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

[Zachs]

i just wanted to test for audio differences.

but i run into my general audio issues again this time way worse than usually.
the first thing that changed is that the fullness is not the problem it is now staying at stable 90%.
but now i see the ref clock changing when i hear an audio issue and some times a frame is delayed at the same time. the refclk deviations is skyrocketing after that. i even saw numbers like 0.35 % after 60 sec playback.
that's without rate tuner. but i got this issue with it too.

of cause i don't have that issue with madVR.

Without the rate tuner, MPDN is a slave to the ref clock which is usually derived from the audio card by the audio renderer. As such, video smoothness is affected when you get an audio problem. TBH I'm not sure what could be the problem seeing as I can't replicate your problem even on my very low end Atom tablet, or my very old and dusty Core2Duo laptop with ATI Mobility Radeon X2300. My HTPC runs a Creative X-Fi and that has never encountered any problems too.

mpdn+lav video intel quciksync hardware gpu deinterlacing mpeg2 vob files = image gets stuck..:(
mpc+madvr+lav video intel quciksync hardware gpu deinterlacing mpeg2 vob files =no problem......

I personally watch a lot of MPEG2 interlaced materials with MPDN and haven't had any problem. With madVR are you sure you're using LAV's quicksync decode+deint hardware and not madVR's DXVA decode?

[ryrynz]

I notice some render script examples get installed with the latest release, is that intentional? Could be tider if there was an option to install 'em.

[Zachs]

That wasn't intentional. I'll have a look!

EDIT: Extensions v1.6.1 should have the examples removed.

[Magik Mark]

Zachs,

How do i put 1080p & 720p videos both 24fps in the video type found in the rate tuner? Thanks

[Zachs]

24.000fps or 23.976?
If it's the latter it's "23p"

[Magik Mark]

24.000fps or 23.976?
If it's the latter it's "23p"

Is this the right format:

w1920 h1080 p23? Gives me an error

[Zachs]

There's no need to specify height or width if you want to match all p23 materials. If you want to match 1080p 23.976, use w1920p23. w1280p23 for 720p 23.976hz materials.

[Magik Mark]

There's no need to specify height or width if you want to match all p23 materials. If you want to match 1080p 23.976, use w1920p23. w1280p23 for 720p 23.976hz materials.

Thank you that did it for tv shows. I'm getting a lot of frame drops when watching movies. This is really odd. The frame rate & resolution are the same as that of tv shows. Its really moving a little bit fast for both audio & video. Like the chipmunks

[Zachs]

Check your render times. Are they the same between the two?

[Magik Mark]

Check your render times. Are they the same between the two?

There is a difference of 2ms. When I switch FSE tv shows are behaving the same as the movies. AUdio & video are moving fast. All I did is copy the speed% found in the calculator

[Zachs]

You can't simply copy the values. If you want it to match 23.976hz of your display refresh rate, you need to first switch to that mode. Otherwise you're asking it to play at whatever your refresh rate was when you opened the calculator.

I will write a guide on how to use it when I get some time.

[ryrynz]

Its really moving a little bit fast, like the chipmunks

You can't simply copy the values.

:stupid: :D

[Zachs]

A short guide using 23.976FPS materials, assuming you have a monitor / TV capable of 24 input.

First, set your display refresh rate to 24Hz.
Run your 23.976FPS video for 5 minutes (make sure you don't seek / pause -- just let it run continuously).
Once you have let it run for 5 minutes, open the calculator with the video still running.
At this point, the calculator will give you very accurate values which will make your 23.976FPS input run at exactly your display refresh rate.

If you have a display capable of 25Hz, you can use the same steps as above to run your 23.976FPS input as 25FPS so you don't get any dropped / repeated frames. Most TVs don't support 25Hz but can do 50Hz. In which case, all you have to do is manually divide your Display Rate value by 2 in the calculator. This will give you 25FPS on a 50Hz refresh rate.

What this means is, you can disable fluid motion and still have perfectly fluid video playback. For the 24Hz refresh rate case, you can now enable your TV's FRC feature. With absolutely no repeated / dropped frames, your TV's FRC feature won't interpolate the wrong frames - IOW the FRC feature will work in its most ideal manner.

[ryrynz]

Would it be perfectly? or could it stand to do an automatic correction every say 5/10 minutes or something, as going by the OSD the there is a bit of fluctuation going on.

