madVR - high quality video renderer (GPU assisted)

[madshi]

Please have a look. This is file http://dl.dropbox.com/u/43376972/Cadence%2032322.mkv (DVD content) where Madvr detects (film mode) very exotic cadence (3:2:3:2:2).
Is the source content actually 25p transformed to 59,94i?
Shoud it be deinterlaced and decimated to 25p?
Another sample of 3:2:3:2:2 cadence in NTSC dvd (in mkv). If madvr in film mode we see combing and everything is good in video mode.

http://dl.dropbox.com/u/43376972/Cadence%2032322%20num2.mkv
Both samples are field blended, so IVTC won't work.

Update: A new sample for you:
http://www.mediafire.com/?cicxx9u4crfkb5o

Cadence is detected as 3:2, but there's a lot of flickering/aliasing and combing. Not sure if this is something going wrong with the deinterlacing or if it's a really bad source. (it's a really bad film, so I wouldn't be surprised!)
The source itself seems to contain all the problems.

here a 3 examples of my NTSC DVD Library where i am not able to IVTC with madVR or DScaler mod.

https://rapidshare.com/files/576191865/Examples.zip
The first and 3rd samples are field blended. IVTC doesn't work for such DVDs. The 2nd sample is detected by madVR as 4:2:2:2 and IVTCing appears to work just fine here. The 4:2:2:2 cadence seems to be correct, too, based on a quick check on some consecutive fields.

Another deinterlacing sample. This is a weird one, don't know if it will be useful or not. The film itself was perfect, staying in 2:2, but when the credits came up, it went to unknown cadence, and then switched to 4:2:2. (which is wrong)

http://www.mediafire.com/?l85rwl1ss2cghv5
The end credits in this sample are natively interlaced, sadly, so IVTC won't work for them.

[kalston]

Month old I know, but he may had been on to something.

http://support.microsoft.com/kb/2653312

It's a hotfix for distorted audio file sounds after you convert the sample rate in Windows 7 or Windows Server 2008 R2

Oooh nice, I had been looking at various threads mentioning this problem and thought no hotfix had been released yet. Glad to see Microsoft has looked into it. The distortion going from 44,1khz to 48khz (or the other way around) was just plain horrible and easy to notice.

@ 6233638
Thanks a lot for your post, I'll be sure to adjust my settings according to your recommendations once I get back home.

[e-t172]

The Windows Visa/7 audio engine does not introduce distortion. In fact it should be bit-exact if the input sample rate matches the configured sample rate.

Month old I know, but he may had been on to something.

I was wrong when I said that: the audio engine dithers the output. This is inaudible but obviously not bit perfect.

[Trib]

I recommend that you open these in separate browser tabs (middle click, or ctrl+click) and swap between them to compare. (ctrl+tab to go forwards, ctrl+shift+tab to go back) They're posted in order of sharpness.


Sick post! Thanks a lot! =)

[dansrfe]

I was wrong when I said that: the audio engine dithers the output. This is inaudible but obviously not bit perfect.

What does Windows 7 dither the output to? 48khz?

[nevcairiel]

What does Windows 7 dither the output to? 48khz?

Sample rate is not dithered.
The mixer does all its processing in float, so when it needs to convert that float back to 16 or 24-bit int, it uses dithering. Rather have dithering then no dithering, of course, but its not untouched because its internally converted to float once.

[kerman]

Is it advisable enable dithering adding on filters? I have a plasma which already produces dithering by its own; CoreAVC adds dithering, madVR adds dithering... isnt it much or enough dithering for a native dithering display technology as plasma?

Would it be fine or any upgrade on quality disabling dithering on CoreAVC and madVR?

[Razoola]

Ohh my, a new reclock version has turned up.. Pity it only looks to add power dvd12 support.

[Andy o]

Is it advisable enable dithering adding on filters? I have a plasma which already produces dithering by its own; CoreAVC adds dithering, madVR adds dithering... isnt it much or enough dithering for a native dithering display technology as plasma?

