madVR - high quality video renderer (GPU assisted)

[turbojet]

Nachbar madlevelstweaker didn't work for me either but this (http://blog.metaclassofnil.com/?p=83) did.

[Nachbar]

Nachbar madlevelstweaker didn't work for me either but this (http://blog.metaclassofnil.com/?p=83) did.

Thank you very much for this! Now windows is using the same darker color level as madvr and I can adjust the brightness accordingly.

[MistahBonzai]

However, I don't understand what you mean by "Zoom Player doesn't honor those settings", the only madVR setting Zoom Player overwrites is smooth motion which is controlled by Zoom Player's own setting. I get 1000s of frame drops on my overclocked Titan if I use 256 neuron luma doubling for 720p -> 1440p or NNEDI3 chroma scaling with 4K@60p video or something.

I apologize for stating non-existent limitations associated with the Zoom Player. It wasn't my intent to unjustly disparage it's performance. My lack of familiarity with the set-up requirements coupled with the limitations of the free introductory version led me to some false conclusions :(

I see..now - after giving it a more thorough test session. When I first tried it I didn't immediately see the effects and assumed... This was after trying a '4k' video in which Zoom Player automagically downscaled the input to 1080p prior to MadVR rendering. MPC-HC will report it as native resolution (4K) although MadVR will render it to match the display. However performing an image capture (alt+i) with MadVR does capture the 4K image - not so with Zoom Player.

This time I let it play longer with a 720p/24fps video and saw it was dropping (lotsa) frames (as expected) although the NNEDI3 sharpening effects of 'image doubling' didn't appear for 20+ seconds with the first attempt. After a few more attempts they began to switch with the changes in MadVR settings. No idea what was going on there...

I was also frustrated by the lack of concise information regarding the technical limitations of the 'free version' and the lack of status reporting when an 'out of bounds' selection was made. Example: I spent maybe 20 minutes trying to replace the 'native' audio renderer with the AC3 Filter (I use 7.1 virtual surround head phones - they can't deal with 32-bit PCM Zoom Player feed me by default - only to discover the settings weren't holding and to utilize it I needed to upgrade (still not knowing if AC3 would work as I expected). I never did discover if/how I could insert ffdshow in the video decoding stream (I need that to support real time effects supported by avisynth).

All in all I was quite frustrated by the time I began testing the MadVR tweaks and had lost the urge to explore further. Sorry Zoom Player...

I'm not comfortable with a video player that tries to do the 'right thing' rather than what I ask it to do. I can see how Zoom Player would be appreciated by those who actually view media content rather than 'pixel peep' as I do. :)

[Asmodian]

I'm not comfortable with a video player that tries to do the 'right thing' rather than what I ask it to do. I can see how Zoom Player would be appreciated by those who actually view media content rather than 'pixel peep' as I do. :)

Odd, maybe it is an issue with limitations of the free version. I have used the MAX version for over a decade now and I am annoyed by all other players because they do not do what I ask or are missing the option entirely.

If Smart Play is off (Smart Play is the main thing not available in the free version) Zoom Player uses the same Windows filter merit system of every other directshow based player. It is very easy to change audio or video renderers and it even has a built in filter merit editor that still works in Windows 8.1. It sounds like you don't know how filter merits work? All the examples you state are not something you configure in Zoom Player (without smart play) but something you configure in ffdshow or AC3filter or adjust the merits so the filters you want are loaded. e.g. it isn't Zoom Player that sets the audio format, it is the audio decoder.

I too am a pixel peeper, spending more time messing with settings than actually watching media. :)

Sorry for caring so much, it is just that madVR + LAV + Zoom Player has been such a good combination for me that I cannot help myself. :o

This was after trying a '4k' video in which Zoom Player automagically downscaled the input to 1080p prior to MadVR rendering. MPC-HC will report it as native resolution (4K) although MadVR will render it to match the display. However performing an image capture (alt+i) with MadVR does capture the 4K image - not so with Zoom Player.

This doesn't happen on my system. Zoom Player will not resize the input before it gets to madVR, that isn't even possible as I understand it (ffdshow might but not Zoom Player). I don't know anything about MPC-HC's screen capture but it sounds like it happens before madVR if the image is captured at a different resolution than madVR is displaying.