[bacondither]

I have found a strange bug.

When testing my adaptive sharpen shader in a script chain with deband and the video was displayed at original size the image flickers heavy with a pattern of fast dark and normal frames.

I have located the core of the problem to outputting a float3 as a float4 with padding for the w component which results in flickering.
//"flicker.hlsl"
sampler s0 : register(s0);

float4 main(float2 tex : TEXCOORD0) : COLOR
{
float3 c0 = tex2D(s0, tex).rgb;

return c0.rgbb;
}

Example script chains with flickering:
deband --> flicker.hlsl
deband --> resizer --> flicker.hlsl
flicker.hlsl --> deband
chroma scaler --> flicker.hlsl

Flickering only occurs when the image is at the original size, if the image is scaled the flickering goes away.

"flicker.hlsl" by itself in the chain gives no problems at all.

Tested with MPDN v2.31.0 and extensions v1.6.1, all the settings in MPDN is the default.
AMD GPU running driver version 14.4.

[Anime Viewer]

You can't simply copy the values. If you want it to match 23.976hz of your display refresh rate, you need to first switch to that mode. Otherwise you're asking it to play at whatever your refresh rate was when you opened the calculator.

I will write a guide on how to use it when I get some time.

That might explain why when I originally tested Rate Tuner it was giving me reports of ~249%-250% even when I clicked the calculator and input 23.976 and 24 in the display rate section of the calculator.

This raises a problem I've noticed. I have two displays: one that is the default notebook screen which supports 59hz and 60hz only, the other is a TV that supports 23hz,24hz, 59hz, and 60hz. I used to be able to go into Windows screen resolution, pick the TV under display, click the Advanced Settings link, List All Modes, select the mode I want to run in (ex: 1920x1080, True Color (32bit), 23Hertz, click ok and it would switch the tv screen to that setting which I could confirm with the CTRL+J listing for display as the video would run. However now while it looks like it took the setting in Windows display modes (if I go back in the one I chose is still highlighted/selected MPDN CTRL+J inside the video reports the display as 59hz regardless. The videos I've tested are 23.976hz source files. I've tested with and without Display Changer Activate automatic display refresh rate changer checked, and still MPDN OSD is reporting the display as 59hz.
Any thoughts as to why the OSD is reporting 59hz when playing videos on the TV regardless or what hz I've selected for the screen to run in? (I have both "Use highest" and "Activate only" both unchecked when testing in Display Changer). Is it possible the screen is changing, but the OSD isn't reporting it right?

[Shiandow]

I have found a strange bug.

When testing my adaptive sharpen shader in a script chain with deband and the video was displayed at original size the image flickers heavy with a pattern of fast dark and normal frames.

I have located the core of the problem to outputting a float3 with padding for the z component which results in flickering.

[...]

Example script chains with flickering:
deband --> flicker.hlsl
deband --> resizer --> flicker.hlsl
flicker.hlsl --> deband
chroma scaler --> flicker.hlsl

Flickering only occurs when the image is at the original size, if the image is scaled the flickering goes away.

"flicker.hlsl" by itself in the chain gives no problems at all.

Tested with MPDN v2.31.0 and extensions v1.6.1, all the settings in MPDN is the default.
AMD GPU running driver version 14.4.

That's probably caused by the alpha channel. MPDN doesn't like it when it is not 1. Some processing fixes the alpha channel again so it only causes problems sometimes.

[bacondither]

That's probably caused by the alpha channel. MPDN doesn't like it when it is not 1. Some processing fixes the alpha channel again so it only causes problems sometimes.

Thanks! Setting the w channel to 1 seems to work.

[huhn]

Without the rate tuner, MPDN is a slave to the ref clock which is usually derived from the audio card by the audio renderer. As such, video smoothness is affected when you get an audio problem. TBH I'm not sure what could be the problem seeing as I can't replicate your problem even on my very low end Atom tablet, or my very old and dusty Core2Duo laptop with ATI Mobility Radeon X2300. My HTPC runs a Creative X-Fi and that has never encountered any problems too.

i only have this problem with MPDN. EVR, foobar2000, computer games and web browser they all work totally fine.

and the next thing is if MPDN is the slave to the audio clock why is audio distorted and not a frame repeated/dropped? in best case i get a frame delay with audio distortion in most cases just audio distortion even through MPDN shows a frame repeat of a couple of secs with clock deviations of 0.35 %.

and of cause is no real work for my HTPC to use MPDN.

