWIW, there was a time when I wasn't very friendly to him, and I kinda regret that. Sadly, in forums one often writes things without really thinking it through, and then it's hard to take it back, because it's all "written in stone", so to say.

[jkauff]

There are pluses and minuses to both open and closed source applications (Richard Stallman notwithstanding). Microsoft, for example, recently has been open sourcing a lot of their work, but you're not going to get the source code for Windows or Office anytime soon. Too many salaried people working on those products.

madshi, bless him, has been sharing the fruits of his labor for many years now for free--and in fact won't take donations. For my part, I wouldn't want madVR to be a community project. It's nice to have one person in charge, especially one who takes as much user feedback as madshi does. I look forward to the day when I can give him my money.

There's plenty of room for madVR with its large base of supported media players, MPDN, mpv, and any other innovative video solution that comes along. The real shame is that most people still use Windows Media Player or VLC because they don't know or care that there are superior alternatives.

[aufkrawall]

It's possible there's a more advanced API somewhere that I'm not aware of.
I must confess that I got a bit lazy and got satisfied with good ol' Jinc scaling (and mpv's ewa_lanczossharp looks quite sharp for downscaling if scaling factor isn't too high).
But there also is e.g. NNEDI3 available as a usershader, doesn't seem like GLSL is a showstopper (I gues there is nothing with OGL without annoyance :p ).
Like I said, MPDN still has its very own unique features, which is totally fine and great.
It's just that, afterall, it's only about watching videos. Refining scaling quality etc. is cool, but it's just annoying how hard it is to use MPDN or madVR (+ players) for e.g. online streaming. It just doesn't meet my media habits anymore. With mpv + youtube-dl I basically can watch anything which is available for free by just pointing the URL to the player (YouTube, YouTube live, YouTube HDR, general HLS, files located online...).

[Anime Viewer]

After updating to the most recent version I'm getting the following error message:

===================================

Specified argument was out of the range of valid values. (MediaPlayerDotNet)

------------------------------
Error Type = System.ArgumentOutOfRangeException
Error Message = Specified argument was out of the range of valid values.
Error Source = MediaPlayerDotNet
Error Site = Mpdn.IScaler C(Mpdn.IScaler)
Error occurred = at a.C.C(IScaler )
at a.C.b(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.P(Object , EventArgs )
at System.EventHandler.Invoke(Object sender, EventArgs e)
at System.Windows.Forms.ToolStripItem.RaiseEvent(Object key, EventArgs e)
at System.Windows.Forms.ToolStripMenuItem.OnClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleMouseUp(MouseEventArgs e)
at System.Windows.Forms.ToolStrip.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.ToolStripDropDown.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.Control.WmMouseUp(Message& m, MouseButtons button, Int32 clicks)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.ToolStrip.WndProc(Message& m)
at System.Windows.Forms.ToolStripDropDown.WndProc(Message& m)
at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)
OS Version = Microsoft Windows NT 6.3.9600.0
Total Physical Memory = 15.9 GB
Available Physical Memory = 11.2 GB
Date Time = 9/30/2017 11:50:13 AM

I've checked and applied the one Windows Update I didn't have, but even after applying that and restarting the computer I still get the error.

[Shiandow]

But there also is e.g. NNEDI3 available as a usershader, doesn't seem like GLSL is a showstopper (I gues there is nothing with OGL without annoyance :p ).

GLSL isn't the problem. It's just that MPV's usershaders aren't much more than GLSL with some extra metadata. In MPDN the metadata is replaced with C#, which is a heck of a lot more powerful.

The only reason it's unfortunate that they use GLSL (although portability kind of forced that decision on them) is that this means you can't just automatically convert MPDN's renderchains into usershaders.


It's just that, afterall, it's only about watching videos. Refining scaling quality etc. is cool, but it's just annoying how hard it is to use MPDN or madVR (+ players) for e.g. online streaming. It just doesn't meet my media habits anymore. With mpv + youtube-dl I basically can watch anything which is available for free by just pointing the URL to the player (YouTube, YouTube live, YouTube HDR, general HLS, files located online...).

You don't like the 3DYD Youtube filter?

For what it's worth, I have my old player extension that downloads a stream using youtube-dl and loads it into MPDN. I didn't work on it much because I didn't think it was too useful when there's a nice DirectShow filter for youtube urls. But if there's a demand for it I could see if I could polish it up a bit.

[aufkrawall]

You don't like the 3DYD Youtube filter?

It doesn't work reliably for me with 4k 60fps video, playback often stalls after some time because it's not streaming fast enough.
So far I can only achieve stable playback with mpv + youtube-dl (and browsers, of course).


For what it's worth, I have my old player extension that downloads a stream using youtube-dl and loads it into MPDN. I didn't work on it much because I didn't think it was too useful when there's a nice DirectShow filter for youtube urls. But if there's a demand for it I could see if I could polish it up a bit.
I already made a fuss about it in the MPC HC thread. :devil:
It has several advantages like expanding general HSL functionality, special media site functionality or allowing streaming of YouTube live.
I'd think if more users knew about this, they'd find it quite useful.

Though I'm not sure if some basic youtube-dl integration alone is the holy grail. I also tried piping youtube-dl's output to MPC HC and it's still limited in terms of e.g. seeking compared to mpv.

[Shiandow]

Though I'm not sure if some basic youtube-dl integration alone is the holy grail. I also tried piping youtube-dl's output to MPC HC and it's still limited in terms of e.g. seeking compared to mpv.

Out of interest, how exactly do you use youtube-dl with mpv? Do you just pipe the output from youtube-dl into mpv?

[aufkrawall]

You drop its executable into mpv's folder (or install it as a dependency on Linux) and it's automatically used when opening URLs with mpv.
It's not used for the actual streaming but for source discovery. At least that's my conclusion from seeing its process closing after playback start.

[Cy4n1d3]

Ok. Bit of an update regarding this particular bug. It's 100% on Microsoft. If you are up-to-date with your Windows Update, this bug will magically go away.
I am fully updated using latest Win10 (1703) and still get this error.
After downgrading the player from 2.49 to 2.48 (still using 1.31 renderscripts) the error goes away and everything is working as before.
So even if this error / bug is somehow occuring due to Windows malfunctioning, some code change inbetween 2.48 and 2.49 seems to trigger it.

Regards

[mringis]

After updating to the most recent version I'm getting the following error message:

===================================

Specified argument was out of the range of valid values. (MediaPlayerDotNet)

------------------------------
Error Type = System.ArgumentOutOfRangeException
Error Message = Specified argument was out of the range of valid values.
Error Source = MediaPlayerDotNet
Error Site = Mpdn.IScaler C(Mpdn.IScaler)
Error occurred = at a.C.C(IScaler )
at a.C.b(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.P(Object , EventArgs )
at System.EventHandler.Invoke(Object sender, EventArgs e)
at System.Windows.Forms.ToolStripItem.RaiseEvent(Object key, EventArgs e)
at System.Windows.Forms.ToolStripMenuItem.OnClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleMouseUp(MouseEventArgs e)
at System.Windows.Forms.ToolStrip.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.ToolStripDropDown.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.Control.WmMouseUp(Message& m, MouseButtons button, Int32 clicks)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.ToolStrip.WndProc(Message& m)
at System.Windows.Forms.ToolStripDropDown.WndProc(Message& m)
at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)
OS Version = Microsoft Windows NT 6.3.9600.0
Total Physical Memory = 15.9 GB
Available Physical Memory = 11.2 GB
Date Time = 9/30/2017 11:50:13 AM

I've checked and applied the one Windows Update I didn't have, but even after applying that and restarting the computer I still get the error.

same with me :scared:

I'm on win 7 64. go back to 2.48 for now

[Zachs]

Hmm I'll investigate. I'm no longer getting the error but I did at some stage.

Have you guys tried uninstalling MPDN fully and reinstalling? Make sure you backup your settings before you do that. And try doing a windows restart after reinstalling.

[Anime Viewer]

Hmm I'll investigate. I'm no longer getting the error but I did at some stage.

Have you guys tried uninstalling MPDN fully and reinstalling? Make sure you backup your settings before you do that. And try doing a windows restart after reinstalling.

I did try a complete uninstall and reinstall of MPDN as well as the extensions, rebooted, reinstalled (clean install of both MPDN and extensions), rebooted again, and then launched MPDN. The problem was still present.

[thewildsun]

I'm having a problem with video play speed and subtitle sync. If I set the rate to 0.75x, the subtitles are too slow, and if I set the rate to 1.25x, the subtitles are too fast. Is this a bug?

I'm on 2.48 since I get the "Specified argument was out of the range of valid values. (MediaPlayerDotNet)" error on 2.49, on Win10.

[Heaud]

I have an odd issue in which MKV muxed from a MP4 is not playing at the intended 60FPS. The source is from a recording of a game which plays fine from the MP4 container, but I get lots of repeated frames and a reported average video rate of 39.9Hz based on the renderer overlay (https://img.warywolf.net/k0VYm.png). I am using MPDN 2.49.0.3606 and the latest version of MKVToolNix to mux.

Any ideas as to what might be going on? YouTube was able to render a video upload at 60FPS with the same MKV file so I am guessing the muxing isn't causing a problem with playback.

[huhn]

do you have a short sample?

[Heaud]

do you have a short sample?

Here is a link to the sample (http://www113.zippyshare.com/v/d8BtOYNO/file.html) (warning: there is audio) which is only a 18 second highlight of a match.

[huhn]

ok that a little bit strange it seems like the file is VFR with relative stable 40 FPS. maybe the file is stable 40 FPS and the game was using something like adaptive V-sync to stay stable at 40.
but the file claims to be constant 60 FPS in media info.

i'm not really sure but this line for audio : Frame rate : 46.875 FPS (1024 SPF)
but that's not something i know alot about.

[mikezilla2]

so why should i use this over MPC HC BE / modern codecs ?

[TaurezAG]

These two features will be very useful

1. Screenshot
A screenshot feature will be very useful, even more so if before/after processing(like scaling/RenderScript) screenshots can be taken separately.

2.
Additional input file format support for the 3DLUT RenderScript.
Currently, it only supports the .3dlut format which from what I've seen is supported only by a few software, and also just by calibration software. I would like to use 3DLUT for color correction. I'm using DaVinci Resolve. This is the format specification for the .cube files generated by Resolve: Link (https://forum.blackmagicdesign.com/viewtopic.php?f=21&t=40284)
I've attached a sample 3DLUT file generated by Resolve.

[huhn]

the 3D LUT implementation never worked correctly.

DisplayCAL supports both IRIDAS and 3D LUTs.
so you just needs a profiling of your screen (it may even work with profiling from a different program) and you can create a LUT for both programs in a very short time with the "same" results.

[TaurezAG]

the 3D LUT implementation never worked correctly.

DisplayCAL supports both IRIDAS and 3D LUTs.
so you just needs a profiling of your screen (it may even work with profiling from a different program) and you can create a LUT for both programs in a very short time with the "same" results.
That's the issue here, my 3DLUT is not from a screen profile, it's from color grading. It's unrelated to display calibration.

[The Dark Lord]

Win 10 1709 (with latest updates), MPDN 2.49 still throws out of range error, any specific KB I should install to get it to working?

[foozoor]

I guess this player is completely dead.
Maybe it should be mentioned on the first post.

[Shiandow]

How on earth did you reach that conclusion?

[foozoor]

Nice to see you again Shiandow. I always hope I'm wrong. ;)
But it seems there are many problems with the latest Windows updates and that development is still rather slow.

Could you port RAVU and maybe FSRCNN(X) for MPDN?
https://github.com/zachsaw/MPDN_Extensions/issues/239

[Shiandow]

I'm only aware of one problem with windows 10, which has something to do with a change to the .Net runtime, that one has been going on for a while, but it looks some updates may or may not stop it from happening. Also since it's not a problem in the code itself it's somewhat hard to track down.

I already commented on RAVU, and will again when there's something to comment about. The reason you haven't heard anything is because I haven't changed my mind about implementing it.

So yeah, not much has changed since last month, but MPDN didn't suddenly die either.

[sauma144]

Foozoor is a little rude sometimes. ;)
Are you going to use .net standard soon?
Will you create an alternative to NGU/NGU-AA or FSRCNN/RAVU?
Are we going to see an open source version of MPDN or is it impossible?

[Shiandow]

What exactly do you mean by 'use .net standard'? Do you mean make the extensions target .Net standard rather than .Net framework? I haven't really looked in to it, but I do plan to see if the extensions can at least be made to support C# 6.0 or higher. This depends on how well CS-script can handle C# 6.0 now.

The author of RAVU has kindly offered to help port it to MPDN, so that's likely to happen. Depending on how well this goes, other MPV scripts might be ported to MPDN as well.

The long term goal of MPDN is to become completely open source, as well as fully extendable. Unfortunately the core code for the renderer and the bridge to DirectShow is proprietary, and will need to be rewritten (and probably some other parts here and there, but those two are the most difficult).

[huhn]

The author of RAVU has kindly offered to help port it to MPDN, so that's likely to happen. Depending on how well this goes, other MPV scripts might be ported to MPDN as well.

nice to hear i can't wait to check it out together with your deband algorithm.

[foozoor]

nice to hear i can't wait to check it out together with your deband algorithm.

And a totally new D3D12 renderer and no more directshow! :D

[madshi]

And which practical benefits for end users would either of those 2 changes bring, foozoor?

[Shiandow]

I guess you could attempt to use the new DirectX renderer on an operating system that doesn't support Directshow?

[aufkrawall]

Async Compute might be able to improve performance with GCN GPUs, there is a pull request for a new Vulkan renderer of mpv which supports it.
However, there is no proof yet for actual performance benefits over DX11. I suspect it will be faster at some point than DX11 with GCN, question is just if it really will make a difference for advanced video shaders.

If I remember correctly, mpv Windows dev Rossi told me that there is an interop possible for D3D11VA -> Vulkan. So really no need for DX12, it's a completely redundant API.

[huhn]

did anyone even read what DX12 or Async Compute even does?

[aufkrawall]

There is no optimal GPU utilization without Async Compute, simple as that.
There are always utilization bubbles when the shaders process tasks serialized. Question is just how big they are.

[huhn]

and this is a video renderer.
it's pretty much about shaders. the benifit of async compute is most likely nothing at all because "only" the shaders matter no combination of other GPU parts.

[aufkrawall]

It doesn't matter that only shaders are involved for the fact that utilization isn't optimal. GPU is btw. also busy with other stuff like presentation, HW decoding etc.

[Shiandow]

I wouldn't expect any miracles from processing shaders asynchronously when 99% of the time MPDN is just rendering shaders at (hopefully) full speed, with little opportunity to do anything else in the meantime.

Unless you're telling me the GPU can process two shaders faster concurrently than it can process them sequentially, but that seems really weird.

[huhn]

yeah very unlikely it will do anything useful for video rendering.

[aufkrawall]

Unless you're telling me the GPU can process two shaders faster concurrently than it can process them sequentially, but that seems really weird.
Why not? That's also true for running tasks on the CPU with hyperthreading.
Games can profit by it the most (when they can utilize enough threads), but also other applications with constant high load like video encoding or ray-/pathtracing profit by it.

[huhn]

you can't compare a CPU multi threading with a GPU multithreading. a GPU has thousand of cores you can blindly assume they can multi thread fine.

the most profit from vulkan DX12 is the CPU which is a no issue here.

async compute is not made to utilized shaders better it is made to use shaders, rasterizer and other GPU part at the same time. while for video playback we are mostly using the shaders (correct me if i'm wrong here Shiandow).

if the shaders get better utilisation because the rasterizer are not standing in there way you can argue it is increasing shader utilisation but i don't see this happening for such a program.

[Shiandow]

Why not? That's also true for running tasks on the CPU with hyperthreading.
Games can profit by it the most (when they can utilize enough threads), but also other applications with constant high load like video encoding or ray-/pathtracing profit by it.

Why wouldn't it just run the same shader twice on the left and right halves of the image?

[aufkrawall]

Why wouldn't it just run the same shader twice on the left and right halves of the image?
Wouldn't that be quite an amount of work to adjust the shaders to work accordingly?
New Vulkan revision is not even mainlined yet, I guess more optimizations will come later.

[Shiandow]

Wouldn't that be quite an amount of work to adjust the shaders to work accordingly?
New Vulkan revision is not even mainlined yet, I guess more optimizations will come later.