Would it be fine or any upgrade on quality disabling dithering on CoreAVC and madVR?

The decoder should add dithering only if it's processing (as in RGB conversion). If you let LAV or (I assume) CoreAVC pass the unprocessed NV12 output for instance, only madVR will dither. And yes, it's better to enable it in madVR. Your display is doing its own processing, so it needs to do its own dithering.

[chros]

Ok. Can you please check this build to confirm that the problem is gone for you:

http://madshi.net/madVRmaybeFixed.rar

Hi! If I'm right, I have the same problem than Razoola, this build doesn't fixes the problem for me...

Geforce v295.73 driver, 9600m GT (in laptop), win7x64:
- I have to check overshoot max frame latency to get rid of the presentation glitches in exclusive mode, but in that case render queue is: 3-4/12 (in windowed mode is full).

It's a laptop screen with 60Hz (I don't use any other display), and I don't use any refresh rate changer, the material is 23.976 ...

Thanks!

[madshi]

Hi! If I'm right, I have the same problem than Razoola, this build doesn't fixes the problem for me...
Please describe the problem you have in detail. Let me be the judge if it's the same problem or not. Right now I've no idea what your problem is exactly because you didn't describe it in enough detail.

Geforce v295.73 driver, 9600m GT (in laptop), win7x64:
- I have to check overshoot max frame latency to get rid of the presentation glitches in exclusive mode, but in that case render queue is: 3-4/12 (in windowed mode is full).
So what is the problem exactly? Do you get frame drops? Audio/video async? What exactly? And is it a new problem with v0.80 or did it also occur in older madVR versions?

[Razoola]

Hi! If I'm right, I have the same problem than Razoola, this build doesn't fixes the problem for me...

Geforce v295.73 driver, 9600m GT (in laptop), win7x64:
- I have to check overshoot max frame latency to get rid of the presentation glitches in exclusive mode, but in that case render queue is: 3-4/12 (in windowed mode is full).

It's a laptop screen with 60Hz (I don't use any other display), and I don't use any refresh rate changer, the material is 23.976 ...

Thanks!

I think you have a different issue to me, given you are already using an inconpatable refresh rate with the frame rate. Using the same build above please try the following exclusive mode settings and see if that helps with the presentation glitches count.

uncheck delay switch to exclusive mode
check present several frames in advance.
8 video frames to be presented in advance
then
dont flush
dont flush
dont flush
flush and wait (sleep)
Final three settings unchecked.

I also want to add that some people have reported issues with video and the 295.xx drivers in the nvidia forums so you may want to try 290.53 as thats the driver I'm still on.

[Xaurus]

I also want to add that some people have reported issues with video and the 295.xx drivers in the nvidia forums so you may want to try 290.53 as thats the driver I'm still on.
295.73 doesn't have the memory leak of the previous beta.

[Razoola]

295.73 doesn't have the memory leak of the previous beta.

I was more thinking about the color issues due to the driver not correctly picking up the type of connection is used to the display.. Granted that probably does not effect a laptop though.

[BeNooL]

I recommend that you open these in separate browser tabs (middle click, or ctrl+click) and swap between them to compare. (ctrl+tab to go forwards, ctrl+shift+tab to go back) They're posted in order of sharpness.


Or use CTRL + PgDown or PgUp to quickly navigate tabs.

Thanks for this extensive comparison by the way.

[chros]

Please describe the problem you have in detail. Let me be the judge if it's the same problem or not. Right now I've no idea what your problem is exactly because you didn't describe it in enough detail.


So what is the problem exactly? Do you get frame drops? Audio/video async? What exactly? And is it a new problem with v0.80 or did it also occur in older madVR versions?
Sorry for the lack of detail, I try to explain the problem.
- I get frame drops (0 delayed frame, 0 presentation glitches)
- It exists in 0.79, 0.80, 0.80maybefixed

When 'overshoot...' is checked (GPU flush settings: flush, sleep, don't, don't), after a couple of seconds the render queue is: 2-3/12 , the present queue is: 0-2/4 (all other queue is full) (So I think that's why I get frame drops.)