[Magik Mark]

A short guide using 23.976FPS materials, assuming you have a monitor / TV capable of 24 input.

First, set your display refresh rate to 24Hz.
Run your 23.976FPS video for 5 minutes (make sure you don't seek / pause -- just let it run continuously).
Once you have let it run for 5 minutes, open the calculator with the video still running.
At this point, the calculator will give you very accurate values which will make your 23.976FPS input run at exactly your display refresh rate.

If you have a display capable of 25Hz, you can use the same steps as above to run your 23.976FPS input as 25FPS so you don't get any dropped / repeated frames. Most TVs don't support 25Hz but can do 50Hz. In which case, all you have to do is manually divide your Display Rate value by 2 in the calculator. This will give you 25FPS on a 50Hz refresh rate.

What this means is, you can disable fluid motion and still have perfectly fluid video playback. For the 24Hz refresh rate case, you can now enable your TV's FRC feature. With absolutely no repeated / dropped frames, your TV's FRC feature won't interpolate the wrong frames - IOW the FRC feature will work in its most ideal manner.

I have 60hz Monitor. Does it mean I have to divide the refresh rate in the calculator by 2.5 in order to play 23.976hz or 24 hz video smoothly?

[huhn]

I have 60hz Monitor. Does it mean I have to divide the refresh rate in the calculator by 2.5 in order to play 23.976hz or 24 hz video smoothly?

you should use fluid motion to play it smooth not this new feature.

[Zachs]

Would it be perfectly? or could it stand to do an automatic correction every say 5/10 minutes or something, as going by the OSD the there is a bit of fluctuation going on.

Perfectly. By fluctuation you mean the actual average FPS presented? If so, that's just because of how the average is calculated. With the method I wrote, I've managed to watch a 2.30hr movie without a single repeated/dropped frame with my TV @23.978Hz.

Any thoughts as to why the OSD is reporting 59hz when playing videos on the TV regardless or what hz I've selected for the screen to run in? (I have both "Use highest" and "Activate only" both unchecked when testing in Display Changer). Is it possible the screen is changing, but the OSD isn't reporting it right?

Which monitor did you pick for your "Presentation" graphics adapter? When you move MPDN's window to your TV, does the OSD say "Current Display"? If it says "Other Display", you've set it up to use your laptop screen.

Thanks! Setting the w channel to 1 seems to work.

As long as the final result sets the w channel to 1, MPDN will be happy. Any other values will result in undefined behaviour. You can still use w channel to pass data from an earlier pass in the middle of the processing chain, as long as the w channel gets set back to 1 before returning the result texture to MPDN.

i only have this problem with MPDN. EVR, foobar2000, computer games and web browser they all work totally fine.

and the next thing is if MPDN is the slave to the audio clock why is audio distorted and not a frame repeated/dropped? in best case i get a frame delay with audio distortion in most cases just audio distortion even through MPDN shows a frame repeat of a couple of secs with clock deviations of 0.35 %.

and of cause is no real work for my HTPC to use MPDN.

Have you tried playing just the audio stream to see if MPDN still causes your audio issue? Can you also try playing audio in the background and then make MPDN distort the audio to see if the audio playing in the background is distorted as well?

The point I was trying to make was I won't be able to debug it without you spending some time on it since I (or anyone else for that matter) can't replicate your issue on any of my systems. And I gather from your replies you weren't interested at all in doing so.

I have 60hz Monitor. Does it mean I have to divide the refresh rate in the calculator by 2.5 in order to play 23.976hz or 24 hz video smoothly?

You should not divide it by anything other than an integer number. In your case, if you really want to use the rate tuner, you could look into overclocking your monitor and then get it as close as 24Hz*3 as possible. You then divide the refresh rate by 3. One of my monitors also only does 60Hz but I managed to overclock it to 68.5Hz. This means I end up with 22.833Hz when playing 23.976FPS materials. Your video will play around 5% slower but you won't notice the difference (PAL conversion works the opposite by doing a ~4% speedup). You should also use MPC-BE Audio Renderer (included with MPDN's latest installer) with WASAPI exclusive mode output when using the rate tuner. WASAPI audio renderers can set the sample rate of your audio endpoint directly.