No, by definition pixel shaders can be evaluated on any part of the image in any order. This is what allows GPUs to parallelize the process immensely.

In practice the GPU probably process small clusters of pixels in an orderly fashion, but the exact cluster shape and order depends on the GPU.

[TheFireRed]

I have a question: how can I leave this time code reader always turned on? Or I can only set the status bar to read this information?

https://i.imgur.com/hbUVzjm.png

[Shiandow]

I would actually recommend enabling the status bar as it wouldn't be in front of the video that way.

The alternative would be to make some small modifications to the extensions code. Nothing too difficult really, but it takes some work to set things up so you can recompile the extensions.

[mringis]

Hmm I'll investigate. I'm no longer getting the error but I did at some stage.

Have you guys tried uninstalling MPDN fully and reinstalling? Make sure you backup your settings before you do that. And try doing a windows restart after reinstalling.

Is there update about this bug?

Today i Installes it in my new computer and this problem still exit. i can't play any video.

My previous computer uses AMD GPU (RX 460) and the new one use RX 560. Is this bug only happen to new mid tier AMD GPU???


TITLE: MediaPlayerDotNet Error
------------------------------

===================================

Specified argument was out of the range of valid values. (MediaPlayerDotNet)

------------------------------
Error Type = System.ArgumentOutOfRangeException
Error Message = Specified argument was out of the range of valid values.
Error Source = MediaPlayerDotNet
Error Site = Mpdn.IScaler C(Mpdn.IScaler)
Error occurred = at a.C.C(IScaler )
at a.C.B(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.h.A()
at a.S.A()
OS Version = Microsoft Windows NT 6.3.9600.0
Total Physical Memory = 15,9 GB
Available Physical Memory = 10,1 GB
Date Time = 05/12/2017 15:56:33
User Name = ARIF-HUMAS\Arif
Application Startup Path = C:\Program Files\MPDN
Application Executable Path = C:\Program Files\MPDN\MediaPlayerDotNet.exe
CurrentDirectory = H:\HUT 60
SystemDirectory = C:\Windows\system32
Ran As Admin = False
HasShutdownStarted = False
MachineName = ARIF-HUMAS
ProcessorCount = 12
LogicalDrives = C:\, D:\, E:\, F:\, G:\, H:\, I:\, J:\, K:\, L:\, M:\, N:\, O:\
EnvironmentVariables = COMPUTERNAME = ARIF-HUMAS; USERPROFILE = C:\Users\Arif; HOMEPATH = \Users\Arif; LOCALAPPDATA = C:\Users\Arif\AppData\Local; PSModulePath = C:\Program Files\WindowsPowerShell\Modules;C:\Windows\system32\WindowsPowerShell\v1.0\Modules; PROCESSOR_ARCHITECTURE = AMD64; CommonProgramW6432 = C:\Program Files\Common Files; CommonProgramFiles(x86) = C:\Program Files (x86)\Common Files; ProgramFiles(x86) = C:\Program Files (x86); PROCESSOR_LEVEL = 23; CM2012DIR = C:\Program Files (x86)\Common Files\Autodesk Shared\Materials\; LOGONSERVER = \\ARIF-HUMAS; PATHEXT = .COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC; HOMEDRIVE = C:; SystemRoot = C:\Windows; SESSIONNAME = Console; ALLUSERSPROFILE = C:\ProgramData; PUBLIC = C:\Users\Public; FPS_BROWSER_APP_PROFILE_STRING = Internet Explorer; APPDATA = C:\Users\Arif\AppData\Roaming; Path = C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0\;C:\Program Files (x86)\Autodesk\Backburner\;C:\Program Files\Common Files\Autodesk Shared\;C:\Users\Arif\AppData\Local\Microsoft\WindowsApps;; USERNAME = Arif; OneDrive = C:\Users\Arif\OneDrive; TEMP = C:\Users\Arif\AppData\Local\Temp; CommonProgramFiles = C:\Program Files\Common Files; OS = Windows_NT; USERDOMAIN_ROAMINGPROFILE = ARIF-HUMAS; PROCESSOR_IDENTIFIER = AMD64 Family 23 Model 1 Stepping 1, AuthenticAMD; ComSpec = C:\Windows\system32\cmd.exe; USERDOMAIN = ARIF-HUMAS; SystemDrive = C:; FPS_BROWSER_USER_PROFILE_STRING = Default; ProgramFiles = C:\Program Files; ILBDIR = C:\Program Files (x86)\Common Files\Autodesk Shared\Materials\; NUMBER_OF_PROCESSORS = 12; ILMDIR = C:\Program Files (x86)\Common Files\Autodesk Shared\Materials\; TMP = C:\Users\Arif\AppData\Local\Temp; ProgramData = C:\ProgramData; ProgramW6432 = C:\Program Files; ADSK_3DSMAX_x64_2012 = C:\Program Files\Autodesk\3ds Max 2012\; PROCESSOR_REVISION = 0101; windir = C:\Windows
RuntimeVersion = 4.0.30319.42000
Thread Count = 62
Handle Count = 5030
VM Size = 9,6 GB
Peak VM Size = 9,6 GB
Working Set Size = 287,1 MB
Max Working Set Size = 1,3 MB
Min Working Set Size = 200 KB
Modules = MediaPlayerDotNet.exe, ntdll.dll, MSCOREE.DLL, KERNEL32.dll, KERNELBASE.dll, ADVAPI32.dll, msvcrt.dll, sechost.dll, RPCRT4.dll, mscoreei.dll, SHLWAPI.dll, combase.dll, ucrtbase.dll, bcryptPrimitives.dll, GDI32.dll, gdi32full.dll, msvcp_win.dll, USER32.dll, win32u.dll, IMM32.DLL, kernel.appcore.dll, VERSION.dll, clr.dll, MSVCR120_CLR0400.dll, mscorlib.ni.dll, ole32.dll, uxtheme.dll, System.ni.dll, CRYPTSP.dll, rsaenh.dll, bcrypt.dll, CRYPTBASE.dll, System.Core.ni.dll, MediaPlayerDotNet.ni.exe, System.Drawing.ni.dll, System.Windows.Forms.ni.dll, Mpdn.Core.ni.dll, System.Configuration.ni.dll, System.Xml.ni.dll, comctl32.dll, dwmapi.dll, MSCTF.dll, OLEAUT32.dll, gdiplus.dll, DWrite.dll, WindowsCodecs.dll, CommandLine.ni.dll, clrjit.dll, shell32.dll, cfgmgr32.dll, shcore.dll, windows.storage.dll, powrprof.dll, profapi.dll, YAXLib.ni.dll, System.Xml.Linq.ni.dll, clbcatq.dll, dataexchange.dll, d3d11.dll, dcomp.dll, dxgi.dll, twinapi.appcore.dll, VideoFrameServicesNative.dll, SharpDX.Direct3D9.ni.dll, SharpDX.ni.dll, d3d9.dll, aticfx64.dll, WINMM.dll, winmmbase.dll, atiu9p64.dll, atiumd64.dll, atiumd6a.dll, SharpDX.DXGI.ni.dll, TextInputFramework.dll, CoreMessaging.dll, CoreUIComponents.dll, ntmarta.dll, wintypes.dll, usermgrcli.dll, D3D9NativeServices.dll, SETUPAPI.dll, WTSAPI32.dll, AVRT.dll, dxva2.dll, opencl.dll, DEVOBJ.dll, WINSTA.dll, SharpDX.Direct3D11.ni.dll, atiuxp64.dll, atidxx64.dll, amdocl64.dll, OPENGL32.dll, GLU32.dll, atiadlxx.DLL, PSAPI.DLL, PROPSYS.dll, USERENV.dll, WINTRUST.dll, MSASN1.dll, CRYPT32.dll, atig6txx.dll, amdocl12cl64.dll, dbghelp.dll, Dx11Font.dll, FW1FontWrapper.dll, MediaInfoDotNet.ni.dll, DirectShowLib-2005.ni.dll, MediaInfo.dll, quartz.dll, CSScriptLibrary.ni.dll, WindowsBase.ni.dll, PresentationNative_v0400.dll, jscript.dll, amsi.dll, MpOav.dll, PresentationCore.ni.dll, explorerframe.dll, rasapi32.dll, rasman.dll, rtutils.dll, WS2_32.dll, mswsock.dll, winhttp.dll, LAVSplitter.ax, libbluray.dll, avutil-lav-55.dll, avformat-lav-57.dll, avcodec-lav-57.dll, Secur32.dll, avresample-lav-3.dll, SSPICLI.DLL, IPHLPAPI.DLL, NSI.dll, dhcpcsvc6.DLL, dhcpcsvc.DLL, DNSAPI.dll, qcap.dll, sxs.dll, LAVAudio.ax, WINNSI.DLL, devenum.dll, URLMon.DLL, iertutil.dll, MMDevAPI.DLL, wdmaud.drv, ksuser.dll, rasadhlp.dll, fwpuclnt.dll, AUDIOSES.DLL, msacm32.drv, MSACM32.dll, midimap.dll, DDRAW.dll, DCIMAN32.dll, DSOUND.DLL, AudioTransformFilter.ax, XySubFilter.dll, WINSPOOL.DRV, WININET.dll, LAVVideo.ax, swscale-lav-4.dll, avfilter-lav-6.dll, CallbackFilter.ax, NullRendererFilter.ax, Microsoft.ExceptionMessageBox.ni.dll, Microsoft.VisualBasic.ni.dll, diasymreader.dll


------------------------------
Program Location:

at a.C.C(IScaler )
at a.C.B(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.h.A()
at a.S.A()

[huhn]

try an older build.

[Shiandow]

Is there update about this bug?

Today i Installes it in my new computer and this problem still exit. i can't play any video.

My previous computer uses AMD GPU (RX 460) and the new one use RX 560. Is this bug only happen to new mid tier AMD GPU???


It looks like later versions of the .Net Runtime (in particular versions >= 4.7.1) might not have that bug, but this is not yet entirely certain. It's not related to the GPU by the way, it's purely .Net / Windows that's causing it.

[mringis]

It looks like later versions of the .Net Runtime (in particular versions >= 4.7.1) might not have that bug, but this is not yet entirely certain. It's not related to the GPU by the way, it's purely .Net / Windows that's causing it.

You are right. After update .Net Runtime to version 4.7.1, i don't need to downgrade MDPN.

Thank You sir :)

[Heaud]

Trying to get Streamlink to work with MPDN via the option within Streamlink to pass a URL to the player, but I keep falling into this: https://img.warywolf.net/L0ipC.png

What needs to be done so that Streamlink can work with the player?

[Shiandow]

It is apparently attempting to open your URL with the Directshow D3DY Youtube filter, which isn't installed. What kind of URL are you trying to send MPDN?

[Heaud]

I am not sure which URL Streamlink is sending to the player. The configuration file has options for fifo, streaming via HTTP, and an option to only pass the URL and let the player handle it which I am guessing is to not allow Streamlink to transcode anything. All the options give the error from my previous post when trying to load except for fifo which spits out "File not found '\\. \pipe\streamlinkpipe-5380'".

[Shiandow]

I am not sure which URL Streamlink is sending to the player. The configuration file has options for fifo, streaming via HTTP, and an option to only pass the URL and let the player handle it which I am guessing is to not allow Streamlink to transcode anything. All the options give the error from my previous post when trying to load except for fifo which spits out "File not found '\\. \pipe\streamlinkpipe-5380'".

I gave it a try and you should be able to run streamlink by using the option -a="--stdin", so your command would look something like:


streamlink -p MediaPlayerDotNet.exe -a="--stdin" <url> <quality>


for some weird reason you need to use an equals sign with "-a" rather than just a space, but it should work.

[Heaud]

Thanks for the tip! Got it working with Streamlinks default settings and using the --stdin argument specified. I also decided to toy around to see why the equals sign would be needed to feed the command an argument, and it seems that having a whitespace causes python to not read the command correctly. Giving a more basic -a--stdin was able to work just fine as well.

[dave_in_gva]

Periodically I have audio or subtitle delays that are quite large.

It's a pain to ratchet up to a delay of 1 second or 2 with the audio delay setting of +/- 5 ms, particularly on my laptop without a dedicated + or - button. Similarly, sometimes subtitle files are wildly out and although the F1 and F2 hotkeys make ratcheting up or down by 250 ms at a time much more palatable, the fact remains that sometimes it would just be nice to get a dialog box pop up where the user could enter an audio delay or a subtitle delay value.

Would it be possible Zach to have this integrated into an upcoming release? Otherwise this amazing software is pretty much perfect. :thanks:

Best wishes,

Dave M

[donger]

===================================

Określony argument jest spoza zakresu prawidłowych wartości. (MediaPlayerDotNet)

