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

[Zachs]

Probably because "Initialized()" was never called. It technically should be called whenever a Preset is used, but so far it's largely been ignored. Which was fine since it didn't really do anything, but then "ActiveOptions" was introduced, which was only ever intialized when "Intialized()" was called.

Edit: seems it's also initialized in the constructor, but only if "Options" isn't null, which it won't be because we're in the constructor.

Looks like we need a bit more discussion so let's continue on GitHub.

[Zachs]

Very nice work, now it's definitely smooth.
There's btw. no difference for me between FSE or windowed, even DX11 works absolutely glitch-free here.

Some questions:
What exactly does "improve chroma reconstruction" do?
I don't notice any difference whether it's off or on with render script chroma upscaling.

When we do chroma and luma upscaling via render script, main MPDN's scalers are still used for offset correction. Should bilinear be enough for this?

Unfortunately, I have a severe problem with render script level 8 API and my old script chain:
I'm just getting lots of errors with it with any file:
http://abload.de/thumb/mpdntdu7l.png (http://abload.de/image.php?img=mpdntdu7l.png)
My script:
http://pastebin.com/cpQMnqLC

Latest commit on GitHub fixes this issue.

[Zachs]

Wow, so after ages the Remote Control finally got an update :)
The remote control now supports video tracks, opening of playlist files and URLs and the interface has finally gotten a long overdue overhaul (no more fugly! :o)

Please note that the Remote will not work with extensions v1.22.0 or below.

For the full set of release notes see https://github.com/DeadlyEmbrace/MPDN_RemoteControl/releases/tag/v1.3.0.67

Or grab the release https://github.com/DeadlyEmbrace/MPDN_RemoteControl/releases/download/v1.3.0.67/MPDNRemoteControl.zip

Thank you so much!

The next MPDN release is going to be a massive step forward!

[Zachs]

I mentioned this in the Guru3D forums as I originally thought it was the display driver. A member there suggested I post here about it in case it's a compatibility issue with MPDN and Windows 10 TH2 build 10565+.

I have an issue that seems to be pretty random when playing MPDN in Windows 10 build 10565. It seems to cause a crash in quartz.dll (Directshow runtime) which in turn crashes the desktop explorer, which reloads quickly. If it happens to anyone else, it just looks like the taskbar is refreshing along with the icons (but it does show in event viewer). It has only happened since build 10565, which is why I believe it is Windows related. It is an Insider Preview, hence beta, so I wouldn't expect it to be perfect! I've already reported it through the Insider Preview feedback tool.

Thanks for the heads up!

[Zachs]

Very nice work, now it's definitely smooth.
There's btw. no difference for me between FSE or windowed, even DX11 works absolutely glitch-free here.

Some questions:
What exactly does "improve chroma reconstruction" do?
I don't notice any difference whether it's off or on with render script chroma upscaling.

When we do chroma and luma upscaling via render script, main MPDN's scalers are still used for offset correction. Should bilinear be enough for this?

On some systems windowed mode will cause a lot of DWM glitches so it's not a good base to test FluidMotion, which was why I asked for it to be tested in FSE mode only - doesn't mean it won't work in windowed mode.

"Improve chroma reconstruction" only works with MPDN's internal chroma scalers, so if you have render script chroma scaling, it has no effect.

What linear interpolation to use for offset correction really depends on how sensitive you are to chroma. I haven't tested this myself, so would appreciate it if you could give it a proper test to see if you could see any difference at all going to bicubic or higher. Since you know what to look for in chroma, you'd be a good candidate to give us some feedback.

[Shiandow]

"Improve chroma reconstruction" only works with MPDN's internal chroma scalers, so if you have render script chroma scaling, it has no effect.

That's not quite true, sure chroma scalers can override it, but the default behaviour is to use the same setting as MPDN's internal scalers.

[Zachs]

That's not quite true, sure chroma scalers can override it, but the default behaviour is to use the same setting as MPDN's internal scalers.

Oh yes that's right - Renderer.ConvertToRgb() has supported that for some time now... Totally forgot. :o

[aufkrawall]

Latest commit on GitHub fixes this issue.
Yes.
However, SuperRes is still broken. As soon as I select it (already renewed config, can access SuperRes' settings), I'm getting this:
http://abload.de/thumb/superresu3qxc.png (http://abload.de/image.php?img=superresu3qxc.png)

[Zachs]

We know what the problem is but it'll take some time to figure out how to do it right.

