madVR - high quality video renderer (GPU assisted)

[madshi]

To be honest, NG1 does seem to like Mononoke style images particularly well (although I have mostly optimized for real life photos). In other types of images waifu2x is usually clearly sharper.

Lower levels of NG1 are not finished yet, so I can't comment on that too much yet, but I can say that NG1 should pack a punch at any performance level I can manage to scale it to.

There should be a first version available by next weekend, regardless of whether I'm "finished" by then or not. I'll try very hard to release something for you to play with, at least.

[Klownicle]

Presumably it works with a 900 series card as well, but 4K support is quite limited without full HDMI 2.0, so if thats your goal you should probably upgrade.

It is my understanding that the 9 series is FULL HDMI 2.0. But HDMI 2.0a introduced HDR, and 2.0b introduces another minor modification. So HDMI 2.0 supports 4k at 60hz. Just not officially HDR. Supposedly the 9 series can be upgraded via a Software upgrade but it has been undetermined that nVidia will go that route. You can have Full 4K support with 2.0 is how I've read it.

My question about upgrading is IF I want to send HDR, I'll need a 2.0a card I imagine, BUT the question is how soon will madVR incorporate sending HDR data directly to the TV. If its not for some time, upgrading now is pointless.

[aufkrawall]

There should be a first version available by next weekend, regardless of whether I'm "finished" by then or not. I'll try very hard to release something for you to play with, at least.
Thank you.
Who cares about settings that require less performance than NNEDI3 64? :D

[nevcairiel]

It is my understanding that the 9 series is FULL HDMI 2.0. But HDMI 2.0a introduced HDR, and 2.0b introduces another minor modification. So HDMI 2.0 supports 4k at 60hz. Just not officially HDR. Supposedly the 9 series can be upgraded via a Software upgrade but it has been undetermined that nVidia will go that route. You can have Full 4K support with 2.0 is how I've read it.

https://developer.nvidia.com/high-dynamic-range-display-development


NVIDIA Support for HDR
All NVIDIA GPUs from the 900 and 1000 series support HDR display output. The presence of HDMI 2.0 provides the bandwidth necessary for the higher quality signal desirable for HDR. The 1000 series adds support for HDR over future Display Port revisions.

[JarrettH]

To be honest, NG1 does seem to like Mononoke style images particularly well (although I have mostly optimized for real life photos). In other types of images waifu2x is usually clearly sharper.

Lower levels of NG1 are not finished yet, so I can't comment on that too much yet, but I can say that NG1 should pack a punch at any performance level I can manage to scale it to.

There should be a first version available by next weekend, regardless of whether I'm "finished" by then or not. I'll try very hard to release something for you to play with, at least.

What ever you call it, please make it easily forum searchable :cool:

[Sunset1982]

I'll try very hard to release something for you to play with, at least.

That would be nice. Can't wait to try it out and give you a little feedback.

[StinDaWg]

Who cares about settings that require less performance than NNEDI3 64? :D
I do. 60 fps videos require low frame times (~16ms before frames drop). I have a ton of 720p60 content.

I absolutely cannot do NNEDI3 64 neurons with 60 fps videos on my mid level gpu at anywhere close to 16ms, so I'm hoping that this NG1 "low" still gives good quality at Jinc/super-xbr type gpu loads.

[Warner306]

I don't want to buy a brand new graphics card just to use one algorithm in madVR. NNEDI3 64 is too slow for a GTX 750 Ti. I would rather live with what that card can do than drop $300 on a GPU. Many madVR users are not gamers.

[huhn]

i don't understand why performance is an issue here.
even if your current GPU is to slow the new algorithm isn't running away.
what is so hard to live without this new algorithm for now?

[aufkrawall]

I do. 60 fps videos require low frame times (~16ms before frames drop). I have a ton of 720p60 content.

It was just a joke. madshi said that he's working on lower quality modes which should be performance-equal to Jinc.
Let's just be happy about better quality first.

[Shiandow]

For comparison purposes, here are the results with the latest MPDN version:

MPDN NNEDI3 256 taps + SuperRes (strength 1.0, 10 passes) (http://madshi.net/madVR/NG1/mononoke.mpdn-sr.png) -- | -- MPDN NNEDI3 256 taps + SSSR (sharp, 2 passes) (http://madshi.net/madVR/NG1/mononoke.mpdn-sssr.png)

Also have a look at how the various algos handle the top and bottom letterbox edges.