[romulous]

(Smart Play is the main thing not available in the free version)

Not true - Smart Play is available in Zoom Free. What you may be thinking of is that you can't create your own Smart Play profiles in Zoom Free (you need Zoom Pro or Zoom Max) - but Zoom Free still uses the built-in Smart Play profiles, which is enough for the majority of users (only a small minority need to create their own).

romulous

[Asmodian]

Ah Thanks, I didn't know that. That is probably what is annoying MistahBonzai then.

@MistahBonzai, you should probably just turn off Smart Play and ignore my above post. :)

Or just use MPC-HC if it works for you. I bet no other player has a way to set close to the middle mouse button though. ;)

[secvensor]

Whether it will be difficult to create temp a video file with completely processed video on the maximum options of quality, for the subsequent viewing with application of effects madvr not in real time?

[LexSfX]

(1) First of all if I wrote a media player, I would *not* add a "keep aspect ratio" option at all. It doesn't make sense. Why would anyone watch a video with distorted aspect ratio? The only situation where "keep aspect ratio" is useful is if the video was encoded incorrectly, *and* if at the same time the correct aspect ratio actually happens to match the resolution of the video (as in your case). That is a very rare situation, though.The video I linked is not encoded incorrectly. It's encoded with the 4:3 aspect ratio flag in hopes of giving what is thought to be the most common viewing preference, which is what Nintendo Entertainment System footage is usually displayed in for most gamers who played on their CRT TVs in the 1980s and 1990s. Some NES games do not compensate graphically for displaying the game at this aspect ratio, so I (and others I know in various video game and video game console emulation communities) want to be able to play games or watch NES footage with square pixels. The encoder encoded using the aspect ratio flag knowing that media players typically have the option to ignore it. With this kind of footage, the aspect ratio is just a matter of preference, which is why there is an option in MPC-HC and LAV filters for this.

I hope that clears up any confusion.

Note that this is not a demand for anything, I love madVR's main functionality, and I admire you as a diligent and selfless programmer.

[jkauff]

If you go into the settings the profile that is currently in use is bold. No OSD messages. They are automatic.
If madVR automatically runs my script at startup and recognizes the media type, I'm fine with that. The Help page for scripting doesn't say anything about it one way or the other.

[SamuriHL]

This is for everyone on this thread. I'd like to ask for help in getting some data for my next madVR algorithm graph. Now that 750 Ti is out there and we're going to get some Maxwell refreshes I'm starting to accommodate some settings. NNEDI3 is the only thing I can't test at the moment so some performance numbers with it (ms, dropped frames, anything) would be greatly appreciated. All I need is your GPU and whatever data you can give me so I can parse it.

I'm getting my new laptop tomorrow with an 880m and 8 gigs of ram (LOL). I'll be testing it out for sure.

[markanini]

I do have profile rules set up, and I thought I needed to use the hotkey to call up the Program Group and run the rules. If the rules script gets run automatically, I don't see the OSD message that shows which sub-group has been selected by the rules. There's nothing that tells me the script has run.

I have rules set up for SD video, 720p, and 1080p for doubling. Works perfectly, but I wish I didn't need to use the hotkey every time I play a movie.

If you set it up like in post #3 the currently active profile is highlighted in properties.

[seiyafan]

This is for everyone on this thread. I'd like to ask for help in getting some data for my next madVR algorithm graph. Now that 750 Ti is out there and we're going to get some Maxwell refreshes I'm starting to accommodate some settings. NNEDI3 is the only thing I can't test at the moment so some performance numbers with it (ms, dropped frames, anything) would be greatly appreciated. All I need is your GPU and whatever data you can give me so I can parse it.

I did an extensive test for 270x, 290x and 770 a while ago, check them out. :)

My conclusion is based on 1080 to 1440 upscaling, 270x is the minimum card to do NNEDI3 doubling at 16 neurons, 290x can do that at 32 neurons plus error diffusion. 770 falls in between the two but it's slower with error diffusion.

At the end I settled with a MSI 270x card because the fan speed did not change a bit from idle to hours of movie playing, it was extremely quiet.

[fedpul]

Hi, this is my first post here. I was testing with the chroma resolution test and want to post some pics. I think my TV is 4:2:2 with PC Mode Off and I'm sure it is in 4:4:4 with PC Mode On. I also take two pics comparing the effect on a particular movie (Brave) just for comparison. Colors looks more vivid in 4:2:2 and the picture quality looks superior, less blurry and sharper. The effect is easily perceived in grass.