@Razoola: thanks for your suggestion, but it didn't helped ('overshoot...' is unchecked):
- dont+dont+dont+sleep - 20 pres.glitches , render queue full
- dont+dont+sleep+dont - 0 pres.glitches , render queue: 2-3/12, 1 dropped frame
- flush+sleep+flush+flush - 10 pres.glitches , render queue full
- flush+sleep+dont+dont (default) - 30 pres.glitches , render queue full

[Razoola]

@Razoola: thanks for your suggestion, but it didn't helped ('overshoot...' is unchecked):
- dont+dont+dont+sleep - 20 pres.glitches , render queue full
- dont+dont+sleep+dont - 0 pres.glitches , render queue: 2-3/12, 1 dropped frame
- flush+sleep+flush+flush - 10 pres.glitches , render queue full
- flush+sleep+dont+dont (default) - 30 pres.glitches , render queue full

With my suggested setting you are saying the render queue is staying full. That is what you want isn't it? You can live with presentation glitches providing they are not causing dropped or delayed frames or other onscreen artifacts.

079 was fine for me, the issue I had only came in 080. I feel it would help if you found where the problem started if previous versions were fine and you have not played with the scaling algos etc.

[madshi]

Sorry for the lack of detail, I try to explain the problem.
- I get frame drops (0 delayed frame, 0 presentation glitches)
- It exists in 0.79, 0.80, 0.80maybefixed

When 'overshoot...' is checked (GPU flush settings: flush, sleep, don't, don't), after a couple of seconds the render queue is: 2-3/12 , the present queue is: 0-2/4 (all other queue is full) (So I think that's why I get frame drops.)

@Razoola: thanks for your suggestion, but it didn't helped ('overshoot...' is unchecked):
- dont+dont+dont+sleep - 20 pres.glitches , render queue full
- dont+dont+sleep+dont - 0 pres.glitches , render queue: 2-3/12, 1 dropped frame
- flush+sleep+flush+flush - 10 pres.glitches , render queue full
- flush+sleep+dont+dont (default) - 30 pres.glitches , render queue full
NVidia drivers behave especially bad with 24fps content on 60Hz displays. If your display were 24Hz, you'd have much less problems. A workaround which works for me is activating D3D11 presentation and the option "limit render times to avoid glitches". Everything else set to default.

Alternatively you could try using the old exclusive mode rendering path (by unchecking the option "present several frames in advance"), or if windowed mode works alright for you, you could try disabling exclusive mode altogether.

[chros]

NVidia drivers behave especially bad with 24fps content on 60Hz displays. If your display were 24Hz, you'd have much less problems. A workaround which works for me is activating D3D11 presentation and the option "limit render times to avoid glitches". Everything else set to default.
Not working, either ...

1 last question: when I played the GPU flush settings I discover that the 'render queue' is 2-3/12 (so isn't full) in the following cases:
- when 'overshoot ...' checked
- when 'after copy to backbuffer' is 'sleep' or 'loop' ('overshoot ...' isn't checked)

Is it the intended behaivour of these settings regarding to the 'render queue'?

Thanks!

[madshi]

There is no real "intended behaviour" here. Ideally all queues should be full at all times and there should be no glitches and no drops. Unfortunately the GPU drivers offen are buggy and don't behave as expected. That's especially true for NVidia drivers when playing back 24fps movies on 60Hz displays. As I said, try the old exclusive mode or windowed mode.

[djsolidsnake86]

This is as intended because I don't trust the GPU algorithms to work reliably. I do plan to add some kind of color controls to madVR, but not very soon. If you want to calibrate your display properly, you can use the yCMS functionality. But for that you'd need to have your own meter.


They can't be simple added, and there's no way to calculate a proper "sum", because the processing partially overlaps.


Not yet.


The madVR OSD APIs only work if playback isn't stopped. The OSD only works in paused or playing mode.