[Magik Mark]

Zachs is there a way to reverse engineer reclock then port it to MPDN?

[Zachs]

User guide for Rate Tuner - https://github.com/zachsaw/MPDN_Extensions/wiki/Rate-Tuner

[Zachs]

Zachs is there a way to reverse engineer reclock then port it to MPDN?

There's no need to - reclock doesn't do magic, contrary to what most people believe (you do know reclock won't make 24 fps play buttery smooth at 60hz right?). Read the rate tuner user guide I just posted above.

[Anime Viewer]

Which monitor did you pick for your "Presentation" graphics adapter? When you move MPDN's window to your TV, does the OSD say "Current Display"? If it says "Other Display", you've set it up to use your laptop screen.


You're correct it was listing "Other Display" when playing the video on the TV screen. I used to have the TV as the presentation screen, but changed that after I noticed that I would see the screen flash black (switching modes) sometimes when I was switching between browser tabs (and other tabs on Windows taskbar - it was probably kicking MPDN out of exclusive mode) when running in exclusive mode with the tv as the presentation screen. At the time I switched the notebook to be the presentation screen which would allow exclusive mode on the notebook (so I wouldn't give it up entirely) and let the TV run in full screen windowed mode. I've now changed it back to use the TV as the presentation screen, and just let them both run in full screen windowed mode. The side effect of this is that now it seem when running a 23hz video on the notebook screen it says the display is running at 23hz too (even though the notebook screen is only supposed to support 59hz and 60hz rates).

On a related note there is probably a way to do away with forcing one screen to be the presentation mode and the other to not. I find in MPC with MadVR I can have video playing in Exclusive fullscreen mode on one screen, drag it over to the other screen and maximize it, and that screen will also display as running in exclusive mode. Also MPC/madVR is locking the secondary screen to running at the presentation screens refresh rate. If madVR can do it then MPDN should have the ability to run in such a state too.

[Magik Mark]

User guide for Rate Tuner - https://github.com/zachsaw/MPDN_Extensions/wiki/Rate-Tuner

Thanks a lot! Works great with SVP

I there a way to port SVP settings to MPDN natively? This would be great for the 64bit version. Currently SVP is only 32bit

[Zachs]

You're correct it was listing "Other Display" when playing the video on the TV screen. I used to have the TV as the presentation screen, but changed that after I noticed that I would see the screen flash black (switching modes) sometimes when I was switching between browser tabs (and other tabs on Windows taskbar - it was probably kicking MPDN out of exclusive mode) when running in exclusive mode with the tv as the presentation screen. At the time I switched the notebook to be the presentation screen which would allow exclusive mode on the notebook (so I wouldn't give it up entirely) and let the TV run in full screen windowed mode. I've now changed it back to use the TV as the presentation screen, and just let them both run in full screen windowed mode. The side effect of this is that now it seem when running a 23hz video on the notebook screen it says the display is running at 23hz too (even though the notebook screen is only supposed to support 59hz and 60hz rates).

On a related note there is probably a way to do away with forcing one screen to be the presentation mode and the other to not. I find in MPC with MadVR I can have video playing in Exclusive fullscreen mode on one screen, drag it over to the other screen and maximize it, and that screen will also display as running in exclusive mode. Also MPC/madVR is locking the secondary screen to running at the presentation screens refresh rate. If madVR can do it then MPDN should have the ability to run in such a state too.

Yes this has been on my todo list for a long time but it remains low priority as you could only run one screen in exclusive mode at any given time anyway.

EDIT: BTW, it says the "Other display" is running at 23Hz when you have MPDN on your notebook screen.

EDIT2: You can make MPDN use display 1 or 2 via command line "adapter" option. You can simply create 2 shortcuts - one to launch MPDN for your TV, another when you want to view on your notebook screen.

[ryrynz]

Zach could you allow the rate tuner to only activate either when certain refresh rates are active or tie it in to the display changer?

[Zachs]

I'll have a think about that. It doesn't matter if it's activated even when your refresh rate hasn't been set correctly though does it?