------------------------------
Error Type = System.ArgumentOutOfRangeException
Error Message = Określony argument jest spoza zakresu prawidłowych wartości.
Error Source = MediaPlayerDotNet
Error Site = Mpdn.IScaler C(Mpdn.IScaler)
Error occurred = w a.C.C(IScaler )
w a.C.B(IScaler )
w Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
w Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
w Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
w Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
w B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
w B.M.A(IMedia )
w MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
w MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
w MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
w MediaPlayerDotNet.MainForm.h.A()
w a.S.A()
OS Version = Microsoft Windows NT 6.3.9600.0
Total Physical Memory = 16 GB
Available Physical Memory = 12,2 GB
Date Time = 06.01.2018 02:55:02
User Name = DESKTOP-F812GCS\PJ
Application Startup Path = C:\Program Files\MPDN
Application Executable Path = C:\Program Files\MPDN\MediaPlayerDotNet.exe
CurrentDirectory = F:\FILMY
SystemDirectory = C:\Windows\system32
Ran As Admin = False
HasShutdownStarted = False
MachineName = DESKTOP-F812GCS
ProcessorCount = 4
LogicalDrives = C:\, D:\, F:\
EnvironmentVariables = COMPUTERNAME = DESKTOP-F812GCS; USERPROFILE = C:\Users\PJ; HOMEPATH = \Users\PJ; LOCALAPPDATA = C:\Users\PJ\AppData\Local; JD2_HOME = C:\Users\PJ\AppData\Local\JDownloader v2.0; PSModulePath = C:\Program Files\WindowsPowerShell\Modules;C:\Windows\system32\WindowsPowerShell\v1.0\Modules; PROCESSOR_ARCHITECTURE = AMD64; Path = C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0\;C:\Program Files (x86)\NVIDIA Corporation\PhysX\Common;C:\Users\PJ\AppData\Local\Microsoft\WindowsApps;; CommonProgramFiles(x86) = C:\Program Files (x86)\Common Files; ProgramFiles(x86) = C:\Program Files (x86); PROCESSOR_LEVEL = 6; LOGONSERVER = \\DESKTOP-F812GCS; PATHEXT = .COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC; HOMEDRIVE = C:; SystemRoot = C:\Windows; SESSIONNAME = Console; ALLUSERSPROFILE = C:\ProgramData; PUBLIC = C:\Users\Public; FPS_BROWSER_APP_PROFILE_STRING = Internet Explorer; APPDATA = C:\Users\PJ\AppData\Roaming; PROCESSOR_REVISION = 3c03; USERNAME = PJ; CommonProgramW6432 = C:\Program Files\Common Files; TEMP = C:\Users\PJ\AppData\Local\Temp; OneDrive = C:\Users\PJ\OneDrive; CommonProgramFiles = C:\Program Files\Common Files; OS = Windows_NT; USERDOMAIN_ROAMINGPROFILE = DESKTOP-F812GCS; PROCESSOR_IDENTIFIER = Intel64 Family 6 Model 60 Stepping 3, GenuineIntel; ComSpec = C:\Windows\system32\cmd.exe; SystemDrive = C:; FPS_BROWSER_USER_PROFILE_STRING = Default; ProgramFiles = C:\Program Files; NUMBER_OF_PROCESSORS = 4; TMP = C:\Users\PJ\AppData\Local\Temp; ProgramData = C:\ProgramData; ProgramW6432 = C:\Program Files; windir = C:\Windows; USERDOMAIN = DESKTOP-F812GCS
RuntimeVersion = 4.0.30319.42000
Thread Count = 68
Handle Count = 991
VM Size = 33,6 GB
Peak VM Size = 33,6 GB
Working Set Size = 281,2 MB
Max Working Set Size = 1,3 MB
Min Working Set Size = 200 KB
Modules = MediaPlayerDotNet.exe, ntdll.dll, MSCOREE.DLL, KERNEL32.dll, KERNELBASE.dll, ADVAPI32.dll, msvcrt.dll, sechost.dll, RPCRT4.dll, mscoreei.dll, SHLWAPI.dll, combase.dll, ucrtbase.dll, bcryptPrimitives.dll, GDI32.dll, gdi32full.dll, msvcp_win.dll, USER32.dll, win32u.dll, IMM32.DLL, kernel.appcore.dll, VERSION.dll, clr.dll, MSVCR120_CLR0400.dll, mscorlib.ni.dll, ole32.dll, uxtheme.dll, nvinject.dll, System.ni.dll, CRYPTSP.dll, rsaenh.dll, bcrypt.dll, CRYPTBASE.dll, System.Core.ni.dll, MediaPlayerDotNet.ni.exe, System.Drawing.ni.dll, System.Windows.Forms.ni.dll, Mpdn.Core.ni.dll, System.Configuration.ni.dll, System.Xml.ni.dll, shell32.dll, cfgmgr32.dll, shcore.dll, windows.storage.dll, powrprof.dll, profapi.dll, comctl32.dll, dwmapi.dll, MSCTF.dll, OLEAUT32.dll, gdiplus.dll, DWrite.dll, WindowsCodecs.dll, CommandLine.ni.dll, clrjit.dll, YAXLib.ni.dll, System.Xml.Linq.ni.dll, clbcatq.dll, dataexchange.dll, dcomp.dll, d3d11.dll, dxgi.dll, twinapi.appcore.dll, RMCLIENT.dll, VideoFrameServicesNative.dll, SharpDX.Direct3D9.ni.dll, SharpDX.ni.dll, d3d9.dll, nvldumdx.dll, crypt32.dll, MSASN1.dll, WINTRUST.DLL, imagehlp.dll, nvd3dumx.dll, WINMM.dll, winmmbase.dll, nvspcap64.dll, ntmarta.dll, SharpDX.DXGI.ni.dll, TextInputFramework.dll, CoreUIComponents.dll, CoreMessaging.dll, wintypes.dll, D3D9NativeServices.dll, SETUPAPI.dll, WTSAPI32.dll, AVRT.dll, dxva2.dll, opencl.dll, DEVOBJ.dll, WINSTA.dll, SharpDX.Direct3D11.ni.dll, nvwgf2umx.dll, nvopencl.dll, nvfatbinaryLoader.dll, WS2_32.dll, nvapi64.dll, d3dx9_43.dll, MediaInfoDotNet.ni.dll, usp10.dll, DirectShowLib-2005.ni.dll, MediaInfo.dll, quartz.dll, CSScriptLibrary.ni.dll, WindowsBase.ni.dll, PresentationNative_v0400.dll, jscript.dll, amsi.dll, USERENV.dll, MpOav.dll, WindowsFormsIntegration.ni.dll, PresentationCore.ni.dll, explorerframe.dll, rasapi32.dll, rasman.dll, rtutils.dll, mswsock.dll, winhttp.dll, LAVSplitter.ax, libbluray.dll, avcodec-lav-57.dll, avutil-lav-55.dll, avformat-lav-57.dll, avresample-lav-3.dll, Secur32.dll, SSPICLI.DLL, IPHLPAPI.DLL, NSI.dll, dhcpcsvc6.DLL, dhcpcsvc.DLL, DNSAPI.dll, WINNSI.DLL, rasadhlp.dll, fwpuclnt.dll, psapi.dll, schannel.DLL, qcap.dll, sxs.dll, LAVAudio.ax, MMDevAPI.DLL, PROPSYS.dll, wdmaud.drv, ksuser.dll, AUDIOSES.DLL, msacm32.drv, MSACM32.dll, midimap.dll, DDRAW.dll, DCIMAN32.dll, DSOUND.DLL, AudioTransformFilter.ax, XySubFilter.dll, WININET.dll, WINSPOOL.DRV, LAVVideo.ax, avfilter-lav-6.dll, swscale-lav-4.dll, CallbackFilter.ax, NullRendererFilter.ax, Microsoft.ExceptionMessageBox.ni.dll, Microsoft.VisualBasic.ni.dll, diasymreader.dll


------------------------------
Program Location:

w a.C.C(IScaler )
w a.C.B(IScaler )
w Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
w Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
w Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
w Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
w B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
w B.M.A(IMedia )
w MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
w MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
w MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
w MediaPlayerDotNet.MainForm.h.A()
w a.S.A()


Ok it's dead. Switching to MPV. New installation did nothing.

[amayra]

can you add assFilterMod (http://forum.doom9.org/showthread.php?p=1769770) Support please

[The Dark Lord]

Well the title said it all, just to report after 2018/1 win 10 update, MPDN is working again, I just seen 2 errors in extensions which I'm not sure I should post here or on github.


Superres: can't find file Downscale.hlsl
NNEDI3: Object reference not set to an instance of an object.


Cheers,

[aufkrawall]

yeah very unlikely it will do anything useful for video rendering.
I just tried it with my RX 560 and mpv-git and I can see gains, even on top of Vulkan gains in general. :rolleyes:

[Shiandow]

I just seen 2 errors in extensions which I'm not sure I should post here or on github.

Superres: can't find file Downscale.hlsl
NNEDI3: Object reference not set to an instance of an object.

Officially github would be the best place to put report extension bugs, but it sounds like the extensions weren't installed properly somehow.

[Shiandow]

I just tried it with my RX 560 and mpv-git and I can see gains, even on top of Vulkan gains in general. :rolleyes:

Out of interest, how big of a difference do those options make in your case?

[aufkrawall]

Out of interest, how big of a difference do those options make in your case?
It depends. It seems there needs to be some balance of graphics + compute to have gains with AC.
E.g. I gain at least 5% (in fps) when I use cscale=catmull_rom (pixel shader) and dscale=ewa_lanczos (compute shader) at the same time. But with cscale=ewa_lanczos as well, I don't gain anything or even lose a little (despite of deband being on, which should be pixel shader as well).

Without AC however, the gain over DX11 is already ~12% with vulkan-queue-count=>1 and vulkan-async-transfer.
With d3d11va-copy the gain even is ~30% (with scaling just set to bilinear), so Vulkan seems to be a huge win in general on AMD GPUs.

A bit unfortunate is that I need to force maximum clocks with ClockBlocker, otherwise there is stuttering with Vulkan because the driver doesn't clock up correctly.

[Heaud]

Has anyone found a good combination for upscaling DVD video? So far, I have been using SSSR 3 pass with 0.2 over sharpening, 6 locality, soft mode, and OpenCL NNEDI3. Is it also possible to run another instance of SSSR for image quadrupling?

[Shiandow]

Has anyone found a good combination for upscaling DVD video? So far, I have been using SSSR 3 pass with 0.2 over sharpening, 6 locality, soft mode, and OpenCL NNEDI3. Is it also possible to run another instance of SSSR for image quadrupling?

There's unfortunately currently no way to run SSSR twice as it always scales directly to the final size. You could make SSSR use two NNEDI3 instances though, at least, I'm fairly sure that should work.

[Heaud]

Trying out OpenCL NNEDI3 + Adaptive Sharpen above SSSR in the chain and it looks like it is working out well.

Oddly enough, regular NNEDI3 is giving me an error when placed in a conditional within a render chain. OpenCL NNEDI3 runs fine. Here is what the error states:

Object reference not set to an instance of an object.

~
Stack Trace:
at Mpdn.Extensions.RenderScripts.Shiandow.NNedi3.NNedi3.CreateFilter(TTextureFilter input)
at Mpdn.Extensions.Framework.Chain.FilterChain`1.Process(TFilter input)
at Mpdn.Extensions.Framework.Chain.Preset`2.Process(T input)
at Mpdn.Extensions.Framework.Chain.ScriptChain`2.<Process>b__0(T temp, Preset`2 chain)
at System.Linq.Enumerable.Aggregate[TSource,TAccumulate](IEnumerable`1 source, TAccumulate seed, Func`3 func)
at Mpdn.Extensions.Framework.Chain.ScriptChain`2.Process(T input)
at Mpdn.Extensions.Framework.Chain.Preset`2.Process(T input)
at Mpdn.Extensions.Framework.Chain.ScriptChain`2.<Process>b__0(T temp, Preset`2 chain)
at System.Linq.Enumerable.Aggregate[TSource,TAccumulate](IEnumerable`1 source, TAccumulate seed, Func`3 func)
at Mpdn.Extensions.Framework.Chain.ScriptChain`2.Process(T input)
at Mpdn.Extensions.Framework.Chain.FilterChainScript`2.UpdateFilter()


Is it a problem on my end?


edit: reopening the video with NNEDI3 still loaded in the conditional within a render chain does not cause an error. The error can only be replicated when loading NNEDI3 for the first time inside the conditional without opening the configuration menu for the shader.

[onemoar]

ok I am trying to get my head around the script chain syntax
this is what I am putting in the "conditional config box"

If SourceSize.Width < 1280 and NeedsUpscaling, use script chain

but when I click ok it just spits

Syntax Error

error in script ('conditional'):

Expected ')'

nothing on the gitpage or any examples I have seen use ) in the script .
what I am I missing >

[sneaker_ger]

what I am I missing >
A ")"? :)

Post your script.

[onemoar]

what script I am attempting todo this from the gui the one line I posted is what I was trying to configure

this thing is very poorly documented

there's no clear config for maximum quality or auto config for scaling scripts everything is at least 3 menus down in the ui ...

very confusing

[Heaud]

There's unfortunately currently no way to run SSSR twice as it always scales directly to the final size. You could make SSSR use two NNEDI3 instances though, at least, I'm fairly sure that should work.

Does this also apply to SuperRes? Wondering what would be the better way to sharpen a NNEDI3 upscale before feeding it down the chain to SSSR.

[Shiandow]

Does this also apply to SuperRes? Wondering what would be the better way to sharpen a NNEDI3 upscale before feeding it down the chain to SSSR.

Hi,

SuperRes and SSSR should use the same logic for upscaling, meaning that they'll always scale to the output size. I've been thinking of adding an option to not scale all the way to the output size, but I haven't gotten round to it yet.

[Heaud]

Please do, when possible. Having the option for SuperRes and SSSR to not scale to output size can allow for render chains to do image doubling with refinement more than once.

[foxyshadis]

Hate to bring this recurrent issue up again, but in setting up MPDN on my new laptop (Intel 630+Nvidia 1050), with all default settings, I get this error:
===================================

Specified argument was out of the range of valid values. (MediaPlayerDotNet)

------------------------------
Error Type = System.ArgumentOutOfRangeException
Error Message = Specified argument was out of the range of valid values.
Error Source = MediaPlayerDotNet
Error Site = Mpdn.IScaler C(Mpdn.IScaler)
Error occurred = at a.C.C(IScaler )
at a.C.B(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.y()
at B.M.f(Boolean )
at B.M.A(Action`1 )
at B.M.T()
at MediaPlayerDotNet.MainForm.i(Boolean )
at MediaPlayerDotNet.MainForm.M(Object , EventArgs )
at C.t.B()
at C.t.A(Object , FormClosedEventArgs )
at System.Windows.Forms.Form.OnFormClosed(FormClosedEventArgs e)
at A.e.OnFormClosed(FormClosedEventArgs )
at System.Windows.Forms.Form.CheckCloseDialog(Boolean closingOnly)
OS Version = Microsoft Windows NT 6.3.9600.0
Total Physical Memory = 15.9 GB
Available Physical Memory = 7.9 GB
Date Time = 2/26/2018 3:18:08 AM
User Name = DESKTOP-OB8IKT7\Foxy
Application Startup Path = C:\av\MPDN
Application Executable Path = C:\av\MPDN\MediaPlayerDotNet.exe
CurrentDirectory = D:\anime\Cross Game
SystemDirectory = C:\WINDOWS\system32
Ran As Admin = False
HasShutdownStarted = False
MachineName = DESKTOP-OB8IKT7
ProcessorCount = 8
LogicalDrives = C:\, D:\
EnvironmentVariables = COMPUTERNAME = DESKTOP-OB8IKT7; USERPROFILE = C:\Users\Foxy; HOMEPATH = \Users\Foxy; LOCALAPPDATA = C:\Users\Foxy\AppData\Local; configsetroot = C:\WINDOWS\ConfigSetRoot; PROCESSOR_ARCHITECTURE = AMD64; Path = C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0\;%SystemRoot%\system32;%SystemRoot%;%SystemRoot%\System32\Wbem;%SYSTEMROOT%\System32\WindowsPowerShell\v1.0\;C:\Program Files (x86)\NVIDIA Corporation\PhysX\Common;C:\Users\Foxy\AppData\Local\Microsoft\WindowsApps;; CommonProgramFiles(x86) = C:\Program Files (x86)\Common Files; ProgramFiles(x86) = C:\Program Files (x86); PROCESSOR_LEVEL = 6; LOGONSERVER = \\DESKTOP-OB8IKT7; PATHEXT = .COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC; HOMEDRIVE = C:; SystemRoot = C:\WINDOWS; PSModulePath = C:\Program Files\WindowsPowerShell\Modules;C:\WINDOWS\system32\WindowsPowerShell\v1.0\Modules; ALLUSERSPROFILE = C:\ProgramData; PUBLIC = C:\Users\Public; FPS_BROWSER_APP_PROFILE_STRING = Internet Explorer; APPDATA = C:\Users\Foxy\AppData\Roaming; PROCESSOR_REVISION = 9e09; USERNAME = Foxy; CommonProgramW6432 = C:\Program Files\Common Files; TEMP = C:\Users\Foxy\AppData\Local\Temp; OneDrive = C:\Users\Foxy\OneDrive; CommonProgramFiles = C:\Program Files\Common Files; OS = Windows_NT; USERDOMAIN_ROAMINGPROFILE = DESKTOP-OB8IKT7; PROCESSOR_IDENTIFIER = Intel64 Family 6 Model 158 Stepping 9, GenuineIntel; ComSpec = C:\WINDOWS\system32\cmd.exe; USERDOMAIN = DESKTOP-OB8IKT7; SystemDrive = C:; FPS_BROWSER_USER_PROFILE_STRING = Default; ProgramFiles = C:\Program Files; NUMBER_OF_PROCESSORS = 8; TMP = C:\Users\Foxy\AppData\Local\Temp; ProgramData = C:\ProgramData; ProgramW6432 = C:\Program Files; windir = C:\WINDOWS; SESSIONNAME = Console
RuntimeVersion = 4.0.30319.42000
Thread Count = 47
Handle Count = 1050
VM Size = 29.6 GB
Peak VM Size = 29.6 GB
Working Set Size = 281.5 MB
Max Working Set Size = 1.3 MB
Min Working Set Size = 200 KB
Modules = MediaPlayerDotNet.exe, ntdll.dll, MSCOREE.DLL, KERNEL32.dll, KERNELBASE.dll, ADVAPI32.dll, msvcrt.dll, sechost.dll, RPCRT4.dll, mscoreei.dll, SHLWAPI.dll, combase.dll, ucrtbase.dll, bcryptPrimitives.dll, GDI32.dll, gdi32full.dll, msvcp_win.dll, USER32.dll, win32u.dll, IMM32.DLL, kernel.appcore.dll, VERSION.dll, clr.dll, MSVCR120_CLR0400.dll, mscorlib.ni.dll, ole32.dll, uxtheme.dll, tiptsf.dll, OLEAUT32.dll, System.ni.dll, CRYPTSP.dll, rsaenh.dll, bcrypt.dll, CRYPTBASE.dll, System.Core.ni.dll, MediaPlayerDotNet.ni.exe, System.Drawing.ni.dll, System.Windows.Forms.ni.dll, Mpdn.Core.ni.dll, System.Configuration.ni.dll, System.Xml.ni.dll, shell32.dll, cfgmgr32.dll, shcore.dll, windows.storage.dll, powrprof.dll, profapi.dll, comctl32.dll, dwmapi.dll, MSCTF.dll, gdiplus.dll, DWrite.dll, WindowsCodecs.dll, CommandLine.ni.dll, clrjit.dll, YAXLib.ni.dll, System.Xml.Linq.ni.dll, clbcatq.dll, dataexchange.dll, d3d11.dll, dcomp.dll, dxgi.dll, twinapi.appcore.dll, RMCLIENT.dll, VideoFrameServicesNative.dll, SharpDX.Direct3D9.ni.dll, SharpDX.ni.dll, d3d9.dll, igdumdim64.dll, SETUPAPI.dll, SharpDX.DXGI.ni.dll, TextInputFramework.dll, CoreUIComponents.dll, CoreMessaging.dll, ntmarta.dll, wintypes.dll, D3D9NativeServices.dll, WTSAPI32.dll, WINMM.dll, AVRT.dll, dxva2.dll, winmmbase.dll, opencl.dll, DEVOBJ.dll, igc64.dll, WINSTA.dll, SharpDX.Direct3D11.ni.dll, igd10iumd64.dll, ncrypt.dll, NTASN1.dll, nvopencl.dll, nvfatbinaryLoader.dll, WS2_32.dll, WINTRUST.dll, MSASN1.dll, CRYPT32.dll, nvapi64.dll, IntelOpenCL64.dll, intelocl64.dll, OPENGL32.dll, task_executor64.dll, GLU32.dll, cpu_device64.dll, igdrcl64.dll, igdmcl64.dll, dbghelp.dll, igdfcl64.dll, Dx11Font.dll, FW1FontWrapper.dll, MediaInfoDotNet.ni.dll, DirectShowLib-2005.ni.dll, MediaInfo.dll, quartz.dll, uiautomationcore.dll, Accessibility.ni.dll, OLEACC.dll, CSScriptLibrary.ni.dll, WindowsBase.ni.dll, PresentationNative_v0400.dll, jscript.dll, amsi.dll, USERENV.dll, MpOav.dll, WindowsFormsIntegration.ni.dll, PresentationCore.ni.dll, explorerframe.dll, rasapi32.dll, rasman.dll, rtutils.dll, mswsock.dll, winhttp.dll, IPHLPAPI.DLL, NSI.dll, dhcpcsvc6.DLL, dhcpcsvc.DLL, LAVSplitter.ax, libbluray.dll, avutil-lav-55.dll, avcodec-lav-57.dll, avformat-lav-57.dll, avresample-lav-3.dll, Secur32.dll, SSPICLI.DLL, DNSAPI.dll, WINNSI.DLL, rasadhlp.dll, qcap.dll, sxs.dll, LAVAudio.ax, devenum.dll, URLMon.DLL, imagehlp.dll, iertutil.dll, MMDevAPI.DLL, PROPSYS.dll, wdmaud.drv, ksuser.dll, fwpuclnt.dll, AUDIOSES.DLL, msacm32.drv, MSACM32.dll, midimap.dll, DDRAW.dll, DCIMAN32.dll, DSOUND.DLL, AudioTransformFilter.ax, psapi.dll, schannel.DLL, XySubFilter.dll, WINSPOOL.DRV, WININET.dll, LAVVideo.ax, swscale-lav-4.dll, avfilter-lav-6.dll, CallbackFilter.ax, NullRendererFilter.ax, diasymreader.dll, d3dx9_43.dll, SharpDX.Direct3D11.Effects.ni.dll, SharpDX.D3DCompiler.ni.dll, sharpdx_direct3d11_effects_x64.dll, D3DCOMPILER_43.dll, Windows.UI.dll, mskeyprotect.dll, ncryptsslp.dll, System.ComponentModel.DataAnnotations.ni.dll, System.Runtime.Serialization.ni.dll, System.Data.ni.dll, System.Data.dll, msdmo.dll, Microsoft.ExceptionMessageBox.ni.dll, Microsoft.VisualBasic.ni.dll