FWIW the 'hybrid' mode of SSSR gives a somewhat more nuanced result, it still doesn't come close to NG1 though.

[ryrynz]

I don't want to buy a brand new graphics card just to use one algorithm in madVR. NNEDI3 64 is too slow for a GTX 750 Ti. I would rather live with what that card can do than drop $300 on a GPU. Many madVR users are not gamers.

I simply can't understand this "logic". I'm considering updating my GTX 960 to a 1060 for exactly this purpose.. An extra $100 or whatever to run this new scaler comfortably for most content at low power levels? You'd be mad not to.
Or you know, you could just keep using NNEDI3 or whatever, your choice. Just don't QQ about not wanting to spend money because you can't see the benefit when it's clearly obvious.
There's been enough discussion about more demanding algorithms over the years. Do you really think the upgrade train stops here?

Because Warner306 is having a cry, I vote for:

Next Gen Mad Hardware Upgrade Scaler.

Seriously impressive result there Madshi. Does SuperRes work any good with it, probably not really needed though.

Considering how well this is doing with lines thickness I might actually no longer need to use Avisynth upscaling to achieve similar quality for low res content.
Only a few days to go..

[CarlosCaco]

I simply can't understand this "logic". I'm considering updating my GTX 960 to a 1060 for exactly this purpose.. An extra $100 or whatever to run this new scaler comfortably for most content at low power levels? You'd be mad not to.
Or you know, you could just keep using NNEDI3 or whatever, your choice. Just don't QQ about not wanting to spend money because you can't see the benefit when it's clearly obvious.
There's been enough discussion about more demanding algorithms over the years. Do you really think the upgrade train stops here?

Because Warner306 is having a cry, I vote for:

Next Gen Mad Hardware Upgrade Scaler.

Seriously impressive result there Madshi. Does SuperRes work any good with it, probably not really needed though.

Considering how well this is doing with lines thickness I might actually no longer need to use Avisynth upscaling to achieve similar quality for low res content.
Only a few days to go..



I think madVR goal is to get the most benefit that you can with your hardware...
For sometime I only had laptops with very weak gpus but even with that I can use madVR to get a very good video playback...
Now I have an desktop with an GTX 570 that can do some more things but is quite limited too...
I don t game at all but I consider video playback a passion, a great hobby so in my opinion worth spend some money on a better gpu only for madVR
Here were i live all hardware things are quite expensive comparing to others country so every upgrade that i want to do on my rig takes a lot of time to happen...
But the goal is even with a old hardware you can use madVR to really improve playback
It's a expensive hobby if you want to go deeper
But all human passions have a cost...

But I have in my update list first a decent gpu, second a good colorimeter, third a new tv
It's take time to get all of this... but this don t make my experience with madVR less enjoyable



Enviado do meu iPhone usando Tapatalk

[YxP]

I think madVR goal is to get the most benefit that you can with your hardware...


As a person who is not going to upgrade his gpu just for madvr I have to disagree. Madvr has always been about that +5% or even +1% upgrade to picture quality. The reason the new algorithm is "next generation", is that it might actually be more than just 1% or 5% upgrade. A lot more to my eyes, and I don't even have huge screens like many here do.

[ryrynz]

We'll see how the lower performance NG1 stacks up soon enough. Geforce 960s are going cheap and handle 64 neuron performance cost easily enough for all 24-30 FPS content I've tested.
Pick one up second hand with a decent cooler on it and your away laughing, probably the best card for madVR at their price point.

Consider that people would've spent thousands on hardware to achieve this a decade ago.. Many of us have dedicated HTPCs for this now.

As a person who is not going to upgrade his gpu just for madvr I have to disagree. Madvr has always been about that +5% or even +1% upgrade to picture quality. The reason the new algorithm is "next generation", is that it might actually be more than just 1% or 5% upgrade.

If you have something 750Ti+ level performance then you're going to achieve high neuron NNEDI3 + SuperRes type image quality at low NNEDI3 neuron levels. This will be a good bump up in quality for many and will no doubt be the recommended scaler going forward.


