madVR - high quality video renderer (GPU assisted)

[6233638]

so , what , this is a kodi and potplayer issue afterall? guys you confuse me, setting prerendered frames to 1 or 8 in madvr doesnt do diddly for the lagDon't do this - it's likely to cause lip-sync errors. (http://bugs.madshi.net/view.php?id=261)
I don't know about Kodi or Potplayer's OSD elements, but using anything other than the old Windowed path will be slow to respond to inputs - and it gets slower the more frames you present in advance. That's one of the reasons I have stuck with the old windowed path for so long.

[sexus]

yeah youre reply is abit late as you can see ;)

btw whats this windowed path youre talking about?

is this a lag free 2nd renderer that both potplayer and kodi could have madvr use for OSD elements? cause this is what we need , thanks to nevcairiel for the idea

btw ive got my flush settings set to :

flush
flush and wait
dont flush
flush

for both windows and exclusive , i recon these are right, and ive got both set to prerender frames to 8 as anything above that introduces frame skipping ive noticed aka present several frames in advance for exclusive mode with 16 frames does this, so its been disabled

[Kirk Lazarus]

Hi guys,
I use my HTPC with a 4K TV but If I set MadVR to "enable automatic fullscreen exclusive mode" I see black screen immediatly after the muose arrow vanish.

If I disable it I've no problem.

I've an AMD card and I exit at 2160@23p RGB 4:4:4

If I set my HTPC to output in 1920x1080 I can use "enable automatic fullscreen exclusive mode" without any problem.

My setup is:
Windows 8.1 pro
AMD R9 280X with Catalyst 14.12
MadVR 0.87.14

Sorry If my question is stupid, I'm very new about HTPC and MadVR.

Thank you

[macycat]

I noticed something odd. When I drag a window with a playing (or paused) video from one monitor to a second monitor, madVR crashes. Sometimes I get the following error:


madVR Error:

- creating Direct3D device failed (80070005)


When I select "enable windowed overlay" in general settings, I can move the window with the playing video from one monitor to the other with no crash.

I am using MPC-HC with Windows 7 64 bit and a GTX 670 video card.

[6233638]

btw whats this windowed path youre talking about?Disable Full-Screen Exclusive, Windowed Overlay, and the present several frames in advance option for Windowed Mode to use the old path.

[sexus]

funny cause ive never use any of those in my regular playback which im testing this OSD lag issue on, so the old path doesnt help with this

[huhn]

ok so ive reset my gpu queue to 24 and my cpu queue to 48 and posted nevcairiels answer to kodi dsplayers thread , hope they can fix this , matter of fact maybe madshi can help with this , to have madvr use a seperate renderer for OSD elements , since as of currently even on the latest madvr build there is atleast 4 frames of OSD lag on both kodi and potplayer , as stated by huhn


first of all the default for madVr is 16 CPU and 8 GPU.

and i'm not sure if the madVR osd has a lag of at least 4 frames. it looks like this with MPC-HC. the MPC-HC progress bar has no delay at all.

but the trick is it to not use the madVR OSD to render kodis OSD at all.

[Warner306]

first of all the default for madVr is 16 CPU and 8 GPU.

and i'm not sure if the madVR osd has a lag of at least 4 frames. it looks like this with MPC-HC. the MPC-HC progress bar has no delay at all.

but the trick is it to not use the madVR OSD to render kodis OSD at all.

I know madshi personally involved himself in integrating madVR into Kodi, so I don't know that you are going to get anywhere. Many people won't notice the lag at all. But if it bothers you a lot, you might have to use a separate player with Kodi as an alternative set-up.

Kodi was not designed with madVR in mind. I don't know of a more sophisticated media front-end that supports madVR, so what is there is a significant accomplishment. Player performance, in terms of rendering and picture quality, is excellent. It will never be perfect.

[baii]

Wddm 2.0 AMD driver (windows 10) is pretty broke with madvr now. Not that you can't make to work by changing queue/frame in advance/exclusive mode(completely broke on my machine), just not worth the fiddling. Avoid, avoid, avoid.

Sent from my 306SH

[huhn]

I know madshi personally involved himself in integrating madVR into Kodi, so I don't know that you are going to get anywhere. Many people won't notice the lag at all. But if it bothers you a lot, you might have to use a separate player with Kodi as an alternative set-up.

Kodi was not designed with madVR in mind. I don't know of a more sophisticated media front-end that supports madVR, so what is there is a significant accomplishment. Player performance, in terms of rendering and picture quality, is excellent. It will never be perfect.

someone else has a huge problem with that not me.
and i wouldn't be shocked if this gets fixed sooner or later. if it is a new OSD mode in madVR or by changing the presentation of the kodi OSD to another renderer instance. MPC-HC clearly shows it is possible why shouldn't it possible with kodi?

[sexus]

this is indeed a big issue , thats like saying well , if NEEDI is only half working just use another algorithm , instead of pointing out its issues and getting it fixed , this nonsensical reply can be applied to anything that needs improvement of fixing, and frankly pisses me off, this has nothing to do with being greatful or showing appreciation to madshi or others hard work , they get that from me as well , hell a couple years back warner306 might

as well couldnt be bothered with madvr for dsplayer at all since hes so asskissingly greatfull for current low quality renderers , i know these kind of folks , these are the kind of guys that are happy with whatever they get served instead of giving feedback on issues at all, if everyone would be like this progress would never be made and wed all still be on atari pcs and 56k modems....nough said

as of yet as huhn said we need a proper implementation to get this issue fixed , since again , this exists for both potplayer and kodi ...and according to huhn for mpc-hc as well, btw huhn mpc-hcs progress bar isnt rendered by madvr at all unlike the OSD elements from what ive seen, i could be wrong thou and it just isnt noticeable when frames are lagging

[Warner306]

someone else has a huge problem with that not me.
and i wouldn't be shocked if this gets fixed sooner or later. if it is a new OSD mode in madVR or by changing the presentation of the kodi OSD to another renderer instance. MPC-HC clearly shows it is possible why shouldn't it possible with kodi?

That wasn't directed at you, but the original poster.

I know the developer of DSPlayer is reluctant to create too many exceptions to the Kodi main branch to keep the two versions synchronized. He is only one man versus a team of many people that keep updating the core Kodi code. The two must remain synchronized to keep current. It would have to be a new version of madVR, I'm guessing, to improve this issue. The player is fine. I think that slight lag is the one noticeable compromise. A few crashes, but it works very well.