------------------------------
Program Location:

at a.C.C(IScaler )
at a.C.B(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.y()
at B.M.f(Boolean )
at B.M.A(Action`1 )
at B.M.T()
at MediaPlayerDotNet.MainForm.i(Boolean )
at MediaPlayerDotNet.MainForm.M(Object , EventArgs )
at C.t.B()
at C.t.A(Object , FormClosedEventArgs )
at System.Windows.Forms.Form.OnFormClosed(FormClosedEventArgs e)
at A.e.OnFormClosed(FormClosedEventArgs )
at System.Windows.Forms.Form.CheckCloseDialog(Boolean closingOnly)

Opening MPDN on its own and removing all Bicubic avoids the error. (Unlike an earlier report, lanczos is fine.) Softcubic is fine, spline fine, etc. Only Bicubic specifically doesn't seem to work on my system, and I tried changing the presenter, the OCL, and a bunch of other settings before I isolated it to the upscaling method. Anything I can do to help troubleshoot?

[Shiandow]

Hate to bring this recurrent issue up again, but in setting up MPDN on my new laptop (Intel 630+Nvidia 1050), with all default settings, I get this error:

[...]

Opening MPDN on its own and removing all Bicubic avoids the error. (Unlike an earlier report, lanczos is fine.) Softcubic is fine, spline fine, etc. Only Bicubic specifically doesn't seem to work on my system, and I tried changing the presenter, the OCL, and a bunch of other settings before I isolated it to the upscaling method. Anything I can do to help troubleshoot?

Unfortunately the only known fixes are using windows 7 and/or updating to .Net framework 4.7+. Although it affecting Bicubic is indeed new, I hope that doesn't mean it's an entirely different but similar bug.

The annoying part is that the bug is not actually in the code itself but in the way the code is obfuscated and how this is interpreted by the .Net Framework. So assuming this is the same bug (which it does seem to be, since it's eerily similar, and also impossibly specific) there's unfortunately not much point in debugging any further. Unless someone really wants to spend time tracking down a bug in the raw MPDN bytecode / .Net Framework interpreter.

Please do, when possible. Having the option for SuperRes and SSSR to not scale to output size can allow for render chains to do image doubling with refinement more than once.

I've modified the code on my end to at least add it as a hidden option, which can be enabled by editing the settings manually (which you access by copy/pasting the script to/from a text editor from the script chain window, or by editing the settings file). This should be part of the next extensions update.

[Anime Viewer]

Hate to bring this recurrent issue up again, but in setting up MPDN on my new laptop (Intel 630+Nvidia 1050), with all default settings, I get this error:


Opening MPDN on its own and removing all Bicubic avoids the error. (Unlike an earlier report, lanczos is fine.) Softcubic is fine, spline fine, etc. Only Bicubic specifically doesn't seem to work on my system, and I tried changing the presenter, the OCL, and a bunch of other settings before I isolated it to the upscaling method. Anything I can do to help troubleshoot?

Thanks for sharing your findings. I went and opened it on its own like you mentioned (instead of opening a file to launch it), and checked my settings. Sure enough I had Bicubic as one of my settings (Chroma Upscaler). When I changed it to something else and launched files the error did not occur on my system either.

[mlauzon]

I just downloaded & installed the latest version from:

http://www.zachsaw.com/mpdn/

I added it to the NVIDIA Control Panel so it uses the 1060m instead of the Intel HD -- EDIT: removing it from the NCP still gives the error in the image -- however, trying to play any video I get the following error:

https://i.imgur.com/9rC9MKa.png

Here's the log:

===================================

Specified argument was out of the range of valid values. (MediaPlayerDotNet)

------------------------------
Error Type = System.ArgumentOutOfRangeException
Error Message = Specified argument was out of the range of valid values.
Error Source = MediaPlayerDotNet
Error Site = Mpdn.IScaler C(Mpdn.IScaler)
Error occurred = at a.C.C(IScaler )
at a.C.B(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.o(Object , EventArgs )
at System.EventHandler.Invoke(Object sender, EventArgs e)
at System.Windows.Forms.ToolStripItem.RaiseEvent(Object key, EventArgs e)
at System.Windows.Forms.ToolStripMenuItem.OnClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleMouseUp(MouseEventArgs e)
at System.Windows.Forms.ToolStrip.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.ToolStripDropDown.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.Control.WmMouseUp(Message& m, MouseButtons button, Int32 clicks)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.ToolStrip.WndProc(Message& m)
at System.Windows.Forms.ToolStripDropDown.WndProc(Message& m)
at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)
OS Version = Microsoft Windows NT 6.3.9600.0
Total Physical Memory = 15.9 GB
Available Physical Memory = 10.8 GB
Date Time = 27/04/18 18:20:31
User Name = ADMIN-PC\Admin
Application Startup Path = C:\Program Files\MPDN
Application Executable Path = C:\Program Files\MPDN\MediaPlayerDotNet.exe
CurrentDirectory = C:\Program Files\MPDN
SystemDirectory = C:\WINDOWS\system32
Ran As Admin = False
HasShutdownStarted = False
MachineName = ADMIN-PC
ProcessorCount = 8
LogicalDrives = C:\, D:\
EnvironmentVariables = COMPUTERNAME = ADMIN-PC; USERPROFILE = C:\Users\Admin; HOMEPATH = \Users\Admin; LOCALAPPDATA = C:\Users\Admin\AppData\Local; JD2_HOME = C:\Users\Admin\AppData\Local\JDownloader v2.0; PSModulePath = C:\Program Files\WindowsPowerShell\Modules;C:\WINDOWS\system32\WindowsPowerShell\v1.0\Modules; PROCESSOR_ARCHITECTURE = AMD64; Path = C:\Program Files (x86)\Intel\iCLS Client\;C:\Program Files\Intel\iCLS Client\;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0\;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0\;C:\Program Files (x86)\NVIDIA Corporation\PhysX\Common;C:\Program Files (x86)\Intel\Intel(R) Management Engine Components\DAL;C:\Program Files\Intel\Intel(R) Management Engine Components\DAL;C:\Program Files (x86)\Intel\Intel(R) Management Engine Components\IPT;C:\Program Files\Intel\Intel(R) Management Engine Components\IPT;C:\Users\Admin\AppData\Local\Microsoft\WindowsApps;; CommonProgramFiles(x86) = C:\Program Files (x86)\Common Files; ProgramFiles(x86) = C:\Program Files (x86); NVIDIAWHITELISTED = 0x01; LOGONSERVER = \\ADMIN-PC; PATHEXT = .COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC; HOMEDRIVE = C:; SystemRoot = C:\WINDOWS; SESSIONNAME = Console; ALLUSERSPROFILE = C:\ProgramData; PUBLIC = C:\Users\Public; FPS_BROWSER_APP_PROFILE_STRING = Internet Explorer; APPDATA = C:\Users\Admin\AppData\Roaming; PROCESSOR_REVISION = 5e03; USERNAME = Admin; CommonProgramW6432 = C:\Program Files\Common Files; TEMP = C:\Users\Admin\AppData\Local\Temp; OneDrive = C:\Users\Admin\OneDrive; SHIM_MCCOMPAT = 0x810000001; CommonProgramFiles = C:\Program Files\Common Files; OS = Windows_NT; USERDOMAIN_ROAMINGPROFILE = ADMIN-PC; PROCESSOR_IDENTIFIER = Intel64 Family 6 Model 94 Stepping 3, GenuineIntel; ComSpec = C:\WINDOWS\system32\cmd.exe; SystemDrive = C:; FPS_BROWSER_USER_PROFILE_STRING = Default; ProgramFiles = C:\Program Files; SynaProgDir = Synaptics\SynTP; NUMBER_OF_PROCESSORS = 8; PROCESSOR_LEVEL = 6; TMP = C:\Users\Admin\AppData\Local\Temp; ProgramData = C:\ProgramData; ProgramW6432 = C:\Program Files; windir = C:\WINDOWS; USERDOMAIN = ADMIN-PC
RuntimeVersion = 4.0.30319.42000
Thread Count = 62
Handle Count = 1363
VM Size = 33.7 GB
Peak VM Size = 34.4 GB
Working Set Size = 283.1 MB
Max Working Set Size = 1.3 MB
Min Working Set Size = 200 KB
Modules = MediaPlayerDotNet.exe, ntdll.dll, MSCOREE.DLL, KERNEL32.dll, KERNELBASE.dll, ADVAPI32.dll, msvcrt.dll, sechost.dll, RPCRT4.dll, mscoreei.dll, SHLWAPI.dll, combase.dll, ucrtbase.dll, bcryptPrimitives.dll, GDI32.dll, gdi32full.dll, msvcp_win.dll, USER32.dll, win32u.dll, IMM32.DLL, kernel.appcore.dll, VERSION.dll, clr.dll, MSVCR120_CLR0400.dll, mscorlib.ni.dll, ole32.dll, uxtheme.dll, AWSoundCenterDevProps.dll, SHELL32.dll, cfgmgr32.dll, shcore.dll, windows.storage.dll, powrprof.dll, profapi.dll, WS2_32.dll, AWSoundCenterOSD.dll, System.ni.dll, CRYPTSP.dll, rsaenh.dll, bcrypt.dll, CRYPTBASE.dll, System.Core.ni.dll, MediaPlayerDotNet.ni.exe, System.Drawing.ni.dll, System.Windows.Forms.ni.dll, Mpdn.Core.ni.dll, System.Configuration.ni.dll, System.Xml.ni.dll, comctl32.dll, dwmapi.dll, MSCTF.dll, OLEAUT32.dll, gdiplus.dll, DWrite.dll, WindowsCodecs.dll, CommandLine.ni.dll, clrjit.dll, YAXLib.ni.dll, System.Xml.Linq.ni.dll, clbcatq.dll, dataexchange.dll, d3d11.dll, dcomp.dll, dxgi.dll, twinapi.appcore.dll, RMCLIENT.dll, VideoFrameServicesNative.dll, SharpDX.Direct3D9.ni.dll, SharpDX.ni.dll, d3d9.dll, igd9dxva64.dll, crypt32.dll, MSASN1.dll, WINTRUST.DLL, imagehlp.dll, MMDevApi.dll, DEVOBJ.dll, PROPSYS.dll, AUDIOSES.DLL, wintypes.dll, AVRT.dll, SharpDX.DXGI.ni.dll, TextInputFramework.dll, CoreUIComponents.dll, CoreMessaging.dll, ntmarta.dll, D3D9NativeServices.dll, SETUPAPI.dll, WTSAPI32.dll, dxva2.dll, WINMM.dll, winmmbase.dll, opencl.dll, nvldumdx.dll, nvd3dumx.dll, nvSCPAPI64.dll, nvspcap64.dll, gpapi.dll, cryptnet.dll, WINSTA.dll, SharpDX.Direct3D11.ni.dll, nvwgf2umx.dll, nvopencl.dll, nvfatbinaryLoader.dll, nvapi64.dll, IntelOpenCL64.dll, intelocl64.dll, OPENGL32.dll, task_executor64.dll, GLU32.dll, cpu_device64.dll, d3dx9_43.dll, MediaInfoDotNet.ni.dll, usp10.dll, DirectShowLib-2005.ni.dll, MediaInfo.dll, quartz.dll, CSScriptLibrary.ni.dll, WindowsBase.ni.dll, PresentationNative_v0400.dll, jscript.dll, amsi.dll, USERENV.dll, MpOav.dll, WindowsFormsIntegration.ni.dll, PresentationCore.ni.dll, explorerframe.dll, rasapi32.dll, rasman.dll, rtutils.dll, mswsock.dll, winhttp.dll, IPHLPAPI.DLL, NSI.dll, dhcpcsvc6.DLL, dhcpcsvc.DLL, DNSAPI.dll, WINNSI.DLL, rasadhlp.dll, fwpuclnt.dll, psapi.dll, secur32.dll, SSPICLI.DLL, schannel.DLL, diasymreader.dll, comdlg32.dll, tiptsf.dll, edputil.dll, DUI70.dll, DUser.dll, thumbcache.dll, policymanager.dll, msvcp110_win.dll, Windows.Globalization.dll, Bcp47Langs.dll, globinputhost.dll, xmllite.dll, OLEACC.dll, uiautomationcore.dll, sxs.dll, atlthunk.dll, StructuredQuery.dll, CLDAPI.dll, FLTLIB.DLL, AEPIC.dll, Windows.StateRepositoryPS.dll, Windows.Storage.Search.dll, IconCodecService.dll, ActXPrxy.dll, cscapi.dll, OneCoreUAPCommonProxyStub.dll, twinapi.dll, apphelp.dll, FileSyncShell64.dll, WININET.dll, EhStorShell.dll, cscui.dll, samcli.dll, SAMLIB.dll, netutils.dll, LINKINFO.dll, IcarosPropertyHandler.dll, PhotoMetadataHandler.dll, zipfldr.dll, MPR.dll, drprov.dll, ntlanman.dll, davclnt.dll, DAVHLPR.dll, wkscli.dll, NetworkExplorer.dll, LAVSplitter.ax, libbluray.dll, avformat-lav-57.dll, avutil-lav-55.dll, avcodec-lav-57.dll, avresample-lav-3.dll, qcap.dll, LAVAudio.ax, devenum.dll, URLMon.DLL, iertutil.dll, wdmaud.drv, ksuser.dll, msacm32.drv, MSACM32.dll, midimap.dll, DDRAW.dll, DCIMAN32.dll, DSOUND.DLL, AudioTransformFilter.ax, XySubFilter.dll, WINSPOOL.DRV, LAVVideo.ax, swscale-lav-4.dll, avfilter-lav-6.dll, CallbackFilter.ax, NullRendererFilter.ax, Microsoft.ExceptionMessageBox.ni.dll, Microsoft.VisualBasic.ni.dll


------------------------------
Program Location:

at a.C.C(IScaler )
at a.C.B(IScaler )
at Mpdn.D3D9VideoRenderer.VideoRenderer.ag()
at Mpdn.D3D9VideoRenderer.VideoRenderer.aG()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.o(Object , EventArgs )
at System.EventHandler.Invoke(Object sender, EventArgs e)
at System.Windows.Forms.ToolStripItem.RaiseEvent(Object key, EventArgs e)
at System.Windows.Forms.ToolStripMenuItem.OnClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleMouseUp(MouseEventArgs e)
at System.Windows.Forms.ToolStrip.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.ToolStripDropDown.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.Control.WmMouseUp(Message& m, MouseButtons button, Int32 clicks)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.ToolStrip.WndProc(Message& m)
at System.Windows.Forms.ToolStripDropDown.WndProc(Message& m)
at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)

I haven't used MPDN in about 2yrs, but this never happend before when I used it on Win7 x64, I am currently on Win10 x64.