What you say would be true only if the DXVA based scaling had the same quality as shader based scaling. That is not the case with current hardware. E.g. Intel's scaling currently only can do NV12 -> NV12 scaling, 8bit only, without dithering. I don't know what the future will bring, but you shouldn't praise DXVA scaling quality before it actually proves to produce quality worth praising.


I'm not sure I understand what you mean?


First of all you should check the madVR OSD to make sure that the frame drops are not caused by empty queues. If the frame drops occur even though all queues are full, then it's probably related to a refresh rate mismatch. This problem should not occur when you use Reclock, though. You should disable all VSync related options in Reclock, also make sure you reset the Reclock timing database.


Currently I'm manually blocking the "Pause" message for all languages I know. If you tell me the Polish/Czech "Pause" translation, I can block them, too.


Sadly, CoreAVC does not output any interlaced information. For Blu-Rays it's not much of a problem, just manually activate madVR's deinterlacing + IVTC and it should still work alright. In any case it won't harm to complain in the CoreAVC thread. I've tried to convince BetaBoy to output proper deinterlacing information, but it seems the CoreAVC devs don't like the idea. IMHO they're misguided in this case.


I'll check if I can reproduce this.

regarding colors, or color profiles would be very useful for who use a lcd monitor that make not so simple adjust colours on videos

[fastplayer]

regarding colors, or color profiles would be very useful for who use a lcd monitor that make not so simple adjust colours on videos
You know what would be really useful? You learning how to quote properly.

[Budtz]

I seem to be getting microstutters once in a while, but i am a bit unsure. There are no dropped frames duing a movie and no glitches and queues are rock solid. I am also using reclock. Is it even possible to have stuttering when the numbers show no errors duing af 2 hour period? I have a new tv from 2011 so it should be fine. driverversion and other software versions seems not to change anything.

[Razoola]

It may be repeated frames, those are currently not counted in the stats.

[FlashGordon]

That sounds quite interesting. Can I have a sample, or maybe even a couple of samples?

Is there a preferred tool to use to cut a sample from a VOB file? What about file size and length of the sample? Where do I upload it? Mediafire?

I also have a DVD that seems to have issues with IVTC, madvr recognizes some type of 3:2:3:2:2 pattern and applies IVTC but there still seems to be some combing. I know it could be the way the DVD was transferred, but it's worth a look I think.

[dansrfe]

Is there a preferred tool to use to cut a sample from a VOB file? What about file size and length of the sample? Where do I upload it? Mediafire?

I also have a DVD that seems to have issues with IVTC, madvr recognizes some type of 3:2:3:2:2 pattern and applies IVTC but there still seems to be some combing. I know it could be the way the DVD was transferred, but it's worth a look I think.

Use DGIndex to cut the VOB and yes mediafire is preferred by madshi along with everyone else that appreciates good file-sharing hosts.

[madshi]

I seem to be getting microstutters once in a while, but i am a bit unsure. There are no dropped frames duing a movie and no glitches and queues are rock solid. I am also using reclock. Is it even possible to have stuttering when the numbers show no errors duing af 2 hour period?
As Razoola says, it might be repeated frames. Try resetting the Reclock timing database. Are we talking about progressive or interlaced movies? In the latter case it might also be a deinterlacing issue.

What about file size and length of the sample?
The key thing is that when playing the sample alone, I should be able to see what you're seeing with the full DVD. For deinterlacing tests, at least 20 seconds of playtime per sample would be nice. Personally, I don't mind if you provide bigger samples than that. After all, I can cut them down myself, if I feel the need. So if you're in doubt, go bigger. Thanks!

I also have a DVD that seems to have issues with IVTC, madvr recognizes some type of 3:2:3:2:2 pattern and applies IVTC but there still seems to be some combing. I know it could be the way the DVD was transferred, but it's worth a look I think.
I can have a look, no problem. FWIW, until now all DVDs which madVR detected as 3:2:3:2:2 have been field blended. Meaning that some of the interlaced fields were a blend of the two surrounding fields. With such content IVTC doesn't really work. Anyway, samples are always appreciated, and I can have a look, just to make sure...