[huhn]

Have you tried playing just the audio stream to see if MPDN still causes your audio issue? Can you also try playing audio in the background and then make MPDN distort the audio to see if the audio playing in the background is distorted as well?

EDIT:i tried to reproduce this while MPC-HC was running muted and at 5% and i still get the issue but i didn't get a delayed frame in 10 min testing i tested if a running and muted MPDN instance has audio effects on MPC-HC and a muted instance of MPDN most likely had an effect. heard an huge audio issue and a lot of frames where dropped at one point in madVR could be an issue of two program using GPU but unlikely in this case. double check MPC-HC only afterwards and no issue what so ever.

my clock deviation at 23p is 0.00333 to 0.00340 usually after ~3 mins this doesn't sound like a to bad audio clock in general.


EDIT:
direct audio:
15 mins audio only didn't present me any issue.

mpc-hc audio renderer in MPDN:
audio video playback:
extrem audio distortions combined with delayed repeated and dropped frames from time to time way worse than with directaudio but the ref clock stays at a stable 23.976 something clock unlike direct audio.
second try ref clock was at 23.97602-23.97610 after 40 sec no audio problems but delayed frames.
12 min audio only didn't present me any issues.
2 audio programs at the same time is pretty much pointless with exclusive mode wasapi or?

The point I was trying to make was I won't be able to debug it without you spending some time on it since I (or anyone else for that matter) can't replicate your issue on any of my systems. And I gather from your replies you weren't interested at all in doing so.

i pointed the old issue at the fullness but this can't be the issue anymore. the old issue need some concentrated hearing. i don't even kno if this is the same issue or i'm "lucky" that it triggers this hard. and sometimes it doesn't trigger at all and the ref clk deviations stays normal.

[huhn]

bug:
open a flac file most likly all audio only files:

===================================

An unexpected error 'System.NullReferenceException' has occurred.

------------------------------
Error Type = System.NullReferenceException
Error Message = Object reference not set to an instance of an object.
Error Source = Mpdn.Extensions
Error Site = Void PlayerStateChanged(System.Object, Mpdn.PlayerStateEventArgs)
Error occurred = at Mpdn.Extensions.PlayerExtensions.DisplayChanger.PlayerStateChanged(Object sender, PlayerStateEventArgs e)
at System.EventHandler`1.Invoke(Object sender, TEventArgs e)
at Mpdn.PlayerControl.OnPlayerStateChanged(PlayerState playerState, PlayerState oldState)
at MediaPlayerDotNet.MainForm.<>c__DisplayClass3d.<SetPlayerState>b__3c()
at Mpdn.Utilities.ControlHelpers.HandleAsync()
OS Version = Microsoft Windows NT 6.3.9600.0

...stuff...


===================================

Object reference not set to an instance of an object. (Mpdn.Extensions)

------------------------------
Program Location:

at Mpdn.Extensions.PlayerExtensions.DisplayChanger.PlayerStateChanged(Object sender, PlayerStateEventArgs e)
at System.EventHandler`1.Invoke(Object sender, TEventArgs e)
at Mpdn.PlayerControl.OnPlayerStateChanged(PlayerState playerState, PlayerState oldState)
at MediaPlayerDotNet.MainForm.<>c__DisplayClass3d.<SetPlayerState>b__3c()
at Mpdn.Utilities.ControlHelpers.HandleAsync()

[Zachs]

EDIT:i tried to reproduce this while MPC-HC was running muted and at 5% and i still get the issue but i didn't get a delayed frame in 10 min testing i tested if a running and muted MPDN instance has audio effects on MPC-HC and a muted instance of MPDN most likely had an effect. heard an huge audio issue and a lot of frames where dropped at one point in madVR could be an issue of two program using GPU but unlikely in this case. double check MPC-HC only afterwards and no issue what so ever.

my clock deviation at 23p is 0.00333 to 0.00340 usually after ~3 mins this doesn't sound like a to bad audio clock in general.