[Warner306]

this is indeed a big issue , thats like saying well , if NEEDI is only half working just use another algorithm , instead of pointing out its issues and getting it fixed , this nonsensical reply can be applied to anything that needs improvement of fixing, and frankly pisses me off, this has nothing to do with being greatful or showing appreciation to madshi or others hard work , they get that from me as well , hell a couple years back warner306 might

as well couldnt be bothered with madvr for dsplayer at all since hes so asskissingly greatfull for current low quality renderers , i know these kind of folks , these are the kind of guys that are happy with whatever they get served instead of giving feedback on issues at all, if everyone would be like this progress would never be made and wed all still be on atari pcs and 56k modems....nough said

as of yet as huhn said we need a proper implementation to get this issue fixed , since again , this exists for both potplayer and kodi ...and according to huhn for mpc-hc as well, btw huhn mpc-hcs progress bar isnt rendered by madvr at all unlike the OSD elements from what ive seen

Do you want to do the coding yourself? Aracnoz is doing this in his spare time. He has mentioned the headache it is to keep DSPlayer relevant.

Team Kodi abandoned DSPlayer a long time ago. Instead of campaigning here, you should be asking someone from Team Kodi to donate some time to DSPlayer. They are so focused on smaller boxes and Linux-based operating systems that you might not get a response. But it is a worthy cause. Kodi already includes so many features that promote a true home theatre experience. Why skimp on picture quality when a solution is already available to many? DSPlayer needs more assistance to be the perfect everything. Aracnoz does this as a hobby; that's why there are disclaimers on the main forum.

[sexus]

well , first off i didnt know which one to notify about this first so ive posted this issue for madvr and ill be posting and contacting kodi devs on this issue as well as post some in the dsplayer thread about this and see what aracnoz says about this

[huhn]

as of yet as huhn said we need a proper implementation to get this issue fixed , since again , this exists for both potplayer and kodi ...and according to huhn for mpc-hc as well, btw huhn mpc-hcs progress bar isnt rendered by madvr at all unlike the OSD elements from what ive seen, i could be wrong thou and it just isnt noticeable when frames are lagging


the madVR OSD is notable delayed with present 16 frames in advance and 24 GPU queue. i set it this high to make sure it's delayed. so this is something in madVR.

but the hole MPC-HC control panel isn't delayed at all it is simple not rendered on the frames from madVR. this is showing clear that none delayed control can be achieved with the current version of madVR and why should you even render control on a OSD?

i simply don't get why kodi is rendering everything on the madVR frames/OSD. even the mouse can be delayed like crazy. i guess this is simply the way the none DS player version works and the creator of the DS player version has to build everything up on this "mess". but why did they create it this way...

i just did a small test with the default kodi. it is delayed too not as hard as the ds player version by using madVR but it is delayed by at least a frame when playback starts. this is a serious issue with kodi in general.

[huhn]

I heard there might be issues with native vs loopback GPU decoding. If I have no issues with native should I keep using it? Shouldn't it be faster?

if it's faster doesn't matter in real playback cases. madVR gets the frames fast enough or not.

http://forum.doom9.org/showpost.php?p=1715427&postcount=28552

and don't forget rule 8.

[tobindac]

DXVA native doesn't work with other filter like vsfilter. you can't use force film mode or the decimation feature of madVR.

I noticed you are mixing subtitles and decoding and rendering in your logic. If you choose LAV decoding to be GPU assisted, why does it affect madvr?

I expected they all use a different GPU context.

[Warner306]

the madVR OSD is notable delayed with present 16 frames in advance and 24 GPU queue. i set it this high to make sure it's delayed. so this is something in madVR.

but the hole MPC-HC control panel isn't delayed at all it is simple not rendered on the frames from madVR. this is showing clear that none delayed control can be achieved with the current version of madVR and why should you even render control on a OSD?

i simply don't get why kodi is rendering everything on the madVR frames/OSD. even the mouse can be delayed like crazy. i guess this is simply the way the none DS player version works and the creator of the DS player version has to build everything up on this "mess". but why did they create it this way...

i just did a small test with the default kodi. it is delayed too not as hard as the ds player version by using madVR but it is delayed by at least a frame when playback starts. this is a serious issue with kodi in general.

I am not that bothered by the lag, but other options do exist. I have used MPC with Kodi and the experience is pretty seamless. Users can push for new features, but like everything in life, there is always a compromise.

If you like Kodi and DirectShow players, use either DSPlayer, MPC or any other madVR-supported player. Otherwise, use Emby, MyMovies or whatever your preference. None of them are perfect, either.

[Asmodian]

I noticed you are mixing subtitles and decoding and rendering in your logic. If you choose LAV decoding to be GPU assisted, why does it affect madvr?.

Because it changes where in memory the frames are when madVR gets them. DXVA2 Native leaves them in the GPU memory while all the other options leave them in the system memory.

I am not sure what you mean by "mixing subtitle and decoding and rendering" but DXVA Native doesn't allow vsfilter to draw subtitles on the frames because vsfilter draws subtitles on frames in system memory, it cannot draw on frames in the GPU memory. It is the same with madVR's force film (IVTC) implementation.

[sexus]

warner306 i respect your oppinion as every individual is entitled to one , but i think wheve all got it , you dont care about current issues such as the one ive pointed out , and thats good for you , now if you dont have anything to add to this issue to solve it id appreciate it if youd refrain from commenting on the forementioned issue any further, and again your so called "other options" have OSD lag as well as long as they use madvr renderer dont matter if you use them as external player plugged into kodi , its the goddamned same issue!, its the way madvr

renders the OSD elements thats the issue here , as pointed out by not just me but by huhn as well , and NO there is no compromise on issues needing to be solved , again this kind of attitude gets nobody nowhere except with current buggy implemented playback and issues that could be solved , this kind of attitude is what makes youre previous oppinions non valid to the current issue , again i have nothing against your oppinions but if you dont give a shit , then dont post , cause frankly its this kind of crap that gets me pissed off , thank you

[e-t172]

sexus: software development, like every engineering project, is all about tradeoffs and priorities. Unless of course you have an infinite budget along with infinitely talented developers, a perfect codebase completely devoid of technical debt, and a target platform with infinite performance. If you don't have those things, then you need to accept that some things are only going to get done, if at all, after other priorities have been dealt with.