Chroma Resolution Test PC Mode Off
http://i61.tinypic.com/r1c6di.jpg

Chroma Resolution Test PC Mode On
http://i62.tinypic.com/1418m7t.jpg

Brave PC Mode Off
http://i61.tinypic.com/121en3n.jpg

Brave PC Mode On
http://i60.tinypic.com/2zzsxog.jpg

My PC Specs:
CPU: i5 3570k @ 4.2Ghz
GPU: HD 7870 @ 1100/1375
RAM: 8Gb @ 1600Mhz

My TV: Philips 32PFL3017D

Using MPC-HC+LAV+madVR+Reclock

Both modes (PC Mode Off and On) are calibrated using AVS BT.709 Calibration Disc. Waiting for filters to calibrate Color and Tint.

Greetings from Argentina.

[huhn]

Yes..I miswrote...here are no blue lines inside the 4:4:4....it is clear...the lines are around the rest of the image....the only thing that has got me confused is that that the 4:4:4 looks light colored inside it and and the outside of it is darker..almost the reverse coloring of the image below..what I mean is that I DO see 4:4:4 clearly visible, except it is light colored and the background is dark. In the image below it is the opposite in brightness, which confuses me:

http://abload.de/img/444-examplewpjy0.png



this souunds like your display is not showing 1:1 pixelmapping?

So if one had a 4:4:4 display what do you suggest for chroma up scaling in madvr
and if one had a 4:2:2 display what are the recommended chroma settings?

what looks best to your eyes. most people say jinc 3 ar. i personally use spline 3 ar. the effect is just higher on a 4:4:4 screen or it's more worth it.

[huhn]

Hi, this is my first post here. I was testing with the chroma resolution test and want to post some pics. I think my TV is 4:2:2 with PC Mode Off and I'm sure it is in 4:4:4 with PC Mode On. I also take two pics comparing the effect on a particular movie (Brave) just for comparison. Colors looks more vivid in 4:2:2 and the picture quality looks superior, less blurry and sharper. The effect is easily perceived in grass.
Specs:
CPU: i5 3570k @ 4.2Ghz
GPU: HD 7870 @ 1100/1375
RAM: 8Gb @ 1600Mhz

My TV: Philips 32PFL3017D

Using MPC-HC+LAV+madVR+Reclock

Both modes (PC Mode Off and On) are calibrated using AVS BT.709 Calibration Disc. Waiting for filters to calibrate Color and Tint.

Greetings from Argentina.

force full range color output in pc mode this is a normal "bug" with philiphs tv in pc mode is needs a unlimited signal not a limited.

just go to the ccc and change to untimited 4:4:4 full rgb

[fedpul]

Sorry, I forgot to mention i am alreafy using Full RGB. In ATI Control Panel Pixel Format is set to 4:4:4 Full RGB. Also I'm using PC Levels 0-255 in madVR. that's the way I calibrate my display too, using full RGB with PC Levels.

[6233638]

Unfortunately on a 1440p monitor with 270x I can choose to do either NEDDI3 luma doubling, NEDDI3 chroma upscaling, or Error diffusion dithering. I think the doubling has the greatest visual improvement, not sure about the other two though. I doubt it's possible to do all three on a 290x, but two is possible on 280x and above.NNEDI3 Luma doubling should have the most obvious visual improvements.
NNEDI3 Chroma doubling had very little benefit at all in my testing. (though NNEDI3 Chroma scaling is very nice)

I would use Ordered Dithering rather than Error Diffusion if it lets you use higher quality settings elsewhere. Madshi has done a very good job with it.

noob question from me as I'm still learning so much....My projector shows 4:4:4 in Game mode...but 4:2:0 in Cinema1 mode which is what I use to watch movies. Are there any recommended settings for madVR (chroma wise) when watching 4:2:0 display?If it's 4:2:0 then your chroma settings almost don't matter at all. You could probably even reduce them to Bilinear (the worst quality) with no noticeable effect - though I would prefer to at least use Bicubic 75 at a minimum.