I don't even have huge screens like many here do.
Huge screens aren't necessary for best results, it does depend on how close you are to it though.

[CarlosCaco]

As a person who is not going to upgrade his gpu just for madvr I have to disagree. Madvr has always been about that +5% or even +1% upgrade to picture quality. The reason the new algorithm is "next generation", is that it might actually be more than just 1% or 5% upgrade. A lot more to my eyes, and I don't even have huge screens like many here do.



At moment my old tv is broken so I using a tv that's not even full hd... but with the material that I have I can see clearly much more than 5% improve... for example downscaling with ssim in case of lower resolution tv... but like I said it's a hobby to me... some people just use the standard settings and get a very good quality picture, some people are more sensible to some aspects like smoothness or artifacts... other not, like for example my brother only starts to perceive stutter and judder because I said and show him... until this he aways watched movies with evr... now his starts to use madVR e became very sensible to this things


Enviado do meu iPhone usando Tapatalk

[aufkrawall]

Until now our problem has been that we can't have everything good at the same time (little ringing & aliasing, good sharpness, no bloating/other artifacts).
NG1 seems to fully aim for that gap. And considering that NNEDI3 32/64 + SR3 is considered pretty much the most reasonable compromise regarding coming close to ground truth and being a good option for bad sources at the same time, it should also save us a lot of GPU power.
Imagine the deringing filter would work better with high SR strength. It would be extremely expensive, combined with NNEDI3 64 + SR3 (or even 4). And probably still look worse than NG1, compared to the screenshots.

[nautilus7]

Why Next Generation? It will be an improvement, not a next generation thing. It would suggest to name it after the creator. If the scaler is aiming at 4K then I would suggest something like madshi4K1 or madScaler4K1 or simpler mad4K1. In a few years we might get 8K with new possibilities and a new madScaler8K1 scaler. Just a thought.

madshi called it NG1, Next Generation 1. You have to pay attention on these things, if you're about to make a comment.

[ryrynz]

I'm curious about how NG1 came about more than anything. It does look to be based on NNEDI3.. did tritical have a hand in this? Hopefully a somewhat in depth post is coming..

In similar news (http://www.androidauthority.com/google-raisr-machine-learning-image-enhancement-728942/) "Google has shared details on a new image enhancement technique it has developed called RAISR. RAISR, or rapid and accurate image super-resolution, incorporates machine learning to produce high-quality versions of low-quality images.

With RAISR, the team employs machine learning to train on pairs of images, one low quality, one high, to find filters that, when applied to selectively to each pixel of the low-res image, will recreate details that are of comparable quality to the original. Essentially, RAISR uses high-quality images as a benchmark from which its algorithm can establish ways to adjust a low-resolution image."

Likely similar to NG1, considering how close the upscaled image is to ground truth.

[aufkrawall]

I think the upscaled Mononoke image would be even closer to ground truth if a sharper downscaler would have been used.

[Shiandow]

I think the upscaled Mononoke image would be even closer to ground truth if a sharper downscaler would have been used.

Well, maybe but it's kind of pointless to optimize the downscaling since that's the part we don't have usually any control over (otherwise, why would we even downscale?). Catmull-Rom in gamma light is pretty realistic. You could argue that a box filter in linear light is closer to how cameras work, but I'd be surprised if it resulted in a huge difference in the relative quality of the scaling algorithms.

[ryrynz]

I was wondering if a post process sharpener specifically designed based upon the differences from an NG1 upscaled image and ground truth could be used to bring it as close as possible to ground truth, it does seem to be quite close as it is though.

[Warner306]

I simply can't understand this "logic". I'm considering updating my GTX 960 to a 1060 for exactly this purpose.. An extra $100 or whatever to run this new scaler comfortably for most content at low power levels? You'd be mad not to.
Or you know, you could just keep using NNEDI3 or whatever, your choice. Just don't QQ about not wanting to spend money because you can't see the benefit when it's clearly obvious.
There's been enough discussion about more demanding algorithms over the years. Do you really think the upgrade train stops here?

Because Warner306 is having a cry, I vote for:

Next Gen Mad Hardware Upgrade Scaler.

Seriously impressive result there Madshi. Does SuperRes work any good with it, probably not really needed though.

Considering how well this is doing with lines thickness I might actually no longer need to use Avisynth upscaling to achieve similar quality for low res content.
Only a few days to go..