[foxyshadis]

I just downloaded & installed the latest version from:

http://www.zachsaw.com/mpdn/

I added it to the NVIDIA Control Panel so it uses the 1060m instead of the Intel HD -- EDIT: removing it from the NCP still gives the error in the image -- however, trying to play any video I get the following error:


Here's the log:



I haven't used MPDN in about 2yrs, but this never happend before when I used it on Win7 x64, I am currently on Win10 x64.

Open MPDN on its own, and in the scaler options change Bicubic to pretty much anything else. It's a known issue.

[mlauzon]

Open MPDN on its own, and in the scaler options change Bicubic to pretty much anything else. It's a known issue.

Well that didn't work, new error:

https://i.imgur.com/tUZQ9Zj.png

===================================

HRESULT: [0x80070057], Module: [General], ApiCode: [E_INVALIDARG/Invalid Arguments], Message: The parameter is incorrect.
(SharpDX)

------------------------------
Error Type = SharpDX.SharpDXException
Error Message = HRESULT: [0x80070057], Module: [General], ApiCode: [E_INVALIDARG/Invalid Arguments], Message: The parameter is incorrect.

Error Source = SharpDX
Error Site = Void CheckError()
Error occurred = at SharpDX.Result.CheckError()
at SharpDX.Direct3D9.Device.CreateTexture(Int32 width, Int32 height, Int32 levels, Int32 usage, Format format, Pool pool, Texture textureOut, IntPtr sharedHandleRef)
at A.v..ctor(Device , c )
at c.m.A(c[] )
at Mpdn.D3D9VideoRenderer.VideoRenderer.am()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.h.A()
at a.S.A()
OS Version = Microsoft Windows NT 6.3.9600.0
Total Physical Memory = 15.9 GB
Available Physical Memory = 12.3 GB
Date Time = 5/3/2018 7:44:09 AM
User Name = DESKTOP-5V2IM7T\Admin
Application Startup Path = C:\Program Files\MPDN
Application Executable Path = C:\Program Files\MPDN\MediaPlayerDotNet.exe
CurrentDirectory = C:\Windows\system32
SystemDirectory = C:\Windows\system32
Ran As Admin = False
HasShutdownStarted = False
MachineName = DESKTOP-5V2IM7T
ProcessorCount = 8
LogicalDrives = C:\, D:\
EnvironmentVariables = COMPUTERNAME = DESKTOP-5V2IM7T; USERPROFILE = C:\Users\Admin; HOMEPATH = \Users\Admin; LOCALAPPDATA = C:\Users\Admin\AppData\Local; JD2_HOME = C:\Users\Admin\AppData\Local\JDownloader v2.0; PSModulePath = C:\Program Files\WindowsPowerShell\Modules;C:\Windows\system32\WindowsPowerShell\v1.0\Modules; PROCESSOR_ARCHITECTURE = AMD64; Path = C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0\;C:\Windows\System32\OpenSSH\;C:\Program Files (x86)\NVIDIA Corporation\PhysX\Common;C:\Users\Admin\AppData\Local\Microsoft\WindowsApps;; CommonProgramFiles(x86) = C:\Program Files (x86)\Common Files; ProgramFiles(x86) = C:\Program Files (x86); NVIDIAWHITELISTED = 0x01; LOGONSERVER = \\DESKTOP-5V2IM7T; PATHEXT = .COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC; HOMEDRIVE = C:; SystemRoot = C:\Windows; ProgramData = C:\ProgramData; SESSIONNAME = Console; ALLUSERSPROFILE = C:\ProgramData; DriverData = C:\Windows\System32\Drivers\DriverData; APPDATA = C:\Users\Admin\AppData\Roaming; PROCESSOR_REVISION = 5e03; USERNAME = Admin; OneDrive = C:\Users\Admin\OneDrive; SHIM_MCCOMPAT = 0x810000001; CommonProgramFiles = C:\Program Files\Common Files; OS = Windows_NT; USERDOMAIN_ROAMINGPROFILE = DESKTOP-5V2IM7T; PROCESSOR_IDENTIFIER = Intel64 Family 6 Model 94 Stepping 3, GenuineIntel; ComSpec = C:\Windows\system32\cmd.exe; SystemDrive = C:; TEMP = C:\Users\Admin\AppData\Local\Temp; ProgramFiles = C:\Program Files; SynaProgDir = Synaptics\SynTP; NUMBER_OF_PROCESSORS = 8; PROCESSOR_LEVEL = 6; TMP = C:\Users\Admin\AppData\Local\Temp; CommonProgramW6432 = C:\Program Files\Common Files; ProgramW6432 = C:\Program Files; windir = C:\Windows; USERDOMAIN = DESKTOP-5V2IM7T; PUBLIC = C:\Users\Public
RuntimeVersion = 4.0.30319.42000
Thread Count = 65
Handle Count = 1056
VM Size = 37.8 GB
Peak VM Size = 37.8 GB
Working Set Size = 281.9 MB
Max Working Set Size = 1.3 MB
Min Working Set Size = 200 KB
Modules = MediaPlayerDotNet.exe, ntdll.dll, MSCOREE.DLL, KERNEL32.dll, KERNELBASE.dll, ADVAPI32.dll, msvcrt.dll, sechost.dll, RPCRT4.dll, mscoreei.dll, SHLWAPI.dll, combase.dll, ucrtbase.dll, bcryptPrimitives.dll, GDI32.dll, gdi32full.dll, msvcp_win.dll, USER32.dll, win32u.dll, IMM32.DLL, kernel.appcore.dll, VERSION.dll, clr.dll, MSVCR120_CLR0400.dll, mscorlib.ni.dll, ole32.dll, uxtheme.dll, AWSoundCenterDevProps.dll, SHELL32.dll, cfgmgr32.dll, shcore.dll, windows.storage.dll, profapi.dll, powrprof.dll, FLTLIB.DLL, WS2_32.dll, AWSoundCenterOSD.dll, clrjit.dll, System.ni.dll, System.Drawing.ni.dll, System.Windows.Forms.ni.dll, CRYPTSP.dll, rsaenh.dll, bcrypt.dll, CRYPTBASE.dll, System.Core.ni.dll, System.Configuration.ni.dll, System.Xml.ni.dll, comctl32.dll, dwmapi.dll, MSCTF.dll, OLEAUT32.dll, gdiplus.dll, DWrite.dll, WindowsCodecs.dll, System.Xml.Linq.ni.dll, clbcatq.dll, dataexchange.dll, d3d11.dll, dcomp.dll, dxgi.dll, twinapi.appcore.dll, RMCLIENT.dll, VideoFrameServicesNative.dll, d3d9.dll, igdumdim64.dll, SETUPAPI.dll, TextInputFramework.dll, CoreUIComponents.dll, CoreMessaging.dll, ntmarta.dll, wintypes.dll, D3D9NativeServices.dll, WINMM.dll, dxva2.dll, WTSAPI32.dll, AVRT.dll, winmmbase.dll, opencl.dll, MMDevApi.dll, DEVOBJ.dll, PROPSYS.dll, AUDIOSES.DLL, crypt32.dll, MSASN1.dll, WINTRUST.DLL, imagehlp.dll, nvldumdx.dll, nvd3dumx.dll, nvSCPAPI64.dll, PSAPI.DLL, nvspcap64.dll, igc64.dll, WINSTA.dll, nvwgf2umx.dll, IntelOpenCL64.dll, intelocl64.dll, task_executor64.dll, OPENGL32.dll, GLU32.dll, cpu_device64.dll, igdrcl64.dll, igdfcl64.dll, igdmcl64.dll, dbghelp.dll, nvopencl.dll, nvfatbinaryLoader.dll, nvapi64.dll, MediaInfo.dll, d3dx9_43.dll, usp10.dll, quartz.dll, WindowsBase.ni.dll, PresentationNative_v0400.dll, jscript.dll, amsi.dll, USERENV.dll, MpOav.dll, PresentationCore.ni.dll, explorerframe.dll, rasapi32.dll, rasman.dll, rtutils.dll, mswsock.dll, winhttp.dll, IPHLPAPI.DLL, NSI.dll, dhcpcsvc6.DLL, dhcpcsvc.DLL, DNSAPI.dll, WINNSI.DLL, rasadhlp.dll, LAVSplitter.ax, libbluray.dll, avutil-lav-55.dll, avcodec-lav-57.dll, avformat-lav-57.dll, avresample-lav-3.dll, Secur32.dll, SSPICLI.DLL, qcap.dll, sxs.dll, LAVAudio.ax, devenum.dll, fwpuclnt.dll, URLMon.DLL, iertutil.dll, wdmaud.drv, ksuser.dll, msacm32.drv, MSACM32.dll, midimap.dll, DDRAW.dll, DCIMAN32.dll, DSOUND.DLL, AudioTransformFilter.ax, schannel.DLL, XySubFilter.dll, WINSPOOL.DRV, WININET.dll, LAVVideo.ax, avfilter-lav-6.dll, swscale-lav-4.dll, CallbackFilter.ax, NullRendererFilter.ax, diasymreader.dll, Microsoft.VisualBasic.ni.dll, System.Management.ni.dll


------------------------------
Program Location:

at SharpDX.Result.CheckError()
at SharpDX.Direct3D9.Device.CreateTexture(Int32 width, Int32 height, Int32 levels, Int32 usage, Format format, Pool pool, Texture textureOut, IntPtr sharedHandleRef)
at A.v..ctor(Device , c )
at c.m.A(c[] )
at Mpdn.D3D9VideoRenderer.VideoRenderer.am()
at Mpdn.D3D9VideoRenderer.VideoRenderer.e(Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(String , FrameBufferInputFormat , Int32 , Int32 , Int32 , Int32 , InterlaceFlags , Double )
at B.M.A(String , FrameBufferInputFormat , Int32 , Int32 , YuvColorimetric , InterlaceFlags , PointF )
at B.M.A(IMedia )
at MediaPlayerDotNet.MainForm.A(IMedia , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(IMedia , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.OpenMedia(String , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.h.A()
at a.S.A()


I think the log also screwed up as well.

[foxyshadis]

I've been trying to get MPDN to work at all since updating to 1803, and I think it's just not going to happen. That obfuscation layer has obviously completely broken the program in newer Windows. Sucks, this was my favorite player. :(

[Greenhorn]

If anyone reading this would still like to use MPDN on the latest Win10 update, the 32bit and AnyCPU versions seem to work perfectly.

[MysteryX]

Is it possible to use MPDN as a component within a .NET program? Or any plan on adding this support?

[Shiandow]

Using MPDN as a component inside another .Net program is probably not possible at the moment. That said, I have done some work in trying to separate MPDN into its components. The end goal is to have a fully open source and modular version of MPDN, but progress is a bit slow, and some rather important parts aren't allowed to be made open source hence the obfuscation.

I'm still hoping to find a nice option for the video-decoding part. Ideally something that doesn't necessarily rely on Directshow.

[madshi]

JFMI, why open sourcing some parts but not others?

Don't get me wrong, obviously I understand very well why a developer would not want to open source his solution (after all almost all my stuff is closed source). I'm just not sure what the reason would be for open sourcing some parts of a project like this, but not others?

[huhn]

Zachs said for "legal reason" from the country he lives in AFAIK.
i can't find the post right now.

[Anime Viewer]

I'm receiving a render script error as soon as I try to launch the player now.

"Operation is not valid due to the current state of the object. (Mpdn.RenderScript)"

Is there a way to start the program in a "safe" type mode to get into it with out it running RenderScript, or somewhere I can remove the setting that is using renderscript when starting the player?

I was using the x64 version, after reading about the AnyCPU version working after the Windows Patch. I was able to use it once, but then the next time I tried to use it I encountered this error.

[Shiandow]

JFMI, why open sourcing some parts but not others?

Don't get me wrong, obviously I understand very well why a developer would not want to open source his solution (after all almost all my stuff is closed source). I'm just not sure what the reason would be for open sourcing some parts of a project like this, but not others?

Unfortunately the current implementation contains code that is not allowed to be made open source. Which is annoying in C# where a 'compiled' application is barely different from its source code.

Is there a way to start the program in a "safe" type mode to get into it with out it running RenderScript, or somewhere I can remove the setting that is using renderscript when starting the player?

In options>debugging you can enable the debug dialog, that will give more information on what went wrong. If you empty the extensions folder you should at least be able to reach the options menu. If not something else has gone wrong, which is entirely possible with the way Microsoft is apparently messing about with the .Net framework.

[Anime Viewer]

In options>debugging you can enable the debug dialog, that will give more information on what went wrong. If you empty the extensions folder you should at least be able to reach the options menu. If not something else has gone wrong, which is entirely possible with the way Microsoft is apparently messing about with the .Net framework.

First I tried copying everything in the extension folder out of it, to a temp folder elsewhere. I ran MDPN and still got the error, so I then went to where I had copied the extension folder files, ran the uninstall from there, then deleted everything that was left behind in the folder, and finally tried running MPDN again. Still encountered the same error, and everything grayed out and no way to get into options after clicking "ignore" after the error message.

Guess I'm stuck waiting on a Microsoft/.Net fix that might never come unless anyone has any other ideas.

[mlauzon]

If anyone reading this would still like to use MPDN on the latest Win10 update, the 32bit and AnyCPU versions seem to work perfectly.

Where do I get the LAV filters for MPDN, as it won't let me watch anything without them..?!

[huhn]

here you go: https://forum.doom9.org/showthread.php?t=156191

[mlauzon]

here you go: https://forum.doom9.org/showthread.php?t=156191

Do I get the installer version, as I am trying the AnyCPU version of MPDN that I quoted..?!

[huhn]

sorry no clue i can't get this working.

[foxyshadis]

Using MPDN as a component inside another .Net program is probably not possible at the moment. That said, I have done some work in trying to separate MPDN into its components. The end goal is to have a fully open source and modular version of MPDN, but progress is a bit slow, and some rather important parts aren't allowed to be made open source hence the obfuscation.

I'm still hoping to find a nice option for the video-decoding part. Ideally something that doesn't necessarily rely on Directshow.

I know most obfuscators let you work on a method, class, module, or whole program level; I wonder if reliability could be increased by applying it only to the proprietary parts that can't be open-sourced anyway. The UI is relatively boilerplate compared to the image processing algorithms.

[Anime Viewer]

Looks like I was able to resolve my issue with the "Operation is not valid due to the current state of the object. (Mpdn.RenderScript)" error.

First I installed both the versions (x64 and x86) of MPDN that comes with an Installer (and Uninstaller), and then the Extensions that come with an Installer (and Uninstaller). Then I ran the uninstaller for the extensions first (deleting all settings), and then ran the uninstaller for MPDN (both versions) and selected the check boxes to remove extenisons and configurations. Next I installed MPDN (x64) from the installer with reset settings checked. Finally I installed the Extensions from the installer and checked for a clean install.