Yes..I miswrote...here are no blue lines inside the 4:4:4....it is clear...the lines are around the rest of the image....the only thing that has got me confused is that that the 4:4:4 looks light colored inside it and and the outside of it is darker..almost the reverse coloring of the image below..what I mean is that I DO see 4:4:4 clearly visible, except it is light colored and the background is dark. In the image below it is the opposite in brightness, which confuses meIf you could post a photograph, that might be helpful. It sounds like your display is doing some sort of processing (image scaling, sharpening, dynamic contrast or something else?) which is affecting the results.

Hi, this is my first post here. I was testing with the chroma resolution test and want to post some pics. I think my TV is 4:2:2 with PC Mode Off and I'm sure it is in 4:4:4 with PC Mode On. I also take two pics comparing the effect on a particular movie (Brave) just for comparison. Colors looks more vivid in 4:2:2 and the picture quality looks superior, less blurry and sharper. The effect is easily perceived in grass.Thank you for the photographs - they're perfect examples of what should be seen with 4:4:4 and 4:2:2 displays.
Unfortunately, many televisions will disable a lot of processing options when you put them into PC mode.

On some displays like my Sony TV, the main thing that gets disabled is the motion interpolation/backlight scanning options.
On other displays though, it will disable things like the color management system or may even limit the refresh rate to 60Hz.

In your case, it looks like it is either changing the gamma, disabling some kind of dynamic contrast processing, or possibly changing the input level range from 16-235 to 0-255. It's difficult to say.

But for video purposes, it's usually not worth losing these options just to display 4:4:4 chroma.
If you're using the display as a monitor, however, then 4:4:4 chroma becomes very important.

[seiyafan]

A dumb question, if I use computer monitor, do I need to concern about whether it's 4:2:2 or 4:4:4? It seems all the talk recently is about TV.

[James Freeman]

Most if not all PC monitors are 4:4:4.

4:4:4 has more color resolution than 4:2:2.

[flashmozzg]

sweet! a benchmark :p

http://s8.postimg.org/7o578j0zp/madvr_test01.png (http://postimage.org/)
capture (http://postimage.org/app.php)

system specs here:
http://valid.canardpc.com/cache/banner/hicw7d.png (http://valid.canardpc.com/hicw7d)
(it's a 7950)

Some inconsistency with the results. 7850 here.

http://i.imgur.com/y5JEiXc.png

[StinDaWg]

Sorry, I forgot to mention i am alreafy using Full RGB. In ATI Control Panel Pixel Format is set to 4:4:4 Full RGB. Also I'm using PC Levels 0-255 in madVR. that's the way I calibrate my display too, using full RGB with PC Levels.
For what it's worth, 4:4:4 Full RGB doesn't look right on my Panasonic plasma either, it's blurry. I use 4:4:4 YCbCr.

[e-t172]

I can't say if madVR is triggering it, at least I wouldn't know which madVR behaviour would/could be responsible for that. But one thing is probably important to mention: There are quite a lot of madVR users, many of them posting regularly in this thread, and I think the majority of them have NVidia GPUs, but it seems you're the only one who has this specific problem (at least that I'm aware of). That indicates that there is probably something special about your PC.


I just tried the following:
- Moved GPU to another PCIe port -> no difference
- Used DisplayPort instead of DVI-D (heh, you never know) -> no difference

And then finally I did a fresh clean Windows install on a separate drive, installed DirectX, nVidia drivers, MPC-HC and madVR (nothing else) and tried again. The problem is still there, and XPerf still shows the same peaks...

So I guess at this point a purely hardware issue (either GPU or motherboard) is the only possible cause. I was planning to upgrade my motherboard and CPU anyway, so I just accelerated my plans and placed my order today.

Just to confirm though: to any 780 Ti owners reading this, can you explicitly confirm that you're able to use madVR for 30+ minutes (preferably with Smooth Motion on) without any frame drops or presentation glitches? Just making sure that this is not a widespread defect.

[scollaco]

If you could post a photograph, that might be helpful. It sounds like your display is doing some sort of processing (image scaling, sharpening, dynamic contrast or something else?) which is affecting the results.



I made sure all post processing that I could find was turned off. I loaded the image in paint and made sure it was full screen at 100%. How else can I make sure it 1:1 pixel mapping?
My Nvidia card is set to RGB 0-255 using the registry hack tool.