Sounds terrible. NG...anything, really, works for me. Why not NG1 or NG373, NG69?

I not crying. I'm just hopeful I don't have to open up my computer and drop some money into it. The most recent card cost me $100, used. It is good enough for my purposes, about equivalent to a PS4.

I like what I have been able to do with old hardware. I am also a madVR nut, now, and that wasn't the case when I started with HTPCs. I do want to try NG1 and get some enjoyment from it. I still publish a blog on this software that is updated every week. You can't publish 2,000 words on madVR and not care about its features.

But I really don't watch that much content these days? A movie every week or two and a couple of TV shows. That isn't worth a fortune or a GTX 1080.

[JarrettH]

You can't search for three letter words (NG1) in the forum.

[aufkrawall]

Well, maybe but it's kind of pointless to optimize the downscaling since that's the part we don't have usually any control over (otherwise, why would we even downscale?). Catmull-Rom in gamma light is pretty realistic. You could argue that a box filter in linear light is closer to how cameras work, but I'd be surprised if it resulted in a huge difference in the relative quality of the scaling algorithms.
I agree that it's realistic. My intention was to point out that it isn't necessarily the upscaler's fault if the image is a bit softer than ground truth.

[jerryleungwh]

This may be a very dumb question but I'm new to this so please bear with me. May I ask how can you guys capture an exact same frame from a video with different scaling algorithms? I want to try doing the comparisons by myself but I can don't know how to

[Warner306]

This may be a very dumb question but I'm new to this so please bear with me. May I ask how can you guys capture an exact same frame from a video with different scaling algorithms? I want to try doing the comparisons by myself but I can don't know how to

Pause the video and switch algorithms.

It is easier if you use still frame images from the Internet. Do a Google search for 1280 x 720 images, or something like that. Seeing differences in live video is more difficult compared to a good still frame image. Drag the image into the MPC-HC window and press pause. Run your tests.

[CarlosCaco]

Pause the video and switch algorithms.



And how to take screenshots that have upscaling refinements applied? If take mpc-be the refinements don t appear on the image


Enviado do meu iPhone usando Tapatalk

[HelmedHorror]

Hi, I've been trying for over an hour to fix a subtitle display problem that I'm pretty sure has to do with madVR rather than MPC-HC.

I have an "ultrawide" 21:9 aspect ratio monitor, and I use madVR's settings under Processing -> Zoom Control to zoom/crop past the hardcoded top & bottom "black bars" of most movies.

The problem is that subtitles are still rendered as if those black bars are visible.

The nature of the problem becomes clear if I just show you screenshots:

This (http://i.imgur.com/1EHOYk6.jpg) is what I see in fullscreen. (Notice the few pixels of black bar at the top and bottom - that's because 21:9 monitors aren't precisely the same aspect ratio as movies, but pretty darn close.)

This (http://i.imgur.com/GNNDvEw.png) is what I see in windowed mode.

I thought that the fix would be easy: check the "move subtitles" option in madVR and choose one of the two options in the corresponding dropdown menu. However, neither of them seemed to have any effect at all.

So I next tried all conceivable combinations of enabling/disabling the following options within madVR:

Move subtitles (to bottom of screen)
Move subtitles (to bottom of video)
Zoom small black bars away
Crop black bars

No combination of those options seemed to have any effect on the subtitles at all, even when I closed MPC-HC and played the video all over again.

Am I right that this is a problem with madVR rather than MPC-HC? Because none of the subtitle options in MPC-HC seem to have any effect whatsoever (although, neither did any of madVR's subtitle settings). If that's the case, then is there any way I can fix it?

Let me know if there's any more information that would help diagnose this, or if there's anything you'd like me to test out for you in terms of a possible solution.

Thanks!

[Asmodian]

Subtitle formats that specify position (e.g. ass) cannot be moved by madVR. What format are your subtitles?

If they are ass you can convert them to srt but it is a bit of work, especially if you aren't used to demuxing and remuxing already.

[HelmedHorror]

Subtitle formats that specify position (e.g. ass) cannot be moved by madVR. What format are your subtitles?

If they are ass you can convert them to srt but it is a bit of work, especially if you aren't used to demuxing and remuxing already.