That got it where I could get into the options area and play videos again.

I found that if I change presenation API to anything other than Direct3D 9Ex (Direct3D 10.1 or Direct3D 11) then I would get the RenderScript error the next time I tried to open MPDN, and would be completely locked out again where I'd have to resort to uninstalling and reinstalling as detailed above to be able to get into MPDN again.

I also found that if I had full screen exclusive mode checked/enabled that the video would freeze (but audio still played) when I went to full screen.

I didn't have any problems prior to the latest Windows update using Direct3D 10 or 11, nor full screen exclusive mode, so it looks like whatever they implemented in their update broke both of those features.

I also have to uncheck Improve Chroma Reconstruction (which is checked by the default install settings) otherwise black, purple, indego, etc don't display correctly, but I don't believe that to be related to any of the other issues and it doesn't lead to an inability to play videos like the Direct3D and Full Screen Exclusive settings do now.

Now I just need to remember what settings (upscalers, downscalers, render scripts, etc) looked and worked best on my system...


Edit:
Looks like SuperRes doesn't work now.

[hooddy]

Hi there. I try to run ANYCPU build at WinRT/ARM32. There is errors in it. After run it shows error (01.png) and menus dont clickable. I attach errors logs n screenshots. Can I hope ya add support for WinRT? There is an private VLS ARM32 workin well at WinRT. But I have no idea why VLC huys wont public it.

[louiebh]

is this project under water?

[hajj_3]

is this project under water?
The developer of this hasn't made any posts in 1yr so it looks like it is dead.

[Shiandow]

Hi everyone,

I figured I should give some clarity on the current state of MPDN.

In short, at this point nobody is actively developing MPDN so for lack of a better word it's dead.

In more detail, Zachs has stopped developing MPDN for a few years now, and it seems unlikely that this will change any time soon. Slightly before he stopped completely he did pull me in as one of the developers of MPDN, so development hadn't yet completely stopped at that time. Unfortunately this was also the time we found out that MPDN misbehaves on new versions of windows, and eventually it became clear that the likely cause was the obfuscation software used to disguise the proprietary parts of the MPDN code. And being the only developer at the time I wasn't really prepared to find a (likely temporary) workaround that I'd have to maintain by myself. And that's basically where we are now.

However, for the optimists among you, while this does mean that MPDN is effectively dead, I'm still looking for a way to ensure its extensions can somehow continue to exist (maybe without some of the cumbersome, rarely used features), I have also put some effort into extracting the MPDN UI into a separate MPDN extension, with some success. Unfortunately this still leaves a gap that needs to be filled to turn this into a full functioning application. The main gap left is the I/O, i.e. the parts responsible for opening video files and outputting the processed video/audio. Those have no replacement at this point. The extension API is also currently a tad too complicated to easily integrate the MPDN extensions in other applications, and either way such an API is unlikely to catch on without at least some proof of concept which again means handling the I/O somehow. While I have some ideas, I'm not really seeing an easy way to realise those by myself any time soon.

Anyway thanks everyone for supporting and using MPDN, and until we meet again.

Cheers,
Shiandow

[BetA13]

Thank you for the Clarification :)
Its sad, but its how it is i guess..

Still using the Player since its one of the very few players with an advanced option set to play around with..
And the Quality is also very nice..

Thank you for all the work you guys have done..

until we meet again :) hehe

greetz

[Zachs]

Seeing as I have moved on from that old company for a long time, perhaps there's no need to obfuscate the code any more. :)
I've dropped my domain name though - haven't renewed it and it expired, so the download links are all dead.

[amayra]

Seeing as I have moved on from that old company for a long time, perhaps there's no need to obfuscate the code any more. :)
I've dropped my domain name though - haven't renewed it and it expired, so the download links are all dead.

so there hope for us to make this open source project or share source private at least ?

[Zachs]

so there hope for us to make this open source project or share source private at least ?

Sharing the source privately is definitely ok if anyone wants to continue the project. To make it fully open source might be a bit risky.

For now let's get a raise of hands to see if anyone is interested in continuing the project. By that I mean developing and maintaining it.

Shiandow was trying to rewrite parts of it to make it open source which may be a better idea seeing as the original code interacting with directshow wasn't the best anyway. A bit of refactoring and it's practically a brand new code. My feeling is there's no need to do a full rewrite.

[Zachs]

I've rebuilt all the binaries with nearly no obfuscation and updated the download links - please let me know if it's still not working.

[huhn]

well looks like that went very well for me.
this version works on win 10 older version didn't work at all.

so good job fixing a major.

and xy subfilter isn't by any chance hard code?

[Zachs]

and xy subfilter isn't by any chance hard code?


The support is hardcoded - anyone wants to have a go at making it non-hardcoded?

[huhn]

it's fine i guess it just we now have 2 subtitle renderer that use the same interface that xy subfilter uses.

[Zachs]

Not hard to add them in. Might be a good idea to make them configurable too for future subtitle filters. No time to work on it though but more than happy to add some developers to the repository.

[Zachs]

Can you let me know what the subtitle filters are? I'll try to add them if I have some time.

[Zachs]

Hi there. I try to run ANYCPU build at WinRT/ARM32. There is errors in it. After run it shows error (01.png) and menus dont clickable. I attach errors logs n screenshots. Can I hope ya add support for WinRT? There is an private VLS ARM32 workin well at WinRT. But I have no idea why VLC huys wont public it.

ANYCPU build still needs the native parts which are currently only built for x86 and x64. ARM on windows is very new and I'm not even sure if it's worth supporting seeing as the filters are probably not going to work too. Perhaps running the x86 build would work via emulation?

[Aleksoid1978]

Sharing the source privately is definitely ok if anyone wants to continue the project.

Maybe you can share sourcein private - it's very interesting and can use in other project.

[Zachs]

Maybe you can share sourcein private - it's very interesting and can use in other project.

I can share the source but only for the purpose of continuing development of the project. You can't use it in any other projects because of proprietary code.

[Aleksoid1978]

I can share the source but only for the purpose of continuing development of the project. You can't use it in any other projects because of proprietary code.

But - how you can trust ? :)

[Zachs]

But - how you can trust ? :)

MPDN extension devs can definitely be trusted.

I just don't want to get myself or anyone else on trouble legally.

[amayra]

Can you let me know what the subtitle filters are? I'll try to add them if I have some time.

how about AssFilterMod (https://forum.doom9.org/showthread.php?t=173573)

[amayra]

MPDN extension devs can definitely be trusted.

I just don't want to get myself or anyone else on trouble legally.

well i'm about to ask if you can share it with me but for my only use without redistributed or release it in public

[Zachs]

how about AssFilterMod (https://forum.doom9.org/showthread.php?t=173573)

Is there another one or just this?

[huhn]

these are the 2 known ones that use the same interface.
xy subfilter(vobsub) and assfiltermod(libass)

[Zachs]

Gotta love em asses! Lol!

So just the one ass to support then since MPDN already does xy?

[huhn]

yes assfilter mod is pretty much an libass implementation of xy subfilter and not rarely massively faster. it can render subtitles at UHD where xy barely can do it at 1080. your CPU plays a huge roll too.

[BetA13]

is there a way or an posibillity to watch HDR content? Like MPCBE does? I mean on non HDR monitors.
greetz

[DeadlyEmbrace]

Sharing the source privately is definitely ok if anyone wants to continue the project. To make it fully open source might be a bit risky.

For now let's get a raise of hands to see if anyone is interested in continuing the project. By that I mean developing and maintaining it.

Shiandow was trying to rewrite parts of it to make it open source which may be a better idea seeing as the original code interacting with directshow wasn't the best anyway. A bit of refactoring and it's practically a brand new code. My feeling is there's no need to do a full rewrite.

I'd be interested in seeing MPDN continue. Not sure if I have the time/skill to be a full time maintainer but more than willing to contribute some of my time and even resources (can probably offer hosting space for binaries on my server and domain) to keep things going.
Would be a real shame if MPDN were to die.

Also, a bit early but happy new year to everyone!

[Zachs]

I've moved the source to Bitbucket. Can you please PM your email address to me please?

And happy new year everyone!!!

[Zachs]

Shiandow, I'll need to add you to the new repo on Bitbucket too. You may have to create a branch and cherry pick your commits.

[patul]

I'd be interested in seeing MPDN continue. Not sure if I have the time/skill to be a full time maintainer but more than willing to contribute some of my time and even resources (can probably offer hosting space for binaries on my server and domain) to keep things going.
Would be a real shame if MPDN were to die.

Also, a bit early but happy new year to everyone!

+1. It would be real shame if this wonderful player is no longer developed. Personally, I've been using MPDN since very beginning, and have been following this thread since its early days, hence I would like to see some capable hands will take up and continue the development.

Good to see we have a volunteer already.

Happy new year!

[Shiandow]

Happy new year to everyone!

This sure is a nice way to kick off the new year. Should people be interested in developing it's probably good to know I have managed to extract the UI from the rest of the player logic, that should make it easier to at least make something that looks like MPDN :) but it should also help make some sense of the part that remains.

If anyone's interested in adding HDR support I also have a few pixel shaders that add at least some rudimentary support for HDR images, but they need some tinkering to make sure they get the right parameters from the video (unless you're willing to adjust them manually on the fly, but that's probably just me...)

[Zachs]

Happy new year Shiandow! Can you PM me your Bitbucket email address? I've moved the repo there because it's free for up to 5 devs.

[Zachs]

I'm pretty sure everyone would love to have HDR support! :)

[huhn]

If anyone's interested in adding HDR support I also have a few pixel shaders that add at least some rudimentary support for HDR images, but they need some tinkering to make sure they get the right parameters from the video (unless you're willing to adjust them manually on the fly, but that's probably just me...)

how about running an synthetic created 3D LUT with HDR-SDR conversation at it for know.

the current 3D LUt is broken and very easy to fix (full range -> limited range, 3D LUT, limited range full range conversation).

it's a waste in processing power but what can you do 3d LUT are limited range end of story.

it even kind of works already.
running an 3D LUt with a limited to full range conversation after it looks similar(on a fast view exactly the same to me) to me compared to madVR running with the same HDR->SDR 3D LUT.

so take the best you can get and displaycal developer clearly know what they are doing.

[Zachs]

Last 2 spots available to any potential MPDN Devs.

[Zachs]

If anyone's interested in adding HDR support I also have a few pixel shaders that add at least some rudimentary support for HDR images, but they need some tinkering to make sure they get the right parameters from the video (unless you're willing to adjust them manually on the fly, but that's probably just me...)

Don't we already provide the render scripts with parameters from the video?

[Shiandow]

Don't we already provide the render scripts with parameters from the video?

Don't think we provide any HDR paramaters yet.

It would also be nice if we could figure out a way to make the video parameters extendable, rather than hard coded into the application, but I'm getting ahead of myself.

[Zachs]

That's not entirely impossible. A string dictionary would do the job. The thing is I'm not sure where that HDR info is coming from.

[Shiandow]

A string dictionary would work, but it would be nicer if you could somehow request information by C# interface.

[Zachs]

Any idea where that information comes from? We can decide how to pass that info to the scripts once we know where to get it from.

The thing with interfaces is they must be defined at compile time. I was thinking of something that would allow us to grab all info regardless of how useful they are and dump them into a dictionary. It's the scripts that will decide how useful the info is.

[Zachs]

Or do you mean COM interface? The metadata usually comes from it.

[Shiandow]

No I meant something similar to:



IHDRParameters hdr_params = video.info as IHDRParameters;



although in practice you might end up with something closer to:



IHDRParameters hdr_params = video.requestInfo<IHDRParameters>();



which some other extension providing video information that implements IHDRParameters.

Sure the interfaces need to be known at compile time, but that's no different from needing to know the parameter names that the information would otherwise be in.

I might be overcomplicating things though, it's just that this would prevent a lot of guesswork when reading the parameters.

[nevcairiel]

The thing is I'm not sure where that HDR info is coming from.

From here:
https://github.com/Nevcairiel/LAVFilters/blob/master/developer_info/IMediaSideData.h

Note that if you provide a custom allocator (which renderers typically do), you'll need to make sure the IMediaSamples it allocates actually implement the IMediaSideData interface.

[Zachs]

Cheers Nev

[Zachs]

Nev, do we need to copy and cache pData in STDMETHOD(SetSideData)(GUID guidType, const BYTE *pData, size_t size) PURE;
or can we simply hold the pointer and size to pData and expect pData to be valid throughout the lifetime of the IMediaSample?

[nevcairiel]

You need to copy and save it. The caller of SetSideData has no clue about the ultimate lifetime of the object, since its reference counted.

[Zachs]

Thanks Nev.

[Zachs]

We can't seem to get assfiltermod to call ISubRenderConsumer.Connect. Xysubfilter on the other hand has no problems doing so. :(

Nevermind. Found the problem after reading the source code of assfiltermod - it's expecting the consumer to implement ISubRenderConsumer2. IMHO it should be checking for ISubRenderConsumer interface, not ISubRenderConsumer2.

[huhn]

interesting it just worked out of the box with madVR. madshi didn't even know this renderer was a thing.
that'S the spec if i'm not mistaken:
http://madshi.net/SubRenderIntf.h

[Zachs]

The documentation isn't ambiguous but it was misinterpreted by assfiltermod's author.

MPDN didn't need the Clear method so did not implement the ISubRenderConsumer2 interface. Assfiltermod however only looks for that interface, when the Connect method actually belongs to the ISubRenderConsumer interface.

[Metal-HTPC]

This is by far the greatest player I ever tested. Never got MPC-HC or BE to work properly with madVR.
MPDN has all the usefull settings already integrated and works very stable. The playback is the best possible in my setup.

The only thing I'm missing is a way to play my DVD's. Is there an addon somewhere which makes VIDEO_TS folder playback including menus etc possible ?

I archived my whole music DVD collection in the regular dvd format (VTS/VOB files) and most of those shows will most likely never see a Blu-ray release.

thanks in adavnce for your help
regards

[Shiandow]

Not sure if the VTS files will work since DVD menus are not supported (and will probably remain unsupported for the near future). I'm pretty sure that the LAV filters should be able to handle the VOB files though, so those should work.

[avih]

Hi everyone,
... Both madVR and MPV use portions of code from MPDN Extensions.


Would you mind pointing out some MPDN extensions which are used in mpv (and the mpv code which uses things from MPDN)? I'm somewhat familiar with mpv code, and I don't recall anything which came from MPDN...

[huhn]

here you go: https://gist.github.com/igv
these are the easy ones even i can find.

[avih]

here you go: https://gist.github.com/igv
these are the easy ones even i can find.

But this is not mpv using anything from MPDN. These are shaders which presumably can work in mpv when loaded as external shaders.

It's like I'll write some JS code which can run in Firefox and then advertise/claim that Firefox is using my code...

[huhn]

aren't or at least wasn't they shipped with mpv builds?

[avih]

aren't or at least wasn't they shipped with mpv builds?

Surely not with the official windows distribution at https://mpv.srsfckn.biz/ .

If you or some other third party wants to distribute an mpv build with a collection of shaders from MPDN, you can do that, but it cannot be considered part of mpv itself...

[huhn]

if i'm not mistaken the shader nnedi3 version is written by someone from MPDN a lot of different version for different ways to call openCL nnedi3 where also written by zach again if i remember correctly. it's been a couple of years...

https://github.com/mpv-player/mpv/pull/2555/files

and it's not hard to find evidence as you cna see: * A HLSL port further modified by madshi, Shiandow and Zach Saw could be
* found at (also LGPLv3 licensed):
* https://github.com/zachsaw/MPDN_Extensions
*

and the main post of MPDN is from 2014.

beware i'm not a programmer but MPV has changed over the years and removing shaders from the project that are used fairly as open source is in my novice eyes a degradation of MPV.

[avih]

A nnedi3 port (by bjin iirc) was in mpv for few months and removed some years ago.

[edit]
nnedi3 is, AFAICT, not MPDN code. It was ported to various systems, I guess to MPDN too, and various forms are floating around (avisynth, shader forms, etc). It's possible that the MPDN authors have some lines of code which ended up in mpv, like hundreds of other people do too.

Personally, I don't think the specific nnedi3 subject has enough merit to advertise that "Both madVR and MPV use portions of code from MPDN Extensions", and even if it was, it was only the case for few months, and it's not been the case for years.