[Anima123]

Latest extension doesn't compatible with test version 3535. Is there another test version coming out soon?

[Zachs]

No. We are in the middle of a rather big change so don't use the latest for now.

EDIT: It should now work but keep in mind we're still working on it so it'll break from time to time.

[Anima123]

Thanks for the heads ups. It still brings red errors with no images when using with Fluid Motion right now, basically the texture is not the right type.

Edit: It also doesn't work if set passes of SuperRes to 8.

[Zachs]

Thanks for the heads ups. It still brings red errors with no images when using with Fluid Motion right now, basically the texture is not the right type.

Edit: It also doesn't work if set passes of SuperRes to 8.

Works fine here. You may need to clear your settings.

[Zachs]

MPDN v2.45 will be released soon so we'd appreciate it if you guys could give the latest APILevel8 branch a test to make sure there isn't any outstanding bugs.

Cheers.

[Anima123]

There's Error messages as follow:

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.

I'm using the latest test build 64-bit and latest APILevel8 branch, settings just cleared before start MPDN.

Edit: The script version with no latest [Playlist] changes works. I will test it with my own settings to see if there's any problems.

[Zachs]

Yes the latest playlist changes will require a new test build and it's a work in progress.

[Anima123]

With the last render script Level8 version that worked with 3535, I got the following error messages when increased the pass of SuperRes to 8, SuperRes with NNEDI3 32x16 as image doulber, both steps with EWA ChromaScaler.

https://www.dropbox.com/s/4qdy7u6fx0h5odq/ErrorRenderScript.png?dl=0

Edit: Once the error appears, there's no way to get it back normal with simply revert the settings.

[Zachs]

Can you post your script settings so we could try and replicate the problem?

[Zachs]

Just a heads up. Test build 3550 is up and is compatible with the latest level8 commits on GitHub.

[aufkrawall]

Yes, 3551 + latest render scripts look good, it accepts my old script without issues (just have to reconfigure SuperRes settings, but no errors).

Just two little things:
-main MPDN scaling settings aren't saved on close
-MPDN hangs when closing if a video file was opened and a render script like deband is active

[Olivier C.]

I have encountered a bug when tried to use ffdshow raw filter + Avisynth for video processing :

MPDN :
http://img15.hostingpics.net/thumbs/mini_137513mpdn.jpg (http://www.hostingpics.net/viewer.php?id=137513mpdn.jpg)

MPC-HC :
http://img15.hostingpics.net/thumbs/mini_613863mpchc.jpg (http://www.hostingpics.net/viewer.php?id=613863mpchc.jpg)

MPDN settings :
Render script : None

Without avisynth, i.e. if i enable sharpen filter within ffdshow raw filter, i can not see any difference. For example, i have enabled the post-processing only on a half part and set sharpness to Max value. It seems that MPDN can not manage properly this post-processing filter, but maybe i am wrong in the way i use it ?

[lanzorg]

I have encountered a bug when tried to use ffdshow raw filter + Avisynth for video processing :

MPDN :
http://img15.hostingpics.net/thumbs/mini_137513mpdn.jpg (http://www.hostingpics.net/viewer.php?id=137513mpdn.jpg)

MPC-HC :
http://img15.hostingpics.net/thumbs/mini_613863mpchc.jpg (http://www.hostingpics.net/viewer.php?id=613863mpchc.jpg)

MPDN settings :
Render script : None

Without avisynth, i.e. if i enable sharpen filter within ffdshow raw filter, i can not see any difference. For example, i have enabled the post-processing only on a half part and set sharpness to Max value. It seems that MPDN can not manage properly this post-processing filter, but maybe i am wrong in the way i use it ?

Why are you still using these old things?

[Olivier C.]

Why are you still using these old things?

Because :
- i am using temporal denoisers or other stuff available only in ffdshow / avisynth.
- on first page, i can see this :
"Each chain can be customized by adding postprocessing filters (e.g. AC3Filter, ffdshow, SVP)"

But maybe i am wrong ...

[Zachs]

The biggest difference from your screenshots is both players aren't getting the same input. Take a look at the video resolution.

[Zachs]

Yes, 3551 + latest render scripts look good, it accepts my old script without issues (just have to reconfigure SuperRes settings, but no errors).

Just two little things:
-main MPDN scaling settings aren't saved on close
-MPDN hangs when closing if a video file was opened and a render script like deband is active

Found the problem. Will be fixed in the next build.