Whatever comes with the movie in its BluRay. I think they're all PGS. Certainly the one I've been testing this problem with (The Matrix Reloaded) is PGS.

[Asmodian]

Whatever comes with the movie in its BluRay. I think they're all PGS. Certainly the one I've been testing this problem with (The Matrix Reloaded) is PGS.

PGS does specify position, I believe nothing can move them.

I always OCR my subtitles, Subextractor (https://forum.doom9.org/showthread.php?t=162633) is a good option, but that is quite a lot of work you probably don't want to do. :p

Sorry, but it isn't a madVR issue.

[Aleksoid1978]

HelmedHorror
You can try MPC-BE, can move any subtitles - Ctrl+Shift+Arrow.

[HelmedHorror]

PGS does specify position, I believe nothing can move them.

I always OCR my subtitles, Subextractor (https://forum.doom9.org/showthread.php?t=162633) is a good option, but that is quite a lot of work you probably don't want to do. :p

Sorry, but it isn't a madVR issue.

I see... Well, that's incredibly discouraging. It never ceases to amaze me the new ways I find in which 21:9 monitors aren't supported in some way or another, whether in video games or other software. Not that madVR is one such example; as you point out, this is a problem with PGS technology.

Is there at least a way I can stop madVR from getting rid of the black bars while in windowed mode? Because that way I could at least un-fullscreen if I really need to see a subtitle for some reason.

When I took those screenshots I was under the impression I had it set to keep the black bars in windowed mode, but I think I simply paused the video before it had a chance to crop the black bars. Because no matter what settings I use in madVR I can't seem to get it to stop removing the black bars in windowed mode.

[HillieSan]

madshi called it NG1, Next Generation 1. You have to pay attention on these things, if you're about to make a comment.

Indeed, I pay attention and that's why I made a comment.

[ryrynz]

Indeed, I pay attention and that's why I made a comment.

Madshi called it NG1 and nautilus7 suggested NGS, you then asked why he went with "Next Gen" and quite simply it's because madshi himself called it Next Gen.
Considering NNEDI3 was made in 2009 and this new scaler is likely based upon that to some degree, it is accurate to call it such.

[Betroz]

About the name:
Since Madshi is the creator, the algorithms name should have "Mad" in it. Like "MadSG1" or "MadScaler". Only SG1 is too anonymous.

Edit : SG1 is as we know the codename for the "Stargate SG1" TV series :D

[jkauff]

I think "NGScaler" would work. Keeps the original name, but is searchable. Or "NG1Scaler", though the "1" might be mistaken for a lower-case "L".

[ashlar42]

I still publish a blog on this software that is updated every week. You can't publish 2,000 words on madVR and not care about its features. Where's your blog?

[aufkrawall]

Considering NNEDI3 was made in 2009 and this new scaler is likely based upon that to some degree, it is accurate to call it such.
You're speculating a lot. :devil:

[Warner306]

Where's your blog?

I meant setup guide.

[Kotik]

Does anyone know if besides xySubFilter there is another subtitle renderer that is able to properly parse and show 3D subtitles?

[huhn]

try the internal subtitle filter in MPC-BE.

[kasper93]

Does anyone know if besides xySubFilter there is another subtitle renderer that is able to properly parse and show 3D subtitles?

ISR supports 3D subtitles for madVR for a long time in MPC-HC. Not sure if changes were backported to MPC-BE like huhn suggested.

[madshi]

When a 3DLUT is used for HDR to SDR conversion or HDR processing, does madVR provide Peak Brightnesss metadata to the 3DLUT?
An external 3DLUT is always hard coded to one specific peak brightness value, so it doesn't make sense to provide any metadata to a 3DLUT.

By the way, could you please add a profile variable for “Output Bit Depth” to create profiles for 10 bit output?
Why? 10bit output is an option you can choose yourself. You should know if you have it enabled or not, shouldn't you?

What i recognized is that enabling and disabling the "stereo switch" using the nvidia control panel also activates the windows 3D switch but
activating/deactivating the windows switch does not activate the nvidia switch...thats the issue as it seems.
Madvr seems to trigger the windows switch is that correct?

Last time I checked with the NVidia drivers I was using both switches were "identical", so to say. Changing one switch also changed the other one. If that's no longer the case that would probably classify as a GPU driver bug, I suppose.