Warner306 was simply playing the dissenting voice by pointing out that not all madVR users are bothered by the lag, and therefore not all users want the lag issue to be prioritized over features that are more relevant to their use case. Stating this opinion is certainly not pointless as you seem to think, to the contrary - it provides madshi with valuable feedback as to which users care about what, which can help him prioritize features and fixes.

[Warner306]

warner306 i respect your oppinion as every individual is entitled to one , but i think wheve all got it , you dont care about current issues such as the one ive pointed out , and thats good for you , now if you dont have anything to add to this issue to solve it id appreciate it if youd refrain from commenting on the forementioned issue any further, and again your so called "other options" have OSD lag as well as long as they use madvr renderer dont matter if you use them as external player plugged into kodi , its the goddamned same issue!, its the way madvr

renders the OSD elements thats the issue here , as pointed out by not just me but by huhn as well , and NO there is no compromise on issues needing to be solved , again this kind of attitude gets nobody nowhere except with current buggy implemented playback and issues that could be solved , this kind of attitude is what makes youre previous oppinions non valid to the current issue , again i have nothing against your oppinions but if you dont give a shit , then dont post , cause frankly its this kind of crap that gets me pissed off , thank you

I didn't say I didn't care; I was just pointing out that what you are asking for is a new feature, not a bug. I am fairly certain DSPlayer is designed to work this way. Otherwise, aracnoz would not have released it in that state. If you check the changelog, he is still updating it everyday with improvements.

I was mostly cautioning you to make it people's job to fix these issues. I am sure aracnoz has a full-time job and does this as a hobby. Be careful when labelling your problems as bugs, when they are a matter of personal preference.

I not passionate about this topic, so I won't get in your way of discussing this further...

[XMonarchY]

Why does Exclusive Fullscreen usage results in dropped frames in almost every film I watch? If I turn it off, then there is no frame drop...

[huhn]

no clue you leaving out a ton of information people can only guess.

[Olivier C.]

Hi,

I am developing a .NET application and i am trying to detect a new display.

When a new display is detected, i have to :
- detect all supported display modes
- populate madVR's settings, especially the line parameter "display modes"
- enable options "switch ..." and "restore..." in order to use the auto-switch feature

Question :
How can i use the madVR's API in .NET (C#) in order to ask to set these options ?

I have found a bad alternative (registry edit), but i prefer a cleaner and safer way to achieve this.

Thanks a lot in advance
Olivier

[madshi]

Here are the two logs - the first one created with v0.87.14 and the other one with v0.87.17
Ok, I see the issue. Both the old and new madVR builds fail to detect the refresh rate properly. Even v0.87.14 complains in the log that there's something fishy going on. However, v0.87.14 at least uses a "best guess" to do the rendering while v0.87.17 doesn't make use of the best guess. I *think* that I fixed this problem in this test build, but I'm not 100% sure because I can't reproduce the problem on my PC:

http://madshi.net/madVR8717b.zip

Does it work for you?

Unfortunately I found an issue that seems odd to me:

Issue: Some MKV videos have very dark pictures (wrong, missing colors) when I play them in Kodi using MadVR in either full screen exclusive or full screen overlay. As soon as I press CTRL+Enter to get out of Full screen mode, the picture corrects itself. but if I press CTRL+Enter again, the same problem comes back.

I tested DSPlayer with EVR, MPC-HC with same MadVR / LavFilters and MediaBrowser Theater with LAV & MadVR but I couldn't repro this bug in any of the other environments. So I assume it's something to do with DSPlayer and MadVR integration.
I managed to pin point the issue. I can only see this problem in the following scenario:
1. Only when using DSPlayer (Kodi) with MadVR as a renderer in either Full screen exclusive or Full screen overlay
2. Have luma doubling set (# of neurons doesn't seem to matter)
3. Have the downscaling done by any of the custom pixel shader options (I use Catmull-Rom)

As soon as I change the downscaling to DXVA2 the problem goes away. It repros with version 0.87.17.

Question - do you know what could be causing this behavior only on DSPlayer? Do I lose quality by using DXVA2 instead of Catmull-Rom (looks like DXVA2 equals Bilinear for my NVIDIA card)
Strange. Can you still reproduce the problem with the new test build (see above)? Have you tried resetting the kodi specific madVR settings and to reset madVR to default settings? Which settings do you need to change to make the problem appear then? A debug log might eventually help, although I'm not sure...

I'd like to report a bug.

I'm not sure which program is at fault here but if you are playing a video with mpc-hc and then close it when you have amd gaming evolved (raptr) running then mpc-hc will get stuck running in the background and it will eat up one of your cpu cores until you close it via task manager.
Does it still occur with v0.87.17? If so, I'm not sure what it could be. You could try pressing Ctrl+Alt+Shift+Break. Then wait a few seconds. Do you then get a freeze report text file on your desktop? If so, post it somewhere for me to look it (not here in the forum) and post a link here.

madshi got a big request for madvr , its been bugging me ever since the support for OSD elements by madvr and am suprised this issue hasnt been tackled earlier , so ive noticed madvr has always had issues with rendering OSD elements smoothly aka theres a noticeable delay between OSD elements this has been both noticed on potplayer and kodis latest dsplayer madvr build this can be easily reproduced by simply up and downing the volume in potplayer or kodi or backing into the kodi home navigation screen while a video is being rendered , mind you ive tried already by putting prerendered frames to 1 and deselecting wait for renderqueue is full for window and exclusive as well did i try it with cranked down upscaling ,yet as said all to no avail , id appreciate a fix alot
You should be able to reduce the lag a bit by reducing the size of the GPU queue, the number of "presented frames in advance", and/or maybe even by disabling the "present several frames in advance" options completely for both windowed and exclusive mode. In the end this is a lag that is "by design", though. Changing the options mentioned above should reduce the lag, but it will not totally get rid of it. The reason for that is that madVR tries to render many frames in advance, to make sure playback stays smooth even if the PC gets super busy for short periods of time during playback.

Maybe I can find a way to reduce the lag in a future madVR build, but for now I have different priorities.

Will Nvidia's new drivers speed up NNEDI3 in any way? New drivers support OpenCL 1.2 apparently. I see a lot of "#if __OPENCL_VERSION__ >= 110" in the code ;).
I doubt it, sadly.