I took a photograph of what I'm seeing on my screen here...the 4:4:4 shows up but is light colored not dark colored. I'm not sure what I'm missing? Movies look all properly colored to me through MadVR and my calibration disc all look correct for brightness/contrast/tint and color. Not sure what could be wrong.

http://s24.postimg.org/remwi6msl/HPIM0787.jpg

[soulkeeper]

NNEDI3 Luma doubling should have the most obvious visual improvements.
NNEDI3 Chroma doubling had very little benefit at all in my testing. (though NNEDI3 Chroma scaling is very nice)

I would use Ordered Dithering rather than Error Diffusion if it lets you use higher quality settings elsewhere. Madshi has done a very good job with it.



Is there any benefit using NNEDI3 luma doubling when playing 1080p source in a 1080p screen (4:4:4 screen)?

[Shiandow]

Hi, this is my first post here. I was testing with the chroma resolution test and want to post some pics. I think my TV is 4:2:2 with PC Mode Off and I'm sure it is in 4:4:4 with PC Mode On. I also take two pics comparing the effect on a particular movie (Brave) just for comparison. Colors looks more vivid in 4:2:2 and the picture quality looks superior, less blurry and sharper. The effect is easily perceived in grass.
[...]


The images with PC mode off seem to have been sharpened quite a lot and the saturation seems to be exaggerated, personally I don't really like the result but it's hard to judge without actually seeing the screen. I think you'll be able to achieve better results with PC mode on, but you may need to experiment a bit with the settings.

You may also want to replace those images by links, since they currently take up a lot of space which makes it hard to read your post.

[seiyafan]

Is there any benefit using NNEDI3 luma doubling when playing 1080p source in a 1080p screen (4:4:4 screen)?

In that case you don't need any upscaling. The whole point of MadVR is when output has higher resolution than the source.

[fedpul]

NNEDI3 Luma doubling should have the most obvious visual improvements.
NNEDI3 Chroma doubling had very little benefit at all in my testing. (though NNEDI3 Chroma scaling is very nice)

I would use Ordered Dithering rather than Error Diffusion if it lets you use higher quality settings elsewhere. Madshi has done a very good job with it.

If it's 4:2:0 then your chroma settings almost don't matter at all. You could probably even reduce them to Bilinear (the worst quality) with no noticeable effect - though I would prefer to at least use Bicubic 75 at a minimum.

If you could post a photograph, that might be helpful. It sounds like your display is doing some sort of processing (image scaling, sharpening, dynamic contrast or something else?) which is affecting the results.

Thank you for the photographs - they're perfect examples of what should be seen with 4:4:4 and 4:2:2 displays.
Unfortunately, many televisions will disable a lot of processing options when you put them into PC mode.

On some displays like my Sony TV, the main thing that gets disabled is the motion interpolation/backlight scanning options.
On other displays though, it will disable things like the color management system or may even limit the refresh rate to 60Hz.

In your case, it looks like it is either changing the gamma, disabling some kind of dynamic contrast processing, or possibly changing the input level range from 16-235 to 0-255. It's difficult to say.

But for video purposes, it's usually not worth losing these options just to display 4:4:4 chroma.
If you're using the display as a monitor, however, then 4:4:4 chroma becomes very important.

Exactly, in PC Mode I lose lots of settings in the tv. I just have contrast, brightness and color. I think is preferable to sacrifice chroma a bit as you said.

I hope pics serves well for comparison as i haven't seen them before.

As for image upscaling and other options i will put my madVR setting:

Display: Calibrated to BT.709, Pure Power Curve 2.20
Display Modes: 1080p24, 1080p50, 1080p60.
Chroma Upscaling: NNEDI3 64 Neurons
Image Doubling: Luma and Chroma 32 Neurons only if scaling factor is 2.0x or bigger
Image Upscaling: Jinc AR 4 taps
Image Downscaling: Catmull-Rom AR LL
Dithering: ED 1 with the 2 options disabled

Sharpness set to 4/20 in TV, it goes well up to 7, past this point ringing is too noticeable. Tested with last pattern in Basic Settings with AVS BT.709 Calibration Disc.

Color and Tint to midpoint in TV as I'm waiting for the blue filter to set it.

Thanks for your answer.

PD: I have learned a lot with this forum.