I have a "little" Feature request maybe its easy to do.

1. can you add keyboard shortcut for 3D swap l/r eye.
Why? This should be handled automatically, if the source file has the correct information stored.

2. In madvr you have an detection for black bars aka 16:9(BB top under), 4:3, fullscreen...
is it possible to make an keyboard output for this? 16:9 -> Alt....+y; 4:3-> Alt...+x (just an Example)
so we could use it for automatic Ambilight switching.(with Ambibox for example)
Keyboard output? What do you mean with that?

Hi Guys,
I own an HTPC with Win7 64bit and a GTX970
This bug happens only If I send a 2160p signal with 12bpc color depth...

[...]
Sounds like a GPU driver issue to me. Probably 2160p at 12bpc is not supported by "someone" (either the GPU or receiver or display), so somebody does some conversions somewhere which screw things up. This is most probably outside of my control and not my fault.

MadVR is no longer enabling 3D Stereo in the OS automatically after Win 10 Anniversary Update for me. Resolution automatically switches to 1080 (from my 4k default), however 3D in the OS is not enabled, so video is shown 2D only. To fix, I have to manually go into nVidia settings and enable Stereoscopic 3D myself every time.

I keep saying, Windows 8.1 is the best media OS atm. I don't think this is my fault, probably a GPU driver issue.

can someone tell me what is neccessary to be able to access madvr/madtgp over the network using windows 10 (dispcal or HCFR)?

Both computers are in the same network, LAN access is enabled for madr and firewall rules are set up. But i am still not able to use the pattern generator via network.

I have tried shutting down antivirus and firewall as well... still no luck.
Madvr is only installed on the HTPC and i am trying to access it via my notebook where madvr is not installed.. i have installed it for testing and enabled to show the icon only when there is another madvr instance on the network and it does not show up as well.

Is there any way to test if madvr is working correctly and is reachable in this regard?

Anything special i need to take a look at when using windows 10 ?:thanks:
I'm not aware of any special needs. It usually works just fine for me. I'm not completely sure if dispcal/HCFR allow you to ender an IP address, or if they always rely on the automatic find routine? IIRC Calman allows to enter an IP address, in case the automatic find routine fails for some weird reason...

Is it possible to somehow render MadVR's output to a file or back into vapoursynth / avisynth or something?

I'm basically looking for a way to encode HDR clips / movies to regular SDR files in the way madVr does the converting / compression. It looks awesome (to me) and that way I can take the movies with me on tablet or in the non-HDR tv in the bedroom.

I'm converting it now manually, and I tweak it manually with a final gamma adjustment and a levels/histogram adjustment to get it looking nice. But madVr's way is nicer, and my way probably works fine for shorter clips but for longer clips or full movies it will look nice in one part but horrible dark or bright in another part :S.
It's not possible atm. It's on my to do list to investigate, but it probably won't come too soon.

Fow now you could try using displayCAL to create HDR -> SDR conversion 3dluts and apply them using avisynth. I think there's some avisynth script from tritical available which can handle madVR 3dluts, IIRC? Quality should be identical to madVR, or even slightly better.

I'm currently trying to improve my picture in 3D mode. How can you do a calibration in 3D? Using the glasses and the colorimeter at the same time is already quite problematic but what are the other prerequisites?

Do I just need to activate the stereoscopic mode in Windows and run a calibration through madVR/DispCAL?

I suppose so. Haven't really tried myself. I hope left and right eye will need the same calibration... :sly:

All the easy shader passes take 0.49ms on my Titan X and this is independent of clock speed and resolution.

Please note that in order to measure shader passes I have to flush the GPU and wait for the rendering to succeed. This will actually slow overall rendering speed down. So I wouldn't bet on that each shader pass actually takes that much time when not forcing the GPU to flush after every render pass.

Some 4chan morons would like to sue madshi:

And it seems that another one released all the madshi's algorithms for fun
Annoying, but I'm glad they did this before I published NGU and not after, so I could add a little bit (not too much, though) of protection against reverse engineering.

madVR prejudice exists an algorithm similar to Darbee, I want to remove the radiance between HTPC and projector though the effect of Darbee I like as much as I do?