FWIW, I don't understand why Kodi doesn't just render the video with madVR and any elements needed during playback, but keeps the main GUI on a separate 3D renderer, with no delay at all. But oh well, their design choice I guess.
The main reason for that is that doing what you suggest wouldn't work in fullscreen exclusive mode. The only way to render a GUI in FSE mode is to go through the madVR OSD interfaces - and that is what kodi does. And I'm glad that it does, otherwise we would have no kodi GUI in FSE mode.

Yeah, kodi could probably offer two different ways to render the GUI, one for FSE mode and one for windowed mode. But to be honest, I'm not sure if it's worth the effort. Yes, currently there's a lag, and it's not nice, but as long as you don't use the mouse, I don't consider it a show stopper (even if it might be somewhat annoying).

I use my HTPC with a 4K TV but If I set MadVR to "enable automatic fullscreen exclusive mode" I see black screen immediatly after the muose arrow vanish.

If I disable it I've no problem.

I've an AMD card and I exit at 2160@23p RGB 4:4:4

If I set my HTPC to output in 1920x1080 I can use "enable automatic fullscreen exclusive mode" without any problem.
Is the display even able to sync to the display mode? Do you know that? My projector complains if it can't sync. Does yours? So this could be one potential cause. Another thing is, maybe madVR just fails rendering for some reason. I might be able to see something if you upload a (zipped) debug log somewhere (not here) and link to it here.

I noticed something odd. When I drag a window with a playing (or paused) video from one monitor to a second monitor, madVR crashes. Sometimes I get the following error:

madVR Error:
- creating Direct3D device failed (80070005)

When I select "enable windowed overlay" in general settings, I can move the window with the playing video from one monitor to the other with no crash.

I am using MPC-HC with Windows 7 64 bit and a GTX 670 video card.
Does that still occur with v0.87.17? When that crash occurs, do you get a crash report box? Have you sent a crash report to me yet?

Wddm 2.0 AMD driver (windows 10) is pretty broke with madvr now. Not that you can't make to work by changing queue/frame in advance/exclusive mode(completely broke on my machine), just not worth the fiddling. Avoid, avoid, avoid.
Ouch.

Why does Exclusive Fullscreen usage results in dropped frames in almost every film I watch? If I turn it off, then there is no frame drop...
As huhn already said, you've not provided enough information for us to help. Try making a photo of the debug OSD (Ctrl+J) with your digicam, when those frame drops occur, and upload it somewhere (not here) for us to look at.

I am developing a .NET application and i am trying to detect a new display.

When a new display is detected, i have to :
- detect all supported display modes
- populate madVR's settings, especially the line parameter "display modes"
- enable options "switch ..." and "restore..." in order to use the auto-switch feature

Question :
How can i use the madVR's API in .NET (C#) in order to ask to set these options ?

I have found a bad alternative (registry edit), but i prefer a cleaner and safer way to achieve this.
One potential problem with this approach is that if you modify the settings from within your media player, you might overwrite settings the user has already painfully setup himself. Anyway, that's your decision, of course.

I've zero experience with C#, so I don't know if I can answer your questions. Basically you should create a madVR instance, just like a media player would, then use QueryInterface to turn the madVR instance into an IMadVRSettings interface, then you can use the methods of that interface. Hope that helps?

[noee]

Pretty straight forward, you can just use COM-interop to instantiate a madVR and call the madVR apis. You can just use the WinAPI to get DisplayModes, etc:

"declare" madVR ...
using System;
using System.Runtime.InteropServices;
using System.Text;
using DirectShowLib;

namespace Filters.madVR
{
#region madVR Calibration
// public static class Calibration {
// // Calibration apis
// [DllImport("madhcnet32.dll", EntryPoint = "madVR_ConnectDialog")]
// public static extern bool ConnectDialog(bool lan, IntPtr hwnd);
//
// [DllImport("madhcnet32.dll", EntryPoint = "madVR_BlindConnect")]
// public static extern bool BlindConnect(bool lan, long timeout);
//
// [DllImport("madhcnet32.dll", EntryPoint = "madVR_ConnectToIp")]
// public static extern bool ConnectToIp([MarshalAs(UnmanagedType.LPWStr)] string ip, long timeout);
//
// [DllImport("madhcnet32.dll", EntryPoint = "madVR_Disable3dlut")]
// public static extern bool Disable3dlut();
//
// [DllImport("madhcnet32.dll", EntryPoint = "madVR_SetDeviceGammaRamp")]
// public static extern bool SetDeviceGammaRamp(IntPtr ramp);
//
// [DllImport("madhcnet32.dll", EntryPoint = "madVR_SetOsdText")]
// public static extern bool SetOsdText([MarshalAs(UnmanagedType.LPWStr)] string msg);
//
// [DllImport("madhcnet32.dll", EntryPoint = "madVR_Disconnect")]
// public static extern bool Disconnect();
//
// [DllImport("madhcnet32.dll", EntryPoint = "madVR_ShowProgressBar")]
// public static extern bool ShowProgressBar(int numMeasurements);
//
// [DllImport("madhcnet32.dll", EntryPoint = "madVR_ShowRGB")]
// public static extern bool ShowRGB(double R, double G, double B);
//
// [DllImport("madhcnet32.dll", EntryPoint = "madVR_SetBackground")]
// public static extern bool SetBackground(int areapct, int backgroundcolor);
//
// }
#endregion

#region madVR COM Interfaces

// madVR Filter COM InterOp)
[ComImport, Guid("E1A8B82A-32CE-4B0D-BE0D-AA68C772E423")]
public class madVideoRenderer {}

[ComImport, System.Security.SuppressUnmanagedCodeSecurity,
InterfaceType(ComInterfaceType.InterfaceIsIUnknown),
Guid("88A69329-3CD3-47D6-ADEF-89FA23AFC7F3")]
interface IMadVRExclusiveModeControl
{
void DisableExclusiveMode(bool disable);
}

[ComImport, System.Security.SuppressUnmanagedCodeSecurity,
InterfaceType(ComInterfaceType.InterfaceIsIUnknown),
Guid("ABA34FDA-DD22-4E00-9AB4-4ABF927D0B0C")]
interface IMadVRTextOsd
{
void DisplayMessage([MarshalAs(UnmanagedType.LPWStr)] string msg, long duration);
void ClearMessage();
}