[Zachs]

With the last render script Level8 version that worked with 3535, I got the following error messages when increased the pass of SuperRes to 8, SuperRes with NNEDI3 32x16 as image doulber, both steps with EWA ChromaScaler.

https://www.dropbox.com/s/4qdy7u6fx0h5odq/ErrorRenderScript.png?dl=0

Edit: Once the error appears, there's no way to get it back normal with simply revert the settings.

Can you try the latest source on GitHub? It should now be fixed.

[Anima123]

Can you try the latest source on GitHub? It should now be fixed.

It seems fixed. Thank you developers.

[Zachs]

It seems fixed. Thank you developers.

Fantastic! Thank you for testing.

[Olivier C.]

The biggest difference from your screenshots is both players aren't getting the same input. Take a look at the video resolution.
You are right. The reason is that the avisynth script is supposed to denoise and upscale the movie.

[ryrynz]

Without avisynth, i.e. if i enable sharpen filter within ffdshow raw filter, i can not see any difference. For example, i have enabled the post-processing only on a half part and set sharpness to Max value. It seems that MPDN can not manage properly this post-processing filter, but maybe i am wrong in the way i use it ?

Zach I'm wondering if this is related to the image corruption I experience with Avisynth and NNEDI3.

[Zachs]

Not sure but what can be certain from the two screenshots is ffdshow isn't feeding the same video to MPDN.

[Zachs]

Hi guys

Test build 3555 is now available. Would appreciate if the testers could give it a proper test. We should be quite close to an official release.

Cheers.

[aufkrawall]

Seems to be fine.

Btw: I really don't notice any difference when choosing any algorithm for offset correction.
I scaled a problematic image from 720p to WQHD with SuperRes (NNEDI3 for chroma and luma) and there is basically no difference whether nearest neighbor or Spline is used for offset correction.
So I think it just should be set to bilinear when doing upscaling fully via render scripts. It's cheap and you are someway on the safe side, compared to NN.

[Zachs]

BTW could you post your updated script settings with the different performance presets?
I'm planning to release the new MPDN soon and would like to link to it in the OP. Something like your previous post would do just fine.

[aufkrawall]

Here they are:
http://pastebin.com/3XDMsaei
I made some small changes, e.g. replacing second NNEDI3 pass for highend GPU quadrupling with SuperRes super-xbr.

[Olivier C.]

The biggest difference from your screenshots is both players aren't getting the same input. Take a look at the video resolution.

It's the same video, the only difference between MPDN and MPCHC is that the video is denoised and upscaled on one side and doesn't processed at all on the other side.

But i understand your comment so i have reproduced it again and this time with the same resolution on OSD and with Mediainfo properties in order to convince you this time :) :

MPDN :
http://img15.hostingpics.net/thumbs/mini_799622mpdn.jpg (http://www.hostingpics.net/viewer.php?id=799622mpdn.jpg)

MPCHC :
http://img15.hostingpics.net/thumbs/mini_945502mpchc.jpg (http://www.hostingpics.net/viewer.php?id=945502mpchc.jpg)


Thanks Zachs for your awesome player

[Zachs]

Is this with the latest test build or the latest stable release? If it's not the test build could you give that a try because there was a related bug that was fixed some time ago but hasn't yet made it into the stable build.

EDIT: The test build I mentioned has now been released so once you've updated your current version, please test it again to see if the problem still occurs.

[Zachs]

Here they are:
http://pastebin.com/3XDMsaei
I made some small changes, e.g. replacing second NNEDI3 pass for highend GPU quadrupling with SuperRes super-xbr.

Thanks! Oh and MPDN internal scaler settings for each of the presets would be nice too.

[Zachs]

The following is a collection of render script presets that you can use if you're not familiar with the render script system, or you just wanted something to start off with. Simply copy and paste the text in the link and paste them into the Script Group renderscript.

By aufkrawall:


3 presets (for low / mid / high end GPUs) of scaling and postprocessing
http://pastebin.com/3XDMsaei

Please feel free to contribute your own presets and I'll add them to the collection.

[Olivier C.]

Is this with the latest test build or the latest stable release? If it's not the test build could you give that a try because there was a related bug that was fixed some time ago but hasn't yet made it into the stable build.

EDIT: The test build I mentioned has now been released so once you've updated your current version, please test it again to see if the problem still occurs.

Yes, it was with the latest build.