I don't really like the look the Darbee produces, but I think you can achieve something similar by using the madVR "image enhancements" or "upscaling refinement" option. Furthermore, if you have a 4K display, the Darbee does its work in 1080p and the result is then upscaled somewhere to 4K, while with madVR you have the option to use a better quality upscaler and then sharpen in 4K, which should produce noticeably superior results.

3D picture looks fantastic with madvr, but subtitle depth (plane) goes occasionally wrong resulting my eyes going awry. I'm watching original blu-ray ISO's and using madVR v0.90.20, LAVfilters 0.68.1. MPC-HC + Nvidias 373.06 drivers.

Been testing same movies with Pioneer blu-ray player and the subtitle depth is correct with that.

A few questions:

1) Can you describe in more detail how the subtitle depth goes "wrong"? Are the subs too far in front or too far back or jumping around or something else?

2) Does the same problem always occur at the same play time with the same movie? Or does the problem sometimes not occur if you rewind back and play the same movie section again?

3) If you can perfectly reproduce the problem and if it always occurs at exactly the same movie runtime, maybe you can create a little sample for nevcairiel and me to look at?

Does anybody use the "black bar detection" feature for automatic screen masquerading in your projector setups? MadVR reliable detects the actual aspect ratio and also sends this info optional to the media player (JRiverMC in my case). So it might be possible to adjust motorized screens to respective aspect ratio. I tried jriver http commands, but I don't think this info is available there. In madVR OSD the aspect ratio is shown, but I don't know how to get hands on this info. Maybe someone could give me a hint. It would be great if mediaplayer or better madVR could send UDP Broadcast with current movie aspect ratio. This might be a very cool additional feature for madVR.

One thing you could do is create profiles for different aspect ratios and setup rules to auto activate the matching profiles. Then in each profile you can define a command line which madVR shall execute if said profile gets activated.

Unfortunately the command line execution is currently not implemented yet <sigh>. :(

When can we expect HDR flagging on HDMI?

Based on this looks like it's possible:

https://developer.nvidia.com/displaying-hdr-nuts-and-bolts

The latest drivers have a better API than the one you linked to, but I don't have the API documentation yet. I'm waiting for it to be released.

right now a little off topic from madshis new algo, but here's a little update on my automatic screen masking topic. Madvr is abolute mad awesome! Quite everything I need is already there and right before my eyes. You can easily use profile scripts to detect AR and switch profiles accordingly. There is an option "command line to execute when this profile is activated". Unfortunately there seems to be an issue with this option. I couldn't get a simple command to be executed (e. g. curl command for http get). I also found an old bug ticket: http://bugs.madshi.net/view.php?id=210 which hasn't been resolved. If anyone has experienced the same, do you have a workaround?
I'll increase priority in my to do list.

NG1 is excellent, but waifu does some magic on horizontal lines
Yes, it does. But it's also about a 100x slower than NGU.

I'm curious about how NG1 came about more than anything. It does look to be based on NNEDI3.. did tritical have a hand in this?

I haven't heard from tritical in years. NGU is not based on NNEDI3 in any way. E.g. NNEDI3 needs OpenCL (or D3D11 shaders) to run with decent performance. NGU runs in simple PS3.0 pixel shaders.

Nobody has had a hand in NGU other than me. Well, of course I must have read several dozen scientific papers about image processing and upscaling, and I've certainly taken many good ideas from several such papers, and added a couple of my own ideas.

In similar news (http://www.androidauthority.com/google-raisr-machine-learning-image-enhancement-728942/) "Google has shared details on a new image enhancement technique it has developed called RAISR. RAISR, or rapid and accurate image super-resolution, incorporates machine learning to produce high-quality versions of low-quality images.
Doesn't look all that great to me, to be honest.

Well, maybe but it's kind of pointless to optimize the downscaling since that's the part we don't have usually any control over (otherwise, why would we even downscale?). Catmull-Rom in gamma light is pretty realistic. You could argue that a box filter in linear light is closer to how cameras work, but I'd be surprised if it resulted in a huge difference in the relative quality of the scaling algorithms.
Agreed.

Hi, I've been trying for over an hour to fix a subtitle display problem that I'm pretty sure has to do with madVR rather than MPC-HC.

I have an "ultrawide" 21:9 aspect ratio monitor, and I use madVR's settings under Processing -> Zoom Control to zoom/crop past the hardcoded top & bottom "black bars" of most movies.