To advertise something like that, IMHO, it needs something quite a bit more substantial.

[edit2]
Actually, I'm pretty sure at least with nnedi3, mpv never contained any MPDN code. To quote (https://github.com/mpv-player/mpv/issues/2230#issuecomment-161190654) bjin who wrote the code which landed (and soon later removed) in mpv:

To clarify, the reimplementation is done from scratch without taking any existing code as reference

[huhn]

@zachsaw Thanks for pointing this out. To clarify, the reimplementation is done from scratch without taking any existing code as reference. It's still unclear to me wether this could be considered as derivative work or not.

I will open another PR to add all credits.

adding the full quote makes it look quite different and the fact they are properly named in the shader.

https://github.com/mpv-player/mpv/issues/2752

What is the difference? adaptive sharpen already has been ported and super-res is the same shader from MPDN Extensions.

and that's how it should be.

[avih]

Can you please stop acting like a child? If there's MPDN code in mpv, just show me where it is in mpv.

[huhn]

excuses me?

[Zachs]

http://madshi.net/clown.png
https://imgur.com/download/CDKoAYw
https://imgur.com/download/PfaOev4
https://imgur.com/download/1jmlV63
https://imgur.com/download/0SVl9wx
https://imgur.com/download/Vy90l8M
https://imgur.com/download/3yT2Sym
https://imgur.com/download/b6N5jiE
https://imgur.com/download/Tqs0AdK
https://imgur.com/download/yQwNpb9
https://imgur.com/download/IUrJe5c
https://imgur.com/download/FSRFots
https://imgur.com/download/tlc2cpV
https://imgur.com/download/jzGNF3N

[Zachs]

Any opinions about the quality of the scaler above?

[madshi]

Looks good to me, most probably a neural network? Of course a big question will be speed. You may want to work on reducing ringing artifacts, though.

[huhn]

in the skyscraper image the the skyscraper in the top right has some patter how do patterns like this look in motion?

do i see this correctly or is the ringing very similar to super xbr 150? https://abload.de/img/tqs0adk-imgur.png_snagajqp.png

can you create an x4 image from the castle to see if this get's out of hand: https://abload.de/img/sclossr3j7e.png
i first watched this image on my HTPC and is was not properly possible to judge this image in UHD so i unscaled it using the windows picture app and it got totally out of hand at this part.
on a 1080 screen watched at native resolution it only looks odd.

[Zachs]

I haven't tested it in motion yet but the building image is probably the worst I could find. Wanted to post a mixture of good and bad ones.

What was I supposed to be looking out for in the castle photo? I haven't trained it to scale 4x so it's unreasonable to pile 2x scaler twice and judge it's quality. It's not a scaler where you can pile.

Not sure about speed yet but yes it's NN based. I guess the true test will be when I write the OpenCL code and test it on my 970 GTX.

[Zachs]

Looking closely at the ringing pattern I don't think it resembles the one you posted at all? Not sure if I'm looking at the same thing though.

[huhn]

Looking closely at the ringing pattern I don't think it resembles the one you posted at all? Not sure if I'm looking at the same thing though.

the two black bars ion the top right skyscraper are for me the part with the biggest ringing.

What was I supposed to be looking out for in the castle photo? I haven't trained it to scale 4x so it's unreasonable to pile 2x scaler twice and judge it's quality. It's not a scaler where you can pile.

but what else are you going todo if you have to scale the image further use bicubic?

i'm talking about the white vertical lines in the shadow which are very very sharp if it is scaled further it looked like it is cut there.
i make you an UHD scaled image later which get's cropped for image size reasons.

Not sure about speed yet but yes it's NN based. I guess the true test will be when I write the OpenCL code and test it on my 970 GTX.

so an advanced language like openCL is needed. i'm getting some nnedi3 flashback here and openCL clearly didn't help there in term of functionality.

[Zachs]

No. I'd be training 3x and 4x separately with different models after the 2x one is done. This is the minimum quality version that is meant to work in real time on a decent GPU. But I can't be certain until I actually write the code and test it out. OpenCL has all the functionalities I need to implement it. Just requires some grunt work and time which is very scarce for me right now.

Is the ringing acceptable though? Is it a good compromise between ringing and sharpness and details?

[huhn]

edit: this is maybe not a fair comparison because maybe one is bicubic 60 the other one bicubic 60 AR i look into it sorry.
this is the castle upscaled wiht bicubic to UHD so i have to say the ringing should be fixed:
https://abload.de/img/bicubicupscale49kna.png
here is NGU sharp x2 and bicubic to UHD cropped:
https://abload.de/img/ngubicubicupscale62jq6.png

the castle image is most likely unfair madshi really likes to use it and so NGU will just do great here. of course just an adjudicated guess.

here is one of the vertical "cut" i was talking about too it's the same bicubic upscaled image just with a marker:
https://abload.de/img/cutb1kni.png

here is the sky scraper with NGU as a comparison too:
https://abload.de/img/ngusharpuaj13.png

while your algorithm is generally sharper and that by a lot on the images i took a closer look but NGU doesn't ring.

i personally avoid image scaler that can look odd from time to time even when they are an AVG far better that's why i don't use NGU sharp. so just to be fair.

[Zachs]

Well like I said it shouldn't be upscaled with bicubic to UHD for comparison because you wouldn't stack it with anything else. If anything, you'd use the 4x upscale directly. This is purely a 2x scaler. At 2x, though, is the ringing acceptable?
It also seems to have better detail than NGU.
3x and 4x will be more expensive computationally, and the exercise here is to find out what's the cheapest way to get 2x with high enough quality.
There will be another piece for 3x and 4x respectively.

[huhn]

is pretty hard to see details on UHD if the image 920x600 but i would say yes the ringing is still obvious.

would integar NN scaling for comparison be fair?

while i can see that 2x 3x and 4x scaling fix the issue for additional scalers for resolution like 1080p 720p on an UHD screen but what are you doing about DVDs?

your implementation doesn't look as "flat" as NGU how much of this is from ringing?

[Zachs]

Ringing can contribute to some perceived sharpness so it's not all bad but inspecting the images closely reveals that NGU does remove quite a bit of details, which is likely why it appears "flat".

[huhn]

there is a reason that in my personal opinion NGU sharp should be used with the add grain feature but this is not about NGU here.

what so ever i see the ringing as a major problem in your scaler implementation.

can you maybe add an 1080p image upscaled to UHD?

[hooddy]

ANYCPU build still needs the native parts which are currently only built for x86 and x64. ARM on windows is very new and I'm not even sure if it's worth supporting seeing as the filters are probably not going to work too. Perhaps running the x86 build would work via emulation?
Ashley I'm not sure about Intel x86 emulation but I'm pretty sure WinRT devices is weak to it. There is exist port of VLC for WinRT/32 ARM (not ARM64 pubic available) but its very private and carefully hided from usual ppls (i really dunno whats the reason for this strange behavior). I'm just for populate WinOnARM/32 and freeware code for it. I read later that some part of your player uses specific cpu archs. Its sad tat your palyer wont run on ARM32. The same story with PAin.NET the lead coder wont port it for ARM32 (but yes there is port of older version tat works well on WinRT).

[hooddy]

Maybe some guys that gets sources compile ARM32 build with ANYCPU options in Visual Studio?

[Zachs]

Not that simple. There's a lot of ASM code that won't be compatible.

[trandoanhung1991]

Hi, I've searched around for the cause of this issue but so far I haven't seen a fix.

MPDN crashes when not using DX9 rendering mode:


===================================

Operation is not valid due to the current state of the object. (Mpdn.RenderScript)

------------------------------
Error Type = System.InvalidOperationException
Error Message = Operation is not valid due to the current state of the object.
Error Source = Mpdn.RenderScript
Error Site = Void SetRenderer(Mpdn.IRenderer)
Error occurred = at Mpdn.RenderScript.Renderer.SetRenderer(IRenderer renderer)
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(IWin32Window , IWin32Window , FrameComposerType , Boolean , Boolean , Boolean , Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer..ctor(IWin32Window , IWin32Window , FrameComposerType , Int32 , Boolean , Boolean , Boolean , Boolean )
at B.M.A(Form , FrameComposerType , Int32 , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.K()
at MediaPlayerDotNet.MainForm.g()
at MediaPlayerDotNet.MainForm.m(Object , EventArgs )
at System.Windows.Forms.Form.OnLoad(EventArgs e)
at C.R.OnLoad(EventArgs )
at System.Windows.Forms.Form.OnCreateControl()
at System.Windows.Forms.Control.CreateControl(Boolean fIgnoreVisible)
at System.Windows.Forms.Control.CreateControl()
at System.Windows.Forms.Control.WmShowWindow(Message& m)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.Form.WmShowWindow(Message& m)
at C.R.WndProc(Message& )
at MediaPlayerDotNet.MainForm.WndProc(Message& )
at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)
OS Version = Microsoft Windows NT 6.3.9600.0
Total Physical Memory = 7.9 GB
Available Physical Memory = 4.2 GB
Date Time = 2/18/2019 4:22:20 PM
User Name = FX504GM-PAWNAGE\Hung
Application Startup Path = C:\Program Files\MPDN
Application Executable Path = C:\Program Files\MPDN\MediaPlayerDotNet.exe
CurrentDirectory = C:\Program Files\MPDN
SystemDirectory = C:\Windows\system32
Ran As Admin = False
HasShutdownStarted = False
MachineName = FX504GM-PAWNAGE
ProcessorCount = 12
LogicalDrives = C:\, D:\
EnvironmentVariables = COMPUTERNAME = FX504GM-PAWNAGE; USERPROFILE = C:\Users\Hung; HOMEPATH = \Users\Hung; LOCALAPPDATA = C:\Users\Hung\AppData\Local; PSModulePath = C:\Program Files (x86)\WindowsPowerShell\Modules;C:\Windows\system32\WindowsPowerShell\v1.0\Modules;C:\Program Files (x86)\AutoIt3\AutoItX; PROCESSOR_ARCHITECTURE = AMD64; CommonProgramW6432 = C:\Program Files\Common Files; CommonProgramFiles(x86) = C:\Program Files (x86)\Common Files; ProgramFiles(x86) = C:\Program Files (x86); PROCESSOR_LEVEL = 6; LOGONSERVER = \\FX504GM-PAWNAGE; PATHEXT = .COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC; HOMEDRIVE = C:; SystemRoot = C:\Windows; SESSIONNAME = Console; ALLUSERSPROFILE = C:\ProgramData; DriverData = C:\Windows\System32\Drivers\DriverData; APPDATA = C:\Users\Hung\AppData\Roaming; Path = C:\Program Files (x86)\Intel\Intel(R) Management Engine Components\iCLS\;C:\Program Files\Intel\Intel(R) Management Engine Components\iCLS\;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0\;C:\Windows\System32\OpenSSH\;C:\Program Files (x86)\Intel\Intel(R) Management Engine Components\DAL;C:\Program Files\Intel\Intel(R) Management Engine Components\DAL;C:\Program Files (x86)\Intel\Intel(R) Management Engine Components\IPT;C:\Program Files\Intel\Intel(R) Management Engine Components\IPT;C:\Program Files\Intel\WiFi\bin\;C:\Program Files\Common Files\Intel\WirelessCommon\;C:\Program Files (x86)\NVIDIA Corporation\PhysX\Common;C:\Program Files\NVIDIA Corporation\NVIDIA NvDLISR;C:\Users\Hung\AppData\Local\Microsoft\WindowsApps;; USERNAME = Hung; OneDrive = C:\Users\Hung\OneDrive; CommonProgramFiles = C:\Program Files\Common Files; OS = Windows_NT; USERDOMAIN_ROAMINGPROFILE = FX504GM-PAWNAGE; PROCESSOR_IDENTIFIER = Intel64 Family 6 Model 158 Stepping 10, GenuineIntel; OneDriveConsumer = C:\Users\Hung\OneDrive; USERDOMAIN = FX504GM-PAWNAGE; SystemDrive = C:; TEMP = C:\Users\Hung\AppData\Local\Temp; ProgramFiles = C:\Program Files; NUMBER_OF_PROCESSORS = 12; ComSpec = C:\Windows\system32\cmd.exe; TMP = C:\Users\Hung\AppData\Local\Temp; ProgramData = C:\ProgramData; ProgramW6432 = C:\Program Files; windir = C:\Windows; PROCESSOR_REVISION = 9e0a; PUBLIC = C:\Users\Public
RuntimeVersion = 4.0.30319.42000
Thread Count = 42
Handle Count = 680
VM Size = 5.2 GB
Peak VM Size = 5.2 GB
Working Set Size = 107.3 MB
Max Working Set Size = 1.3 MB
Min Working Set Size = 200 KB
Modules = MediaPlayerDotNet.exe, ntdll.dll, MSCOREE.DLL, KERNEL32.dll, KERNELBASE.dll, ADVAPI32.dll, msvcrt.dll, sechost.dll, RPCRT4.dll, mscoreei.dll, SHLWAPI.dll, combase.dll, ucrtbase.dll, bcryptPrimitives.dll, GDI32.dll, gdi32full.dll, msvcp_win.dll, USER32.dll, win32u.dll, IMM32.DLL, kernel.appcore.dll, VERSION.dll, clr.dll, MSVCR120_CLR0400.dll, mscorlib.ni.dll, ole32.dll, uxtheme.dll, RTSSHooks64.dll, WINMM.dll, winmmbase.dll, cfgmgr32.dll, System.ni.dll, CRYPTSP.dll, rsaenh.dll, bcrypt.dll, CRYPTBASE.dll, System.Core.ni.dll, MediaPlayerDotNet.ni.exe, System.Drawing.ni.dll, System.Windows.Forms.ni.dll, Mpdn.Core.ni.dll, System.Configuration.ni.dll, System.Xml.ni.dll, shell32.dll, shcore.dll, windows.storage.dll, profapi.dll, powrprof.dll, FLTLIB.DLL, comctl32.dll, dwmapi.dll, MSCTF.dll, OLEAUT32.dll, gdiplus.dll, DWrite.dll, WindowsCodecs.dll, CommandLine.ni.dll, clrjit.dll, YAXLib.ni.dll, System.Xml.Linq.ni.dll, clbcatq.dll, dataexchange.dll, d3d11.dll, dcomp.dll, dxgi.dll, twinapi.appcore.dll, RMCLIENT.dll, VideoFrameServicesNative.dll, SharpDX.Direct3D9.ni.dll, SharpDX.ni.dll, d3d9.dll, igd9dxva64.dll, crypt32.dll, MSASN1.dll, WINTRUST.DLL, imagehlp.dll, SharpDX.DXGI.ni.dll, Mpdn.RenderScript.ni.dll, TextInputFramework.dll, CoreUIComponents.dll, CoreMessaging.dll, ntmarta.dll, wintypes.dll, D3D9NativeServices.dll, SETUPAPI.dll, AVRT.dll, WTSAPI32.dll, dxva2.dll, opencl.dll, DEVOBJ.dll, nvldumdx.dll, nvd3dumx.dll, nvspcap64.dll, dbghelp.dll, dbgcore.DLL, gpapi.dll, cryptnet.dll, igdumdim64.dll, igc64.dll, WINSTA.dll, SharpDX.Direct3D11.ni.dll, nvwgf2umx.dll, IntelOpenCL64.dll, intelocl64.dll, task_executor64.dll, cpu_device64.dll, igdrclneo64.dll, WS2_32.dll, igdfcl64.dll, Microsoft.ExceptionMessageBox.ni.dll, Microsoft.VisualBasic.ni.dll, diasymreader.dll, SspiCli.dll, psapi.dll


------------------------------
Program Location:

at Mpdn.RenderScript.Renderer.SetRenderer(IRenderer renderer)
at Mpdn.D3D9VideoRenderer.VideoRenderer.A(IWin32Window , IWin32Window , FrameComposerType , Boolean , Boolean , Boolean , Boolean )
at Mpdn.D3D9VideoRenderer.VideoRenderer..ctor(IWin32Window , IWin32Window , FrameComposerType , Int32 , Boolean , Boolean , Boolean , Boolean )
at B.M.A(Form , FrameComposerType , Int32 , Boolean , Boolean , Boolean , Boolean )
at MediaPlayerDotNet.MainForm.K()
at MediaPlayerDotNet.MainForm.g()
at MediaPlayerDotNet.MainForm.m(Object , EventArgs )
at System.Windows.Forms.Form.OnLoad(EventArgs e)
at C.R.OnLoad(EventArgs )
at System.Windows.Forms.Form.OnCreateControl()
at System.Windows.Forms.Control.CreateControl(Boolean fIgnoreVisible)
at System.Windows.Forms.Control.CreateControl()
at System.Windows.Forms.Control.WmShowWindow(Message& m)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.Form.WmShowWindow(Message& m)
at C.R.WndProc(Message& )
at MediaPlayerDotNet.MainForm.WndProc(Message& )
at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)



Any idea where I should begin troubleshooting?