[6233638]

I took a photograph of what I'm seeing on my screen here...the 4:4:4 shows up but is light colored not dark colored. I'm not sure what I'm missing? Movies look all properly colored to me through MadVR and my calibration disc all look correct for brightness/contrast/tint and color. Not sure what could be wrong.

http://s24.postimg.org/remwi6msl/HPIM0787.jpgI'm not sure why the "4:4:4" text is lighter than the background - I suppose it just depends on how the TV blends things for the chroma subsampling - but that looks like 4:2:0 since both the "4:2:2" text and the background are blurred.

[Qaq]

Unfortunately, many televisions will disable a lot of processing options when you put them into PC mode.Some video driver options are lost as well:- bypasses graphics card's video (damage) algorithmsSo why graphics card's video algorithms are damage a picture and TV's ones do not? :)

[huhn]

Some video driver options are lost as well:So why graphics card's video algorithms are damage a picture and TV's ones do not? :)

i got the same question in my head.

i only see a lose in backlight strobing/scanning and 60 hz only.

[nevcairiel]

In my experience you can just turn them off on a TV without switching to PC mode if you don't like them. It usually doesn't enforce anything.
The only algorithm I use in my TV is its motion processing, in Sony's "Clear" mode, which results in smoother motions without artifacts (although not perfectly smooth like the artifact-riddled "Smooth" mode does)

[scollaco]

I'm not sure why the "4:4:4" text is lighter than the background - I suppose it just depends on how the TV blends things for the chroma subsampling - but that looks like 4:2:0 since both the "4:2:2" text and the background are blurred.

Thanks for the reply...What do suggest for madVR chroma scaling if we assume my display is 4:2:0?

[madshi]

Hey, why do I get low performance with NNEDI3 1080p dxva supported mkv. 720p works fine. Jinc works fine for all too.

Intel I7 720 1.6Ghz
ATI 5870 Mobility
Not enough information. NNEDI3 for what? Chroma upscaling? Luma doubling? Chroma doubling? How many neurons? How are your render times with 720p and 1080p? And what do you mean with "dxva supported mkv"? Do you mean video files that are deinterlaced by DXVA deinterlacing? Or do you mean DXVA decoding?

No no no - madVR 0.86.9 does work. I just thought that since I replaced that folder earlier with newer versions, the older version didn't get uninstalled properly. That is why I re-downloaded it to uninstall properly, but that didn't fix the problem with 0.87.7...

I did try neutral location - no go...

Could you at least tell me whether the gray screen is as intended??? If not, then its not ArgyllCMS...
I already told you about the gray screen. Look back some pages.

Found the cause of the problem. You're probably running the 64bit Argyll files. I've tested with the 32bit Argyll files. The madHcNet64.dll needs to be updated for the 64bit Argyll exes to work properly. I'll do that in the next build. For now you can work around the problem by using the 32bit Argyll exes.

I do have profile rules set up, and I thought I needed to use the hotkey to call up the Program Group and run the rules. If the rules script gets run automatically, I don't see the OSD message that shows which sub-group has been selected by the rules. There's nothing that tells me the script has run.

I have rules set up for SD video, 720p, and 1080p for doubling. Works perfectly, but I wish I didn't need to use the hotkey every time I play a movie.
You don't need to. The rules scripts is run automatically, of course. It would be stupid otherwise.

Whether it will be difficult to create temp a video file with completely processed video on the maximum options of quality, for the subsequent viewing with application of effects madvr not in real time?
It's currently not easily possible. I guess you could use FRAPS to record the madVR output and then record it. But I'm not sure if that's worth the effort and I also don't know if FRAPS does it correctly with exactly the correct framerate etc.

I did an extensive test for 270x, 290x and 770 a while ago, check them out. :)
Unfortunately in the meanwhile user reports show the OpenCL <-> D3D9 interop cost seems to vary greatly from one PC to the next, even with the same GPU. Which means that I fear that your tests might only be true for your PC. It's possible that NNEDI3 will run faster or slower on somebody else's PC... :(

Is there any benefit using NNEDI3 luma doubling when playing 1080p source in a 1080p screen (4:4:4 screen)?
Not really, IMHO, unless you want to sharpen in double resolution and then downscale to the original resolution again. But that's currently not supported by madVR.

The whole point of MadVR is when output has higher resolution than the source.
There are many benefits to madVR which don't have anything to do with upscaling. So I wouldn't say the whole point of madVR is upscaling.