The problem is that subtitles are still rendered as if those black bars are visible.
Does madVR correctly detect the movie aspect ratio? Check the OSD (Ctrl+J) for that.

Which subtitle renderer are you using?

PGS does specify position, I believe nothing can move them.
Actually, madVR is supposed to be able to move PGS subtitles, unless they're encoding in "fullscreen". But most PGS subtitles I've seen are encoded in small rectangles.

I have a 21:9 setup myself and moving all subtitles (including PGS) into the active image area works just fine for me.

[madshi]

madVR v0.91.0 released

http://madshi.net/madVR.zip

* added new NGU (Next Generation Upscaling) algorithm
* added "soften edges" + "add grain" upscaling refinements (atm only for NGU)
* display properties can now also be profiled

A few comments about NGU:

1) Obviously, I've decided to change the name from "NG1" to "NGU". The reason is that "NG1" didn't have any mention of "upscaling" or "super resolution" in the name. So I added the "U" for "upscaling".

2) In this initial version only 2 quality levels (high and medium) are available. A future version will add at least one more level which is faster than medium, maybe there will be even more levels, I don't know yet.

3) NGU is currently available for both doubling and chroma upscaling. The chroma upscaling algo does currently not make use of the luma channel. I plan to try adding that to a future NGU version, but that will take serious development time, so don't expect it too soon.

4) NGU doesn't need OpenCL, DirectCompute or D3D11. It runs fine with D3D9, so it should run fine in XP.

5) A word about "ideal upscaling": When trying to upscale a low-res image, it's possible to get the edges very sharp and very near to the "groundtruth" (the original high-res image the low-res image was created from). However, texture detail which is lost during downscaling cannot properly be restored. This can lead to "cartoon" type images when upscaling by large factors with full sharpness, because the edges will be very sharp, but there's no texture detail. In order to soften this problem, I've added options to "soften edges" and "add grain". Here's a little comparison to show the effect of these options:

low-res image (http://madshi.net/madVR/NGU/clownOrgLanczos.png) -- | -- straight NGU (http://madshi.net/madVR/NGU/clown4xNguSharp.png) -- | -- NGU + "Soften Edges" + "Add Grain" (http://madshi.net/madVR/NGU/clown4xNguSoft.png) -- | -- Jinc AR (http://madshi.net/madVR/NGU/clown4xJinc.png)

I hope these extra options will make NGU acceptable for all of you, including those who preferred Jinc so far.

Now I'd like your feedback:

1) Do we still need NNEDI3 image doubling? Or can NGU completely replace it?
2) Do we still need SuperRes (image doubling)? Or can NGU completely replace it?
3) Do we still need NEDI image doubling?

I'll not ask about super-xbr image doubling yet because I've not scaled NGU down to that performance level yet.

[Dreamland]

Excuse me for my very bad english
Thx for new great release !
I can't install new release MadVr on win 10 Anniversary Update ( 1607 ) 64 Bit

http://t8.pixhost.org/thumbs/1018/35305707_appunti01.jpg (http://pixhost.org/show/1018/35305707_appunti01.jpg)

[madshi]

Did you have an older version installed? In that case you can just copy the files over, no need to uninstall again. Otherwise try right click -> run as admin. If that also doesn't help, you may have to run "regsvr32 madVR.ax" manually in an admin command line window.

[toniash]

madVR v0.91.0 released

http://madshi.net/madVR.zip

* added new NGU (Next Generation Upscaling) algorithm
* added "soften edges" + "add grain" upscaling refinements (atm only for NGU)
* display properties can now also be profiled

...

I'll not ask about super-xbr image doubling yet because I've not scaled NGU down to that performance level yet.

Thanks Madshi! Very good job

Where do I can find soften edges and add grain?

I'm waiting for a more scaled down version , too hungry for me (GTX 730)

[Dreamland]

Did you have an older version installed? In that case you can just copy the files over, no need to uninstall again. Otherwise try right click -> run as admin. If that also doesn't help, you may have to run "regsvr32 madVR.ax" manually in an admin command line window.

I have older version inside MPC HC folder and works very well ..
this release doesnt work (click -> run as admin as always )

edit

copy the files over an older last version and works