EDIT:
direct audio:
15 mins audio only didn't present me any issue.

mpc-hc audio renderer in MPDN:
audio video playback:
extrem audio distortions combined with delayed repeated and dropped frames from time to time way worse than with directaudio but the ref clock stays at a stable 23.976 something clock unlike direct audio.
second try ref clock was at 23.97602-23.97610 after 40 sec no audio problems but delayed frames.
12 min audio only didn't present me any issues.
2 audio programs at the same time is pretty much pointless with exclusive mode wasapi or?



i pointed the old issue at the fullness but this can't be the issue anymore. the old issue need some concentrated hearing. i don't even kno if this is the same issue or i'm "lucky" that it triggers this hard. and sometimes it doesn't trigger at all and the ref clk deviations stays normal.

bug:
open a flac file most likly all audio only files:


Garteal fixed it! :thanks:

Just so I understand your test results complete, can you confirm if I got the following correct?

Directsound
MPDN audio only = no distortion
MPDN audio+video = distortion
MPDN audio+video (muted) with MPC-HC audio+video = distortion
MPDN audio+video (muted) with MPC-HC audio only = distortion?
MPDN audio only with MPC-HC audio only = no distortion?

As for WASAPI exclusive mode, you mentioned you never got it to work with MPC-HC either, is that correct?

Can you try running DPC latency checker to see if you get any obvious problems when distortion occurs? Compare that with DPC latency when there's no distortion, do you see any difference?

[Magik Mark]

Zachs, Is it possible to use DXVA on IMAGE upscaling & downloading? I think this would help low to medium end users

[huhn]

Garteal fixed it! :thanks:

Just so I understand your test results complete, can you confirm if I got the following correct?
Directsound
MPDN audio only = no distortion
MPDN audio+video = distortion
yes
MPDN audio+video (muted) with MPC-HC audio+video = distortion
yes
MPDN audio+video (muted) with MPC-HC audio only = distortion?
MPDN audio only with MPC-HC audio only = no distortion?
didn't tested this combinations. but i can do this test tomorrow.
As for WASAPI exclusive mode, you mentioned you never got it to work with MPC-HC either, is that correct?
i got issue with wasapi but generally i got audio just rarely some distortions. my guess is that i got distortion when a frame drop/repeate should happen but never did. i just used direct audio without any issue. but i can't remember such big issues.
there is a file that sounds very noisy when used with WASAPI and my soundcard so i guess the WASAPI "driver" are simply broken and can be ignored. ASIO works like a charm on the other hand.
Can you try running DPC latency checker to see if you get any obvious problems when distortion occurs? Compare that with DPC latency when there's no distortion, do you see any difference?

this program doesn't properly work on windows 8 and i'm on windows 10.
i used this program back at my windows 7 times with very low numbers.

i will try to record the issue later i may learn something from it.

[Zachs]

Zachs, Is it possible to use DXVA on IMAGE upscaling & downloading? I think this would help low to medium end users

It's on my todo list (as with a lot of other stuff) but I just haven't had any spare time to work on it.

EDIT: When I looked into it, I found it to be rather restrictive on what it could do. This hasn't changed AFAIK. For starters, it only accepts 8-bit input/output, which means it is worse in quality (colour correctness wise - i.e. you'll get banding issues, colour shift etc.) when compared to even hardware bilinear scaling with MPDN's "max performance" settings while requiring more processing power. Even on my low power budget device (e.g. Atom Tablet), I prefer MPDN's hardware bilinear (which has full 32-bit colour accuracy) on 10-bit FSE mode over DXVA. With 10-/16-bit sources, you essentially murder the quality when you use DXVA scaling.

[huhn]

MPDN audio+video (muted) with MPC-HC audio only = distortion?
MPDN audio only with MPC-HC audio only = no distortion?

tested and yes.

Can you try running DPC latency checker to see if you get any obvious problems when distortion occurs? Compare that with DPC latency when there's no distortion, do you see any difference?

i used LatencyMon 6.00: http://www.resplendence.com/downloads

obviously the issue: http://abload.de/img/bingogys3y.png

the high numbers come when the audio distortion appears.
i will let the program run in the background when i next time used madVR for a longer time like 2 h may take some time just to make sure madVR has no problem at all and the issue is limited to MPDN. but a short test with madVR didn't showed any issue.

[Zachs]

Looks like something in MPDN's video renderer (which has nothing to do with audio) is triggering an audio driver bug. 26ms from interrupt to process is definitely a big problem.

[huhn]

so i guess nothing to learn from recoding this issue. let's just hope this is not an general issue.

maybe someone with windows 10/windows 8.1 can do a test for this issue too.
you just have to run LatencyMon 6.00 in the background and watch an 24 min episode or something like that.