[ComImport, System.Security.SuppressUnmanagedCodeSecurity,
InterfaceType(ComInterfaceType.InterfaceIsIUnknown),
Guid("D2D3A520-7CFA-46EB-BA3B-6194A028781C")]
interface IMadVRSeekBarControl
{
void DisableSeekBar(bool disable);
}
.
.
.
.
.
[ComImport, System.Security.SuppressUnmanagedCodeSecurity,
InterfaceType(ComInterfaceType.InterfaceIsIUnknown),
Guid("6F8A566C-4E19-439E-8F07-20E46ED06DEE")]
interface IMadVRSettings
{
//Revision
[PreserveSig]
bool SettingsGetRevision(
[Out] out Int64 revision
);

//Settings file
[PreserveSig]
bool SettingsExport(
[Out, MarshalAs(UnmanagedType.AsAny)] out string buf,
[Out] out int size
);
[PreserveSig]
bool SettingsImport(
[In, MarshalAs(UnmanagedType.AsAny)] string buf,
[In] int size
);

//Sets
[PreserveSig]
bool SettingsSetString(
[MarshalAs(UnmanagedType.LPWStr)] string path,
[MarshalAs(UnmanagedType.LPWStr)] string value
);
[PreserveSig]
bool SettingsSetInteger(
[MarshalAs(UnmanagedType.LPWStr)] string path,
int value
);
[PreserveSig]
bool SettingsSetBoolean(
[MarshalAs(UnmanagedType.LPWStr)] string path,
[MarshalAs(UnmanagedType.Bool)] bool value
);

//Gets
[PreserveSig]
bool SettingsGetString(
[In, MarshalAs(UnmanagedType.LPWStr)] string path,
[Out, MarshalAs(UnmanagedType.LPWStr)] StringBuilder value,
[Out] out int size
);
[PreserveSig]
bool SettingsGetInteger(
[In, MarshalAs(UnmanagedType.LPWStr)] string path,
[Out] out int value
);
[PreserveSig]
bool SettingsGetBoolean(
[In, MarshalAs(UnmanagedType.LPWStr)] string path,
[Out] out bool value
);
[PreserveSig]
bool SettingsGetBinary(
[In, MarshalAs(UnmanagedType.LPWStr)] string path,
IntPtr ptr,
[Out] out int size
);

}

[ComImport, System.Security.SuppressUnmanagedCodeSecurity,
InterfaceType(ComInterfaceType.InterfaceIsIUnknown),
Guid("8FAB7F31-06EF-444C-A798-10314E185532")]
interface IMadVRInfo
{
//Gets
[PreserveSig]
void GetBool(
IntPtr fld,
[Out, MarshalAs(UnmanagedType.Bool)] out bool value
);
[PreserveSig]
void GetInt(
IntPtr fld,
[Out] out int value
);
// TODO: SIZE Struct
[PreserveSig]
void GetSize(
IntPtr fld,
IntPtr size
);
// TODO: RECT Struct
[PreserveSig]
void GetRect(
IntPtr fld,
IntPtr rect
);
[PreserveSig]
void GetULong(
IntPtr fld,
[Out] out ulong value
);
[PreserveSig]
void GetDouble(
IntPtr fld,
[Out] out double value
);
[PreserveSig]
void GetString(
IntPtr fld,
IntPtr str,
[Out] out int size
);

[PreserveSig]
void GetBinary(
IntPtr fld,
IntPtr str,
[Out] out int size
);

}


declare Win32API stuff...

using System;
using System.Runtime.InteropServices;

namespace Noee.Win32Utils
{
public static class NativeMethods
{
#region Structs
[StructLayout(LayoutKind.Sequential, CharSet = CharSet.Unicode)]
public struct DEVMODE
{
// You can define the following constant
// but OUTSIDE the structure because you know
// that size and layout of the structure is very important
// CCHDEVICENAME = 32 = 0x50
[MarshalAs(UnmanagedType.ByValTStr, SizeConst = 32)]
public string dmDeviceName;
// In addition you can define the last character array
// as following:
//[MarshalAs(UnmanagedType.ByValArray, SizeConst = 32)]
//public Char[] dmDeviceName;

// After the 32-bytes array
[MarshalAs(UnmanagedType.U2)]
public UInt16 dmSpecVersion;

[MarshalAs(UnmanagedType.U2)]
public UInt16 dmDriverVersion;

[MarshalAs(UnmanagedType.U2)]
public UInt16 dmSize;

[MarshalAs(UnmanagedType.U2)]
public UInt16 dmDriverExtra;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmFields;

public POINTL dmPosition;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmDisplayOrientation;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmDisplayFixedOutput;

[MarshalAs(UnmanagedType.I2)]
public Int16 dmColor;

[MarshalAs(UnmanagedType.I2)]
public Int16 dmDuplex;

[MarshalAs(UnmanagedType.I2)]
public Int16 dmYResolution;

[MarshalAs(UnmanagedType.I2)]
public Int16 dmTTOption;

[MarshalAs(UnmanagedType.I2)]
public Int16 dmCollate;

// CCHDEVICENAME = 32 = 0x50
[MarshalAs(UnmanagedType.ByValTStr, SizeConst = 32)]
public string dmFormName;
// Also can be defined as
//[MarshalAs(UnmanagedType.ByValArray,
// SizeConst = 32, ArraySubType = UnmanagedType.U1)]
//public Byte[] dmFormName;

[MarshalAs(UnmanagedType.U2)]
public UInt16 dmLogPixels;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmBitsPerPel;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmPelsWidth;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmPelsHeight;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmDisplayFlags;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmDisplayFrequency;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmICMMethod;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmICMIntent;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmMediaType;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmDitherType;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmReserved1;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmReserved2;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmPanningWidth;

[MarshalAs(UnmanagedType.U4)]
public UInt32 dmPanningHeight;

/// <summary>
/// Initializes the structure variables.
/// </summary>
public void Initialize()
{
this.dmDeviceName = new string(new char[32]);
this.dmFormName = new string(new char[32]);
this.dmSize = (ushort)Marshal.SizeOf(this);
}
}

// 8-bytes structure
[StructLayout(LayoutKind.Sequential)]
public struct POINTL
{
public Int32 x;
public Int32 y;
}