Also, currently DX9 crashes upon playback.

[hooddy]

Not that simple. There's a lot of ASM code that won't be compatible.

Well... Got it. Thanks anyway, it's valuable.

[Zachs]

Ringing seems to be nearly non existent now

https://imgur.com/download/nPsPeCm
https://i5.yapics.com/02/19/5319Qp6a02.png

[madshi]

Hard to judge with all those JPEG artifacts! Ringing appears to be much better, but it's also much softer now, overall.

[Zachs]

Yeah didn't realise Imgur saved it as JPG. I'll re-upload later.
I was under the impression users didn't like it too sharp?

[nevcairiel]

Personally I would say that sharpness is good as long as it doesn't come at the expense of visible/distracting artifacts - which is often a problem with previously compressed source material, as source-based artifacts result in post-scaling artifacts as well.

[madshi]

I was under the impression users didn't like it too sharp?
There's no clear consensus among users. Many like it sharp, others like it a bit softer. It can also depend on the scaling factor. E.g. at 4x going very sharp can result in cartoon like images (very sharp edges but no texture detail), while at 2x that's less of a problem. And it can depend on the content, as well. If the content is extremely soft, using an algorithm which rings a bit may not hurt at all.

Among those users who like it sharp, there are some who don't like ringing, and there are some that are ok with some ringing if it results in much sharper results.

[Zachs]

Lol. Looks like it's best to let them decide then. I might do a few and ultimately let the user decides which is best for them.

[Zachs]

Updated the upsampled image with PNG one

[patul]

I compared it to the result of waifu2x, the difference is visible, and indeed your result is bit softer. I know waifu2x was intended for artwork, however for this particular image (castle), it gives surprisingly pleasant result (Noise Reduction: none, Style: photo, Upscaling: 2x).

[Zachs]

Waifu2x is impossibly slow on even modern GPUs. It wouldn't make sense to implement it in a video player.

[madshi]

It's definitely much better in terms of ringing now. Of course it got a bit softer, as a result, but that's the trade you probably have to make when trying to avoid ringing. On the positive side: A clean image can be post-sharpened.

[Zachs]

Part of the problem of the original post was the downscaled image. It had a fair bit of ringing to it so in way it was preserving all the details including the ringing noise. I had to take a high res image and downscaled it myself to produce a cleaner one. The original implementation did have a bit more ringing but the perceived sharpness was largely due to ringing in the downscaled image.

[huhn]

while your original small image had ringing it was heavily amplified by your scaling algorithm.

it looks like this new low ringing image can be scaled without any major issues that's a huge improvement.

[Zachs]

More images to test ringing.

https://i5.yapics.com/02/20/53202XuD02.png
https://i5.yapics.com/02/20/5320GyL002.png
https://i5.yapics.com/02/20/5320Vy3Z02.png
https://i4.yapics.com/02/20/4320yjPZ02.png

[huhn]

did you think about using FP16 which may be accurate enough and is natively supported on new nvidia GPUs and AMD vega?

i know you wanted to test it on a 970 but if this FP16 works as advertised you should get quite a performance boost.

[Aurelian]

Hi. I'm a new member and have been using MPDN for a while with LAV filters. Unfortunately, I can't seem to get it to work when I use the DirectX 11 presentation API. It just crashes and won't start again.

Specs:

Dell Inspiron 7559
Win 10 x64 Home Edition
Nvidia GTX 960M
MPDN ver 2.49.0.3606

Would updating to a newer version solve said issue, or is there another workaround?
Either way, where could I get the latest updated builds? I can't seem to find where the newly united team are posting builds, unless I missed something.
Sorry for the rambling post and thanks everyone for your help.

[Zachs]

did you think about using FP16 which may be accurate enough and is natively supported on new nvidia GPUs and AMD vega?

i know you wanted to test it on a 970 but if this FP16 works as advertised you should get quite a performance boost.

I don't have a newer card to test with. But from what I found, OpenCL on NVIDIA doesn't do fp16. But feel free to correct me.

[huhn]

no they just made a lot of wind with FP16 for tensor and general compute for CUDA.

so much potential wasted by simply not caring for openCL.

this type of code will run terrible on none turing/volta cards anyway.

27 Tflops hidden behind missing openCL support what a waste.

[nevcairiel]

NVIDIA never cared about OpenCL, and they are certainly not going to start now, especially since OpenCL is basically deprecated in favor of Vulkan Compute Shaders - which support FP16 on NVIDIA.
Honestly, any new code in OpenCL is probably a mistake, and you should look into Vulkan instead, which actually has a pretty wide support base on Windows already.

[Shiandow]

PS 3.0 it is then.

[Zachs]

Finally managed to optimise it enough to run at 13ms to do a 2x upscale of DVD res. 1080p to 4k is out for GTX 970 at 67ms but any recent cards like GTX 1070 should do just fine.

Big thanks to Shiandow for his valuable insights!
Now we just need to port it to MPDN renderscript.

[Zachs]

Created an extreme quality version for 3x upscale (most likely will require 2080ti and above to do 720p to 4k uspcale).

https://i2.yapics.com/03/12/2312hNDo03.png
https://i4.yapics.com/03/12/4312GhAk03.png

Normal quality version will run real time without a problem for for 720p to 4k upscale even on my old 970.

[foozoor]

No normal quality screen to compare?

[Zachs]

Normal quality is not nearly as good but for other images it's not as pronounced as this one. This one has lots of high frequency components.

https://i4.yapics.com/03/12/4312GijS03.png

[ryrynz]

The first thing that stands out to me with this upscale is the strength of the darkening of the lines creating an almost cartoonish appearance, can that be reigned in a little?

[Zachs]

That said however, it still fared very well against other state of the art scalers.

[Zachs]

The first thing that stands out to me with this upscale is the strength of the darkening of the lines creating an almost cartoonish appearance, can that be reigned in a little?

That's how the original image looks like. You can artificially blur it though but that would drop SSIM significantly vs ground truth.

[Zachs]

Finally managed to optimise it enough to run at 13ms to do a 2x upscale of DVD res. 1080p to 4k is out for GTX 970 at 67ms but any recent cards like GTX 1070 should do just fine.

Big thanks to Shiandow for his valuable insights!
Now we just need to port it to MPDN renderscript.

After many rounds of optimisations, 1080p to 2160p only needs 18ms on my GTX 970!

720p to 1440p is 9ms.

3x from 720p to 2160p is 35ms.

[Zachs]

3x upscale sample (720p to 2160p):

Original: http://s000.tinyupload.com/index.php?file_id=71145458854064098138
3x: http://s000.tinyupload.com/index.php?file_id=07887740499290546766

[huhn]

any ETA when user will get there and on it?

it's easier to find issues with the scaler this way and if proper way to reproduce are given to you you maybe able to improve on it.

[Zachs]

This will be the baseline model. Any improvements will be looked at after we get this one done.

[Anima123]

Zachs, how's it going on with incorporating your new scaling into MPDN? I'm asking since it's been a while, and am still looking forward to it.

[mlauzon]

https://i.imgur.com/Tqs0AdK.png
https://i.imgur.com/yQwNpb9.png

I know it's completely off topic, but that's 55 King St. W. in Toronto...wonder if anyone else from Toronto noticed that.

[Zachs]

Was too busy with work lately to work on it. Will try to continue when I get some spare time. It just needs to be written as a renderscript.

[louiebh]

whats the latest version and where can we download it because your personal website is down!
im happy this project isnt folded up yet!


Could i also get help with the Reclock script? mine is basically saying "Audioscript could not initialize" and the in brackets (AMD/Intel OpenCL drivers unavailable)?

[Zachs]

I've hosted it on AWS S3. The latest is on the OP. I've had no time to maintain a website so I've shut it down.

You'll need to read OP to setup OpenCL drivers so you can use audio scripts.

[dave_in_gva]

Zach,

Can you unpack some of your acronyms above? What is the URL for downloading the latest installer? I'm afraid it is not clear to me what you mean by:

"I've hosted it on AWS S3. The latest is on the OP."

I've been using MPDN for 3+ years and love it. My laptop died and I am setting things up on a new one and have been getting error messages with my original installer.

Thankyou if you can steer me to the latest build.

Best wishes,

Dave M

[huhn]

https://s3.amazonaws.com/mediaplayerdotnet-releases/Latest/Installers/MediaPlayerDotNet_x64_2_49_0_3606_Installer.exe

you can find all links and information in the first post also known as a original post (OP).

[dave_in_gva]

Thanks Huhn - I had seen AWS meant Amazon and so have uninstalled my prior version as it threw an error and reinstalled from AWS.

Thing is I still am getting the following error on launching MPDN. This is on my x64-based Surface Pro 4/i5:

mscorlib Error
1 is not a supported code page.
Parameter name: codepage (mscorlib)

MPDN comes up but all menus are greyed out and I can't operate the program.

This has been the same error I've been getting with installation of packages from earlier builds. In the most recent installation I removed the prior installation (including the MPDN configuration files) and reinstalled from the AWS package.

Any ideas? I would love to get back to using MPDN and it has never given me troubles before. Only thing to add is that in reinstalling things on this new laptop I have installed the most recent version (4.8) of Microsoft Net Framework.

[dave_in_gva]

Sorry, forgot to add one more behaviour. After launching MPDN a second error message comes up, saying:

MediaPlayerDotNet Error
Object reference not set to an instance of an object. (MediaPlayerDotNet)

Thanks again for any help on this.

[huhn]

sorry i had major problem getting this to work since windows 10 so not much i can help with that.

[Sephiroth_FF]

So glad to see my favorite player is going to get new features!

I do know there was a couple of questions about HDR a few pages ago but, most of it was about HDR to SDR conversion and I was wandering if there is any plan to also support native HDR monitors too? In Windows 10, I can enable "Play HDR games and apps" in "Windows HD Colors settings" and VLC will output HDR video files natively to my monitor. Though, it would maybe require MPDN to have a DirectX 12 renderer?

Speaking of DX12, I have seen some of the amazing new scaler samples and since DX12 has support for explicit multi-gpu, could MPDN use multiple gpus for the rendering and shave off even more more milliseconds or even nearly halve the compute time? My computer has two 1080Ti in SLI and a 980Ti for PhysX and I would love if MPDN could use it all!

Finally, would be awesome to have Dolby Atmos and DTS:X object-based audio output to the new Windows Spatial Sound API like Windows Sonic, etc.

Sorry for making so many requests but MPDN has always been the most cutting edge player for me and I am super happy development might pick up again and a big thank to all who worked on this amazing player!

[louiebh]

hope this project is still going, sadly i still couldnt find a fix for th openCL audio scripts couldnt be initialised error i tried installing latest opencl drivers for intel and latest nvidia drivers and the reclockaudio script still doesnt work lost for options :(

in the op doesnt really direct someone on how to get it working too tried a bit of googling but its not a popular error

[stax76]

It still does not start on UHD 288 DPI.

[ktsbkt]

Is there a chance you will implement screenshot output of various jpeg quality ranges like MPC-HC?

I would like to recommend this player to my friends, but probably won't be able to do so since we like to share screenshots and discuss them etc.

[Kein]

Render Scripts are quite similar to AviSynth scripts and plugins except they run fully on the GPU. You can script Render Scripts in a similar fashion to AviSynth with a C# syntax, or an AviSynth-like syntax with the full power of JavaScript behind it. As opposed to madVR's mostly closed source nature, MPDN's Render Scripts are fully open source and are hosted on GitHub.

I have a question. AviSynth can process YUV420 "natively" in some cases, without any color conversion, for example, some fade-out effect or subtitles I can add without color conversion, but big guys like Vegas or Adobe work in RGB space for all of their effects to be available so the raw recorded in YUV420 video gets converted to RGB regardless with some data loss.

How is NP.NET's Render Script in this regard? I guess essentially I'm asking if I can use it as a replacement to AviSynth to some extent.

[foxyshadis]

I have a question. AviSynth can process YUV420 "natively" in some cases, without any color conversion, for example, some fade-out effect or subtitles I can add without color conversion, but big guys like Vegas or Adobe work in RGB space for all of their effects to be available so the raw recorded in YUV420 video gets converted to RGB regardless with some data loss.

How is NP.NET's Render Script in this regard? I guess essentially I'm asking if I can use it as a replacement to AviSynth to some extent.

Renderscript runs in your choice of colorspace. You can leave it as YUV or convert it to RGB, but it's ALWAYS in float with a minimum of 16-bit. You have to go out of your way to truncate it to 8-bit before going back to a 16-bit processing chain.

[ashlar42]

Has development for this stopped?

[ryrynz]

For the time being.

[amayra]

i wish if dev share source code at least in private I like to see it

[Anima123]

It would really be nice if the developers can share there new integer scaler with other programs like incorporating it into mpc video renderer.

The quality seems nice based on the previous presented pictures, it would be really appreciated if they can be applied in another way.

[Shiandow]

i wish if dev share source code at least in private I like to see it

You'd probably have to contact Zachs for that, the current source code can't really be shared publicly but I think Zachs was allowed to make exceptions for contributors.

It would really be nice if the developers can share there new integer scaler with other programs like incorporating it into mpc video renderer.

The quality seems nice based on the previous presented pictures, it would be really appreciated if they can be applied in another way.

Not sure what happened to scaler Zachs was working on, but the others are all available as open source renderscripts.

[Garteal]

Thanks Huhn - I had seen AWS meant Amazon and so have uninstalled my prior version as it threw an error and reinstalled from AWS.

Thing is I still am getting the following error on launching MPDN. This is on my x64-based Surface Pro 4/i5:

mscorlib Error
1 is not a supported code page.
Parameter name: codepage (mscorlib)

MPDN comes up but all menus are greyed out and I can't operate the program.

This has been the same error I've been getting with installation of packages from earlier builds. In the most recent installation I removed the prior installation (including the MPDN configuration files) and reinstalled from the AWS package.

Any ideas? I would love to get back to using MPDN and it has never given me troubles before. Only thing to add is that in reinstalling things on this new laptop I have installed the most recent version (4.8) of Microsoft Net Framework.
I ran into this after setting up my new system, and it appears to be a problem with MPDN (or its libraries) trying to access a codepage which isn't available in the current regional language settings.
Set your Windows Display Language (keyboard, etc) to English (United States) for it to work until MPDN gets (if ever :p) fixed.

@Shiandow @Zachs any chance one of you could set the codepage to a fixed value?

[huhn]

Thanks Huhn - I had seen AWS meant Amazon and so have uninstalled my prior version as it threw an error and reinstalled from AWS.

Thing is I still am getting the following error on launching MPDN. This is on my x64-based Surface Pro 4/i5:

mscorlib Error
1 is not a supported code page.
Parameter name: codepage (mscorlib)

MPDN comes up but all menus are greyed out and I can't operate the program.

This has been the same error I've been getting with installation of packages from earlier builds. In the most recent installation I removed the prior installation (including the MPDN configuration files) and reinstalled from the AWS package.

Any ideas? I would love to get back to using MPDN and it has never given me troubles before. Only thing to add is that in reinstalling things on this new laptop I have installed the most recent version (4.8) of Microsoft Net Framework.

i found a fix for this.

zou have to set zour windows displaz language to english *united states(

which messes up mz kezboard and no i|m not going to fix this.

[BetA13]

sad... i reinstalled my windows fresh and wanted to use MPDN again (portable) but no matter what i try, i get either and subtitle filter error or the player straight crashes without any message as soon as it loads a video.

anyone knows ho to get it running again with all extensions?
windows 10

[huhn]

the newest version doesn't work try 3603 https://www.videohelp.com/software/MPDN/old-versions

[BetA13]

the newest version doesn't work try 3603 https://www.videohelp.com/software/MPDN/old-versions

hmm, that did not work, same issue..
BUT, when i remove the EXTENSIONS Folder, the issue is gone and the video plays. (Newest MPDN Version) -> 3606

EDIT:

got it working now.. i have no idea how, lol.