[chros]

I seem to be getting microstutters once in a while, but i am a bit unsure. There are no dropped frames duing a movie and no glitches and queues are rock solid. I am also using reclock. Is it even possible to have stuttering when the numbers show no errors duing af 2 hour period? I have a new tv from 2011 so it should be fine. driverversion and other software versions seems not to change anything.
Or maybe the TV's own algorithm? (eg. at LG is TrueMotion)
Is it certainly turned off?

[muu]

Really great project, thanks for all the work.

I'm having a strange issue with .80, maybe some one knows whats wrong?

When I play an h264 mkv it immediately fullscreens(but not like exclusive mode which is unchecked), it sort of plays as the desktop background, the start menu pops over it.
After a while the frame of MPC-HC pops up with an empty window, but its seekbar moves along in time with the video.

I tried unregistering .80 and using .79 but I'm getting the same behavior.
I followed this guide and used its settings for a madVR and LAV config: http://imouto.my/watching-h264-videos-using-compute-unified-device-architecture-cuda/
If I keep everything the same and just switch to Haali renderer then the video plays windowed as normal inside MPC-HC.

Here is a picture of whats happening:
http://i.imgur.com/n1qXls.jpg (http://i.imgur.com/n1qXl.jpg)

Are there any debugging steps to follow? I'm not too sure how to troubleshoot this.

[madshi]

Are you sure madVR is really used? This looks more like a cool Overlay feature to me... :)

[muu]

Yeah, I'm able to click the madVR icon in the tray and change settings, they seem to apply on the fly without restarting the player, and Ctrl+J works.

http://i.imgur.com/RKsDVs.jpg (http://i.imgur.com/RKsDV.jpg)

The MPC-HC right click menus like Filter > madVR seem to pop-under the video and never appear or take a long time to show up, same as MPC-HC's player window with the seekbar etc.
The video itself plays pretty smoothly the whole time, as if in exclusive mode, though its not.

Seems like a strange issue. I was hoping maybe someone had seen it before.

[madshi]

No, I haven't seen this before. Have you tried downloading a different MPC-HC version, just to be safe? Also check whether you activated any fullscreen options in MPC-HC and if so, try turning them off.

[muu]

It does look pretty unusual. I searched around and nobody seems to have reported anything similar, so I'm guessing it must be some conflict or configuration issue.

I tried the MPC-HC version recommended in that guide, but I checked with a few more

mpc-hc SSE tester dfr4087i
mpc-hc SSE2 tester dfr4075
mpc-homecinema.1.6.0.4014.x86
mpc-homecinema.1.6.1.4081.x86

They all behave the same way, and I have to kill the process to get the screen back in the end. I'm able to right click to Exit but it doesn't end the video.

I'm not aware of any fullscreen option I might have enabled, but double clicking in the video or Ctrl+Enter doesn't window it like usual, somehow its permanently fullscreened.

[kasper93]

options->output-> uncheck "D3D Fullscreen"

The problem is really annoying, when we hit ctrl+f in MPC-HC by mistake and didn't recognize it, then in next run this happens and only way to kill mpc-hc is to switch to other user and kill process, becouse everything goes under video and we can't even access task manager :P

[muu]

options->output-> disable "D3D Fullscreen"

The problem is really annoying, when we hit ctrl+f in MPC-HC by mistake and didn't recognize it, then in next run this happens and only way to kill mpc-hc is to switch to other user and kill process, becouse everything goes under video and we can't even access task manager :P

Well, that was it. I guess I should have posted to the MPC-HC thread. What threw me was that switching to Haali renderer would window the video as normal, but in madVR I had the unkillable fullscreen video. I'm still curious why that should be, but the issue is solved, its a player problem, although I never remember hitting Ctrl+F, but who knows.

Great work on madVR madshi and thanks for running into the problem and relating the solution kasper93.