[StructLayout(LayoutKind.Sequential, CharSet = CharSet.Unicode)]
public struct DISPLAY_DEVICE
{
[MarshalAs(UnmanagedType.U4)]
public UInt32 cb;
[MarshalAs(UnmanagedType.ByValTStr, SizeConst = 32)]
public string dmDeviceName;
[MarshalAs(UnmanagedType.ByValTStr, SizeConst = 128)]
public string dmDeviceString;
[MarshalAs(UnmanagedType.U4)]
public UInt32 StateFlags;
[MarshalAs(UnmanagedType.ByValTStr, SizeConst = 128)]
public string dmDeviceID;
[MarshalAs(UnmanagedType.ByValTStr, SizeConst = 128)]
public string dmDeviceKey;

/// <summary>
/// Initializes the structure variables.
/// </summary>
public void Initialize()
{
this.dmDeviceName = new string(new char[32]);
this.dmDeviceString = new string(new char[128]);
this.dmDeviceID = new string(new char[128]);
this.dmDeviceKey = new string(new char[128]);
this.cb = (UInt32)Marshal.SizeOf(this);
}


}
#endregion

#region Windows calls
[DllImport("oleaut32.dll", CharSet = CharSet.Unicode, ExactSpelling = true)]
public static extern int OleCreatePropertyFrame(
IntPtr hwndOwner,
int x,
int y,
[MarshalAs(UnmanagedType.LPWStr)] string lpszCaption,
int cObjects,
[MarshalAs(UnmanagedType.Interface, ArraySubType = UnmanagedType.IUnknown)]
ref object ppUnk,
int cPages,
IntPtr lpPageClsID,
int lcid,
int dwReserved,
IntPtr lpvReserved);

[DllImport("User32.dll", SetLastError = true, BestFitMapping = false, ThrowOnUnmappableChar = true)]
[return: MarshalAs(UnmanagedType.Bool)]
public static extern Boolean EnumDisplaySettingsW(
[param: MarshalAs(UnmanagedType.LPTStr)]
String lpszDeviceName, // display device
[param: MarshalAs(UnmanagedType.U4)]
Int32 iModeNum, // graphics mode
[In, Out]
ref DEVMODE lpDevMode // graphics mode settings
);

[DllImport("User32.dll", SetLastError = true, BestFitMapping = false, ThrowOnUnmappableChar = true)]
[return: MarshalAs(UnmanagedType.Bool)]
public static extern Boolean EnumDisplayDevicesW(
[param: MarshalAs(UnmanagedType.LPTStr)]
String lpszDeviceName, // display device
[param: MarshalAs(UnmanagedType.U4)]
Int32 iDevNum,
[In, Out]
ref DISPLAY_DEVICE lpDD, // device settings
[param: MarshalAs(UnmanagedType.U4)]
Int32 dwFlags
);

[DllImport("User32.dll", SetLastError = true, BestFitMapping = false, ThrowOnUnmappableChar = true)]
public static extern int ChangeDisplaySettingsExW(
[param: MarshalAs(UnmanagedType.LPTStr)]
String lpszDeviceName, // display device
[In, Out]
ref DEVMODE lpDevMode,
IntPtr hWnd,
[param: MarshalAs(UnmanagedType.U4)]
int dwflags,
[param: MarshalAs(UnmanagedType.U4)]
int lParam
);

// [System.Runtime.InteropServices.DllImport("kernel32.dll")]
// private static extern int GetSystemDefaultLCID();

[DllImportAttribute("powrprof.dll", EntryPoint = "PowerSetActiveScheme")]
public static extern uint PowerSetActiveScheme(IntPtr UserPowerKey, ref Guid ActivePolicyGuid);

[DllImportAttribute("powrprof.dll", EntryPoint = "PowerGetActiveScheme")]
public static extern uint PowerGetActiveScheme(IntPtr UserPowerKey, out IntPtr ActivePolicyGuid);

[DllImportAttribute("powrprof.dll", EntryPoint = "PowerReadFriendlyName")]
public static extern uint PowerReadFriendlyName(IntPtr RootPowerKey, ref Guid SchemeGuid, IntPtr SubGroupOfPowerSettingsGuid, IntPtr PowerSettingGuid, IntPtr Buffer, ref uint BufferSize);
#endregion

#region MyUtils
[DllImport("utils.dll")]
public static extern int SetDefaultAudioPlaybackDevice([MarshalAs(UnmanagedType.LPWStr)] string Parameter);
#endregion

}

}


Then, you can instantiante and use....
.
.
.
/// LAV Video Instance
/// </summary>
public static LAVVideoSettings LAVVideo {
get
{
if (m_cLAVVideoSettings == null)
m_cLAVVideoSettings = new LAVVideoSettings();
return m_cLAVVideoSettings;
}
}
/// <summary>
/// The madVR instance
/// </summary>
public static madVRSettings madVR {
get
{
if (mVR == null)
mVR = new madVRSettings();
return mVR;
}
}
.
.
.
.

[Tom Keller]

Does it work for you?
It does indeed :) :

http://abload.de/img/madvr9mue3.jpg

Thanks a lot!

[madshi]

The OSD is only one part, though. Does it also really play fluidly again now?

[Tom Keller]

Yes. I did use some videos with long camera pans for testing. With your test build they show no signs of stuttering and play as smooth as they did with v0.87.14 in the past.

[madshi]

Great - good to hear!

[sexus]

You should be able to reduce the lag a bit by reducing the size of the GPU queue, the number of "presented frames in advance", and/or maybe even by disabling the "present several frames in advance" options completely for both windowed and exclusive mode. In the end this is a lag that is "by design", though. Changing the options mentioned above should reduce the lag, but it will not totally get rid of it. The reason for that is that madVR tries to render many frames in advance, to make sure playback stays smooth even if the PC gets super busy for short periods of time during playback.