[MistahBonzai]

It is very easy to change audio or video renderers and it even has a built in filter merit editor that still works in Windows 8.1. It sounds like you don't know how filter merits work? All the examples you state are not something you configure in Zoom Player (without smart play) but something you configure in ffdshow or AC3filter or adjust the merits so the filters you want are loaded. e.g. it isn't Zoom Player that sets the audio format, it is the audio decoder.
It was the filter merits that were misbehaving - The Audio was set to some rather odd value like 800.. rather than 008 (or 004..) and wouldn't stay changed to a lesser value. Setting the merit for AC3 looked OK but going back showed it had reverted. I spent considerable time trying to discover if I was missing some not obvious step in saving the value change and eventually discovered that although I could change it the setting choices were a slightly lighter grey than other settings menus.

In this screen crop (http://www.mediafire.com/view/9niz4063vakdd82/zoom_-_please_upgrade.PNG) you can see that under configure/status it claims it is enabled however clicking on "advanced" says otherwise and prompts me to upgrade. To my way of thinking it should have notified me of that limitation long before getting that deep into configurations

This doesn't happen on my system. Zoom Player will not resize the input before it gets to madVR, that isn't even possible as I understand it (ffdshow might but not Zoom Player). I don't know anything about MPC-HC's screen capture but it sounds like it happens before madVR if the image is captured at a different resolution than madVR is displaying.
Once again something has changed - I am sure I opened up the 4K video 3-4 times to verify, in fact, that 1920-1080 was the MadVR "Movie Resolution" that displayed using Zoom Player/MadVR. I have no idea how I mixed that up but I obviously did - MadVR now displays the native video (4K) size.

This resized full screen capture (http://www.mediafire.com/view/a4iaam3wqpmq83n/zoom_resized_filter_status.png) shows the filter properties. note that Zoom player audio filter appears to be non-negotiable.

The "Pro" version of Zoom Player may be altogether different (in showing a helpful status for what changes one can make/not make) but at this point I am skeptical. I would hate to go with the full version only to discover that inserting ffdshow filters (and avisynth) weren't possible - a full time limited version would be more helpful in an evaluation.

Edit: Sorry about OT diversion. Bottom Line is Zoom Player utilizes MadVR without obvious errors :)

[romulous]

a full time limited version would be more helpful in an evaluation.

That is exactly what Pro and Max are. The download gives you - I think - 20 or 30 days to try them out, and if you like them, you can buy a licence at any time.

For the Zoom Audio Filter - if you disable the following option (need to reload the currently playing media file with Shift+J to make it take effect), does that vanish?
Playback : Audio > Enable Internal DSP Filter (Equalizer / Audio Stream Switching / Audio Synchronization)

(Sorry for the O/T madshi, this is more appropriate over on the Zoom Player forums)

[oddball]

Has anyone managed to get MadVR working with that Virtu passthru (iMode) Intel stuff on a motherboard? It would be nice if I could get that working to save on power when not playing games/video. I only managed to get it working on the crabby default renderer.

[aufkrawall]

Since OpenGL 4.3, it also supports GPU Compute.
Could it be more performant than D3D11 Compute?

[nevcairiel]

Since OpenGL 4.3, it also supports GPU Compute.
Could it be more performant than D3D11 Compute?

You can't mix and match D3D and OpenGL, the overhead would probably kill any potential benefits.

[aufkrawall]

And if madVR was entirely OpenGL?

[nevcairiel]

And if madVR was entirely OpenGL?

Then it would lose quite a bunch of its features which are tightly coupled to D3D, and it would mean a huge rewrite.

And all for what, a potential non-benefit? :)
Until you know its *a lot* faster, its just not worth it.

[aufkrawall]

That's why I ask. ;)
OpenCL NNEDI3 scaling wasn't very usable for me on a R9 290 and with a GTX 670, D3D11 Error Diffusion isn't either with high fps or 4k.
That's why it doesn't seem completely pointless to me to find out about OGL.
Most modern online renderers are being ported to OGL, it's getting more popular for game developers.

[mark0077]

I'm not sure why the "4:4:4" text is lighter than the background - I suppose it just depends on how the TV blends things for the chroma subsampling - but that looks like 4:2:0 since both the "4:2:2" text and the background are blurred.