[huhn]

i get high numbers in madVR and foobar too (~4000). the only difference is that MPDN has audio issues and the rest not. so the driver are kinda broken or this software doesn't work properly on windows 10.

not sure if it is good to waste time on it.

[Magik Mark]

zach can we port super xbr in chroma upscaling? I'm getting nicer result compared to jincs in madvr

[Anima123]

I've tried both NEDI and Super-xBR with SuperRes, the new SuperRes result in a very pleasant picture quality with NEDI, from 1024x576 -> 1920x1080 videos.

Super-xBR is too sharp, the edges of faces sometimes 'harder' than with NEDI.

Also the new algorithm work great without the old rendering time increasing issue, which makes the MPDN an alternative to my choice of daily media players.

[madshi]

Just a heads-up: Downloaded the standalone version of MPDN and the extension binaries and sources into new folders (using the links from the first post), copied the extensions in the right folders. But now both 32bit and 64bit MPDN complain with "One or more extensions failed to load". I had deleted the old "AppData" MPDN folder. Tried emptying the PlayerExtensions folder, didn't help.

[ryrynz]

Just a heads-up: Downloaded the standalone version of MPDN and the extension binaries and sources into new folders (using the links from the first post), copied the extensions in the right folders. But now both 32bit and 64bit MPDN complain with "One or more extensions failed to load". I had deleted the old "AppData" MPDN folder. Tried emptying the PlayerExtensions folder, didn't help.

I've run into this a number of times and clearing out appdata and deleting the extensions folders before reinstalling has always fixed it.

I just downloaded latest x64 and x86 versions of MPDN and extensions from http://mpdn.zachsaw.com/Latest/ after completely uninstalling my old installation and I had no errors with either. You might want to try it again.

[madshi]

I've run into this a number of times and clearing out appdata and deleting the extensions folders before reinstalling has always fixed it.

I just downloaded latest x64 and x86 versions of MPDN and extensions from http://mpdn.zachsaw.com/Latest/ after completely uninstalling my old installation and I had no errors with either. You might want to try it again.
I did all that. But I found what the problem was:

When only using the MPDN standalone version + the binary extensions, everything works fine. But when also extracting the binary source code and copying the binary source code "Extensions" folder over to MPDN, the problem occurs. I guess I'm not supposed to do that? Still, seems like something to look into. The first post looks like I should download and "install" all 3 (standalone version, extension binaries and extension sources). Maybe it would make sense to move the "extension source" to the "Miscellaneous" section on the first page, to make it clearer which downloads are necessary to make the standalone version work with all extensions?

[Shiandow]

Also the new algorithm work great without the old rendering time increasing issue, which makes the MPDN an alternative to my choice of daily media players.

This is just getting weird now, but I'm glad that it works.

[Zachs]

I did all that. But I found what the problem was:

When only using the MPDN standalone version + the binary extensions, everything works fine. But when also extracting the binary source code and copying the binary source code "Extensions" folder over to MPDN, the problem occurs. I guess I'm not supposed to do that? Still, seems like something to look into. The first post looks like I should download and "install" all 3 (standalone version, extension binaries and extension sources). Maybe it would make sense to move the "extension source" to the "Miscellaneous" section on the first page, to make it clearer which downloads are necessary to make the standalone version work with all extensions?

I've changed the wording so hopefully that removes any ambiguities.

[madshi]

Ah, now I understand! I can either use the binary or the source version of the extensions, correct? If that's correct, then the wording should be clear now. Might still make sense to investigate why it gets broken if you install both the binary and source version? I mean, maybe someone installs the binary version first, then wants to switch to the source version, or something. Would be nice if that would still work.

[Zachs]

Yeah that's correct.

I might just remove the source version from the download section actually. Advanced users can always download them from github.

@Shiandow what do you think?

[Shiandow]

Yeah that's correct.

I might just remove the source version from the download section actually. Advanced users can always download them from github.

@Shiandow what do you think?

It might be best to remove them if it prevents confusion, although I'm not too happy with making it harder for people to try playing with the code. If possible I'd prefer if MPDN detected such a conflict and asked to remove the binaries (or the sources but it's hard to tell if those have been edited).

Can you install the sources with the installer by the way? Haven't had a chance to try it out yet.