Maybe I can find a way to reduce the lag in a future madVR build, but for now I have different priorities.




disabling the "present several frames in advance" options completely for both windowed and exclusive mode has never been enabled from the getgo on my setup so thats a none issue, about the gpu queue / cpu queue and the number of presented frames in advance , what settings do you suggest? ive currenlty got gpu queue at 24 and cpu queue at 48 and prerendered frames at 8 and yes id really appreciate if you could fix this issue with a future madvr build indeed , i understand youve got more pressing issues at hand as of currently , ill gladly wait for this annoying issue to be fixed madshi thanks for your time as always

[ryrynz]

what settings do you suggest? ive currenlty got gpu queue at 24 and cpu queue at 48 and prerendered frames at 8 and yes id really appreciate if you could fix this issue with a future madvr build indeed , i understand youve got more pressing issues at hand as of currently , ill gladly wait for this annoying issue to be fixed madshi thanks for your time as always

Just drop them by half for starters and keep lowering them to see what effect they have on the lag if any, personally I think your queues are higher than they likely need to be.

[dbezerra]

From Madshi:

"Strange. Can you still reproduce the problem with the new test build (see above)? Have you tried resetting the kodi specific madVR settings and to reset madVR to default settings? Which settings do you need to change to make the problem appear then? A debug log might eventually help, although I'm not sure..."


After I reported this issue here I found out that the problem only happens when:
a. NNEDI is enabled (even if no downscale is necessary)
b. I have both Kodi and MadVR set to adjust the framerate.

By disabling the adjust framerate on Kodi this problem went away, although I also lost the ability to set the proper frame rate when not using DSPlayer/MadVR. (streaming for example).

I will check your beta build just in case. But now this is a low priority issue for me.

[madshi]

disabling the "present several frames in advance" options completely for both windowed and exclusive mode has never been enabled from the getgo on my setup so thats a none issue, about the gpu queue / cpu queue and the number of presented frames in advance , what settings do you suggest? ive currenlty got gpu queue at 24 and cpu queue at 48 and prerendered frames at 8 and yes id really appreciate if you could fix this issue with a future madvr build indeed , i understand youve got more pressing issues at hand as of currently , ill gladly wait for this annoying issue to be fixed madshi thanks for your time as always
The default is a GPU queue size of 8, IIRC. I'm not sure if that's what I'd recommend. Depends on whether you're using smooth motion FRC, which can always use a few frames more. In any case, as ryrynz suggested, just play with the queue sizes and with the "number of backbuffers". E.g. just try a GPU queue size of 4, and also set the "number of backbuffers" to 3. Does that reduce the lag? If so, check if you still get perfectly smooth playback. If not, increase the queue sizes a bit again. But first check if reducing these options has any effect on the lag at all.

BTW, I think the CPU queue size should have no effect on the issue, so you can keep that high.

After I reported this issue here I found out that the problem only happens when:
a. NNEDI is enabled (even if no downscale is necessary)
b. I have both Kodi and MadVR set to adjust the framerate.

By disabling the adjust framerate on Kodi this problem went away, although I also lost the ability to set the proper frame rate when not using DSPlayer/MadVR. (streaming for example).

I will check your beta build just in case. But now this is a low priority issue for me.
Interesting. I'm not sure what's happening there, to be honest. But it's not good to have 2 different pieces of software fight over which frame rate and display mode should be used. If you use DSPlayer sometimes with and sometimes without madVR then maybe you could ask for a setting in kodi which enables framerate adjustments only when *not* using madVR?

[sexus]

oh hey , sup ryrynz , anyhow ok madshi so ive set gpu queue to 4 even thou huhn recommended against it due to it most likely breaking deinterlacing and introducing dropped frames as well did i reduce prebuffered frames to 3, ok , the OSD lag is noticeably less since aka you now have effectively 4 frames of OSD lag instead of 24 according to huhn , not sure what setting prebuffered frames would do for the OSD lag thou but i set them to 3 as recommended by you anyhow , ill now check for dropped frames , not sure how to check for broken deinterlacing thou ?

[madshi]

As long as you don't get any dropped frames, you'll probably be fine. The good news is that reducing the lag through the settings works for you. Now you just need to find the lowest queue size that plays smoothly for you.

[sexus]

well anything above 4 gpu queue size is noticable in OSD lag so , not sure what exactly to tweak then , say if i now have broken deinterlacing , which i still have no idea how to even notice or have dropped frames , and raising the queue further would fix this logically , but would introduce OSD lag that is noticeable aka anything above 4 frames of lag is noticeable on the OSD, ive tried,the best way to test this is as said, is to quickly go up and down with the volume and the OSD lag becomes very apparent, you can literally see the delay, you must already know this since youve tried

[ryrynz]

oh hey , sup ryrynz , anyhow ok madshi so ive set gpu queue to 4 even thou huhn recommended against it due to it most likely breaking deinterlacing, ill now check for dropped frames , not sure how to check for broken deinterlacing thou ?

Hey there, been awhile.. Anyway, deinterlacing isn't something that's used by many people these days as most media is generally progressive content so unless your watching TV broadcasting or DVDs it's unlikely you
will need deinterlacing.

I've only ever used it a few times since madVR came out, so if you're in the same boat it isn't a deal breaker setting the GPU queue to 4.

[huhn]

oh hey , sup ryrynz , anyhow ok madshi so ive set gpu queue to 4 even thou huhn recommended against it due to it most likely breaking deinterlacing and introducing dropped frames as well did i reduce prebuffered frames to 3, ok , the OSD lag is noticeably less since aka you now have effectively 4 frames of OSD lag instead of 24 according to huhn , not sure what setting prebuffered frames would do for the OSD lag thou but i set them to 3 as recommended by you anyhow , ill now check for dropped frames , not sure how to check for broken deinterlacing thou ?

just play a deinterlaced file at best a broadcast file or interlaced BD.

in my experience a GPU queue of 4 can result in issue with IVTC/decimation or deinterlacing. usually 8 doesn't create any problems at least in my experience.

you can use alt + control +shift + d to force deinterlacing on any file the GPU deinterlace has a very easy job with deinteralcing progressive frames but still better than no tests.
here an example where it doesn't work with a 4 GPU queue and 2 present 2 frames in advanced:
http://abload.de/img/worksbiuph.png

at 5 GPU queue it was stable. i added SM always on top and it was still stable.

i'm not sure about present or buffered frames in advance. but the GPU queue has clearly an effect on KODI and the OSD from madVR in other players like mpc-hc. 8 or 24 GPU queue is day night. but i don't see a reason where you clearly need more than 8 GPU queue but of course not all system are the same.

[ryrynz]

but i don't see a reason where you clearly need more than 8 GPU queue but of course not all system are the same.

Personally I think many people don't have good reason to go beyond the defaults, if there was the defaults would be higher.