Hi 6233638, I have a similar issue on my Samsung Series 9 (LE46A956). In PC Mode I have always believed it supports 4:4:4 however with your png, 4:2:2 is seen fairly lightly, and 4:4:4 is seen relatively darkly. Only non PC Mode does 4:2:2 stand out and 4:4:4 can't be seen. Does this indicate its not truly 4:4:4 or does the way the pixels are layed out have an impact on how it should look when truly 4:4:4. If a picture is useful I can send on here or via PM as I do believe this display is 4:4:4.

[markanini]

What's the most convenient way of displaying a histogram? I'm not sure if I'm misusing the 2x/3x expanded sometimes...

[flashmozzg]

I think move to dx11 (12?) would be more realistic and beneficial for madVR in the long run.

[James Freeman]

Support for Windows XP ends on April 8th 2014 (http://windows.microsoft.com/en-us/windows/end-support-help), and the need for DirectX 9 along with it.
Maybe then madshi will rethink about upgrading to a later Directx as the basis, or not?

Actually, I don't care as long as it works.

[madshi]

OpenCL NNEDI3 scaling wasn't very usable for me on a R9 290
AMD driver issue (high OpenCL <-> D3D9 interop cost). Contact them and ask for a fix.

with a GTX 670, D3D11 Error Diffusion isn't either with high fps or 4k.
Kepler generally doesn't seem to have good compute performance. Switching to a different API won't help there. However, Maxwell seems to run Error Diffusion much faster than Kepler, at least it seems this way atm.

[aufkrawall]

I'm very confident about Maxwell as well, outstanding single precision performance in non-OpenCL apps.
But from what I see in games, Kepler isn't that slow with Compute. That may be because games aren't using this excessively, but for example, in this DX11 CS benchmark, GK110 is much faster than any GCN GPU (scroll down to CS Raytracing):
http://www.computerbase.de/2013-02/test-nvidia-geforce-gtx-titan/12/
Edit: Ok, it's just GK110. Smaller Kepler GPUs really seem to be weak regarding this. Nonetheless, a Tahiti XT GPU has much more FLOPs than a GK104.

Didn't you say that also on AMD, doing the dithering via Compute can be much slower than with OpenCL?

Edit: Btw: Everything I reported to AMD has been ignored so far. I think it would be more promising if the issue was debated by Anandtech etc.

[madshi]

Didn't you say that also on AMD, doing the dithering via Compute can be much slower than with OpenCL?
Nope, last time I compared with identical algorithms, Error Diffusion with DirectCompute was about 20% faster on my PC compared to OpenCL, and that even without taking interop costs into account.

Anyway, I really don't want to discuss which APIs or compute frameworks madVR should use.

[dvzzz]

I have an Intel 4600 IGPU on my Haswell and Benq 1070 projector and here is a problem that I ran into with MadVR.
I am setup:
-MadVR 0.87 with tweak applied to display 0-255
-Intel Driver is set to 0-255
-MadVR is set to 0-255
-Benq is set to PC-levels

When I play 1080p24 materials I get correct blacks according to AVS Cal Basic patterns, when I play Blurays in 1080i29/30 suddenly all blacks are washed out. It only happens on the 29/30 materials that are interlaced. Do you know what should be changed in my setup of MadVR to address it. I have lots of 1080i29 encoded videos with concerts and this is annoying to lose blacks.
Appreciate the help.

[madshi]

When I play 1080p24 materials I get correct blacks according to AVS Cal Basic patterns, when I play Blurays in 1080i29/30 suddenly all blacks are washed out. It only happens on the 29/30 materials that are interlaced. Do you know what should be changed in my setup of MadVR to address it. I have lots of 1080i29 encoded videos with concerts and this is annoying to lose blacks.
With "when I play Blurays in 1080i29/30" do you mean playing back 1080i29/30 content, deinterlaced to e.g. 1080p59/60? Or do you mean outputting this as 1080i29/30 to the display? Is deinterlacing turned on in madVR when the problem occurred and which display mode / refresh rate are you using in that situation?

I think it could either be a different display mode behaving differently - *if* you're using a different display mode for such kind of content.

Or more probably, when activating DXVA deinterlacing, the Intel GPU driver might do damage to the video, screwing up the black levels. Although, madVR should really detect and fix that for you. Anyway, try disabling deinterlacing (there are keyboard shortcuts for disabling/enabling deinterlacing on the fly in madVR). Does switching deinterlacing fix/unfix the black level problem?