I checked with the fresh latest build (2.45.0.3555) and now it works like a charm ! Great job Zachs !

Maybe if i may say one little thing, i noticed that the input resolution is correct until i update the upscale factor while the video is playing. In this case, the input resolution is not refreshed (OSD, Ctrl+J).
I have to admit that it's a minor thing.

Thanks a lot Zachs (and others) for your (very) great job !

[Zachs]

Maybe if i may say one little thing, i noticed that the input resolution is correct until i update the upscale factor while the video is playing. In this case, the input resolution is not refreshed (OSD, Ctrl+J).
I have to admit that it's a minor thing.

Bear in mind that MPDN doesn't support input resolution changes in the middle of playback - you'll need to reopen the file.

[burfadel]

Under Windows 10 Insider Preview build 10565, along with the recent update that pushed the build decriptor to 10565.41737.amd64fre.th2_release_flt.151020-1357 (notice the 20 Oct stamp there), MPDN crashes when changing media files. This did not occur before the update. I tried resetting MPDN and using default settings, changing the decoder settings, colour space etc, nothing stops it from crashing when changing media files.

I tried MPC-HC 64-bit version and it changes media files without hassle, only MPDN is affected.

I also tried the latest nightly for MPC-HC 64-bit, tweaked all the settings, using the latest madVR with settings tweaked, custom shaders, and so on, and no matter what I tried I can not get MPC-HC to crash like MPDN does when changing media files (occurs even with a complete fresh install of MPDN). There must be something in the way that MPDN changes media files that something in the latest Windows 10 TH2 Insider Preview build with the aforementioned update that it does not like!

I suspect that maybe there is a long standing bug in MPDN that has only become apparent with a change in the way Windows does stuff, or a bug in Windows that affects the way MPDN does this, which is different to how MPC-HC etc works? In any case, reopening MPDN after each file is a little annoying!

The bug I mentioned earlier with quartz.dll, or what was causing quartz.dll to crash, has been fixed :).

[Anima123]

I don't use Fluid Motion a lot, but I just did a test with the latest version and it still has increasing number of delayed frames when OSD is off.

[ryrynz]

I don't use Fluid Motion a lot, but I just did a test with the latest version and it still has increasing number of delayed frames when OSD is off.

I'm not seeing that here, using build 3555 with either D3D9, 10.1 or 11 in windowed and exclusive modes on Intel HD graphics.

[Anima123]

I'm not seeing that here, using build 3555 with either D3D9, 10.1 or 11 in windowed and exclusive modes on Intel HD graphics.

Here's on nVidia 880M Optimus, the latest release builds with render script chain applied.

[ryrynz]

Here's on nVidia 880M Optimus, the latest release builds with render script chain applied.

Not seeing it on my 750 Ti either. It is all D3D versions and both windowed and FSE modes? If you disable the renderscript does it still occur?

[Zachs]

It's a problem with Optimus. There's not much that could be done because of the way the technology works. The old version never reported delayed frames when fluid motion is active. The new one does.

[Zachs]

Under Windows 10 Insider Preview build 10565, along with the recent update that pushed the build decriptor to 10565.41737.amd64fre.th2_release_flt.151020-1357 (notice the 20 Oct stamp there), MPDN crashes when changing media files. This did not occur before the update. I tried resetting MPDN and using default settings, changing the decoder settings, colour space etc, nothing stops it from crashing when changing media files.

Does this happen with the older build as well (i.e. API level 7)?

[aufkrawall]

Thanks! Oh and MPDN internal scaler settings for each of the presets would be nice too.
Settings for main MPDN:
low end GPU:
chroma and luma up: bilinear
chroma and luma down: bicubic sharpness 50
(enhanced chroma reconstruction off)

mid range:
chroma and luma up: bilinear
chroma and luma down: spline4 AR

high end:
chroma and luma up: bilinear
chroma and luma down: spline4 AR
(reduction of XySubFilter aliasing turned on)

Since there is no proof that a better scaler for offset correction really makes a difference, imho we shouldn't recommend to burn performance until we might know better one day.

[aufkrawall]

There might be still something a bit weird with fluid motion: OSD reports me delayed frames when it is closed while I'm playing 59,94fps@75Hz, e.g. this video:
http://www.oc-burner.de/ftp/Videos/juddertest/juddertest_59,94.mp4
I'm not sure if this is just a OSD bug or if it's really less smooth, I'm not used to 60Hz.