[sexus]

i see thanks for the elaborate explanation and showing the deinterlacing issue, it appears that i have just about 0 high demanding deinterlaced files , maybe someone could upload one for me to test this, so we can further narrow down what gpu queue and backbuffers amount is needed to get 0 delayed and 0 dropped frames

btw my flush settings are set for both window and exclusive to

flush
flush and wait(sleep)
dont flush
flush

i recon these are correct

[ryrynz]

btw my flush settings are set for both window and exclusive to

flush
flush and wait(sleep)
dont flush
flush

i recon these are correct

Default is 'don't flush' after D3D presentation.

[sexus]

ok ive just corrected that, btw according to huhn we have to set the gpu queue to atleast 5 in order to have deinterlacing not break on us, he just set his queue to 8 for safe measure, i guess having deinterlacing properly working is of importance as well once we start using madvr for live tv streaming and general streaming such as youtube videos on kodi, can anyone confirm huhns setting of 5 as a minimum requirement to have deinterlacing running stable?

[huhn]

i see thanks for the elaborate explanation and showing the deinterlacing issue, it appears that i have just about 0 high demanding deinterlaced files , maybe someone could upload one for me to test this, so we can further narrow down what gpu queue and backbuffers amount is needed to get 0 delayed and 0 dropped frames

btw my flush settings are set for both window and exclusive to

flush
flush and wait(sleep)
dont flush
flush

i recon these are correct

i found this by googling interlaced samples https://dl.dropboxusercontent.com/u/25077543/DIGIANIMALHD.ts

and i found a new bug in madVR time to investigate i guess.

[sexus]

a bug related to deinterlacing , dropped frames , delayed frames or OSD lag?


btw that sample screenshot for failed deinterlacing isnt showing me anything noticeable , if you could point out the issue id appreciate it , since i seem to be a noob when it comes to deinterlacing

update: sweet mother of god , ive heard about a few frames dropped every so often in a movie but this is madness im getting a ton of dropped frames and delayed frames on that sample youve linked me to, plus a weird small thin black bar at the top left of the screen thats blinking, is this supposed to be like this? , its like about 1100 dropped frames and 250 delayed and counting depending on how long i let it loop , wtf! mind you dont use reclock , so all videos use the refresh of my monitors refresh which is at 60 hz

btw changing gpu queue size doesnt affect it either , ive also noticed gpu queue size doesnt affect delayed nor dropped frames whatsoever in a sample movie , alien anthology blueray that ive played , the movie alien, i always have a one or two dropped frames every 3-5 minutes at random, ive only got 1-2 delayed frame thou and 11 dropped frames , by letting it run for 30 minutes, on gpu queue 4, i recon this is good?

it seems to be the blueray the way its been encoded itself , maybe someone could upload an acurate delay and loss test

and

heres a screenshot


and about deinterlacing issues ,oldpoem over at kodi suggested to use lavfilters for deinterlacing instead of madvrs, is this a good idea?

[huhn]

a bug related to deinterlacing , dropped frames , delayed frames or OSD lag?
has something to do with a dual monitor setup where one monitor is 120 hz and the other is 60 hz.

btw that sample screenshot for failed deinterlacing isnt showing me anything noticeable , if you could point out the issue id appreciate it , since i seem to be a noob when it comes to deinterlacing

update: sweet mother of god , ive heard about a few frames dropped every so often in a movie but this is madness im getting a ton of dropped frames and delayed frames on that sample youve linked me to, plus a weird small thin black bar at the top left of the screen thats blinking, is this supposed to be like this? , its like about 1100 dropped frames and 250 delayed and counting depending on how long i let it loop , wtf! mind you dont use reclock , so all videos use the refresh of my monitors refresh which is at 60 hz

btw changing gpu queue size doesnt affect it either , ive also noticed gpu queue size doesnt affect delayed nor dropped frames whatsoever in a sample movie , alien anthology blueray that ive played , the movie alien, i always have a one or two dropped frames every 3-5 minutes at random, ive only got 1-2 delayed frame thou and 11 dropped frames , by letting it run for 30 minutes, on gpu queue 4, i recon this is good?

it seems to be the blueray the way its been encoded itself , maybe someone could upload an acurate delay and loss test
maybe your madVR settings are way to high to render 50 FPS?
the black bar in the top left is part of the video.
heres a screenshot
think about adding a link to the screen attachment may take some time.


and about deinterlacing issues ,oldpoem over at kodi suggested to use lavfilters for deinterlacing instead of madvrs, is this a good idea?
you can try it but in general this is not that good.

[madshi]

madVR v0.87.18 released

http://madshi.net/madVR.zip

* fixed: #181: "deint" tag sometimes didn't work when using profiles
* fixed: #277: DVD Navigator: external subtitles displayed incorrectly
* fixed: target rects with odd widths/heights resulted in a black border
* fixed: SettingsSetBoolean("DebugOSD") didn't take effect immediately
* DXVA scaling is done in RGB again for Intel and NVidia
* DXVA performance improvement under certain conditions
* small performance improvement when movie aspect ratio doesn't match display
* 3dlut split screen mode can now supports mouse drag & drop
* 3dlut can now be temporarily disable (and reenabled) via keyboard shortcut
* some reliability improvements to refresh rate calculation algorithm
Wasn't really planning to do this release, but a couple of bug fixes were necessary/important, and one of the calibration software companies wanted extended 3dlut split screen functionality. So here you go. No further release planned in the next 2-3 weeks.

[sexus]

has something to do with a dual monitor setup where one monitor is 120 hz and the other is 60 hz.

i see


maybe your madVR settings are way to high to render 50 FPS?
the black bar in the top left is part of the video.

so it would seem , perhaps its me using NEEDI64 for chroma upscaling + image doubling set to 16 for all..hmmm..anyone can confirm this?

think about adding a link to the screen attachment may take some time.

a what? you just have to click the attachment itll autolink, its currently showing as pending approval

heres the link

http://forum.doom9.org/attachment.php?attachmentid=14711&d=1428496439


you can try it but in general this is not that good.

whys that so?

[Digiface]

Just installed MadVR v0.87.18. Noticed that MPC-HC (1.7.8) freezes when seeking. This only seems to happen with files that uses Haali.

Edit. Installed earlier version of MadVR and now everything's fine again, so there's certainly something fishy with this new version?