[chros]

@Razoola: thanks for your suggestion, but it didn't helped ('overshoot...' is unchecked):
- dont+dont+dont+sleep - 20 pres.glitches , render queue full
- dont+dont+sleep+dont - 0 pres.glitches , render queue: 2-3/12, 1 dropped frame
- flush+sleep+flush+flush - 10 pres.glitches , render queue full
- flush+sleep+dont+dont (default) - 30 pres.glitches , render queue full
Well, I just want to report back that I don't really know what happend, but now I've got a solid playback with the new exclusive path without the 'overshoot...' option (60Hz). So maybe a driver thing was the cause (295.73) (although I didn't touch a thing)...

My settings are now:
flush+sleep+don't+flush

What I have experienced (since now it's working everything as it should be), that the last option ('...after D3D presentation') is responsible for the presentation glitches:
- if it set to 'don't' : I've got plenty of them
- if it set to any of the flush values: I've got just a small amount or nothing.

Thanks for this great renderer!!!

[madshi]

madVR v0.81 released

http://madshi.net/madVR.zip

* added automatic exception catching functionality
* added manual freeze report functionality (press Ctrl+Alt+Shift+Break)
* added option to switch display mode only when media player is in fullscreen
* added option to restore display mode when media player leaves fullscreen
* display mode changer now reacts on film (24p) vs. video (60p) content type
* improved renderer behaviour when display mode changes
* improved ivtc: faster detection of cadence changes
* improved ivtc: more robust against compression artifacts
* deint + ivtc is now auto activated for 60i movies which are tagged to 24 Hz
* ivtc is now auto activated for 50i movies which are tagged to 25 or 24 Hz
* deinterlacing is only turned on for 50i and 60i movies now
* added patch to MS VC-1 decoder -> deinterlacing off
* modified file name tagging logic (24p, 24i, 24fps, 24Hz, etc)
* improved behaviour with decoders which don't output proper interlaced flags
* improved exclusive -> windowed mode behaviour slightly, mainly for ZoomPlayer
* added "sharpness: 100" option for Bicubic scaling
* queue sizes are now always 16/8 by default, no more automatic adjustments
* keyboard shortcuts: always only change things temporarily by default now
* keyboard shortcuts: some changes (not all) can be stored by pressing F2 now
* the source levels (TV vs PC) can be stored now
* the source content type (film vs video) can be stored now
* key presses are now only "swallowed" if media player process has key focus
* osd keyboard handling runs in its own thread now
* added many more languages to the "Pause" MPC-HC message block
* used GPU RAM is now not shown in osd, anymore
* increased upload thread priority slightly
* fixed: ZoomPlayer eventually froze when leaving exclusive mode
* fixed: some users had composition rate in exclusive mode OSD -> problems
* fixed: crash when video window size was reduced to a very small size
* fixed: frame stepping sometimes resulted in weird play/pause mixed state
* fixed: one case where "delay playback start..." resulted in paused state
* fixed: v210 pitch/stride handling was broken
* fixed: YV24 had swapped chroma channels
* fixed: YV24 DXVA deinterlacing resulted in image corruption
* fixed: OSD API didn't capture mouse events on mouse down
Some additional notes:

(1) When a crash occurs inside of madVR, you should (hopefully) no longer get a crash messages from the OS. Instead you should get a nice exception box from madVR, allowing you to email a bug report to me. If the integrated emailing doesn't work, just close the exception box and you'll find a crash report text file on your desktop which you can then send to me (or upload it somewhere).

(2) If you get a media player freeze which you suspect is the fault of madVR, you can now press Ctrl+Alt+Shift+Break to get a freeze report. The report is stored to your desktop as a text file. Send this file to me, or upload it somewhere, then I'll have a look at it.

(3) If you switch the content type between film and video during playback, the display mode switcher will now automatically switch to a matching display mode (e.g. 1080p24 for film and 1080p60 for video). Of course the display mode changer will do that only if it is configured accordingly.

(4) I've modified the name tagging. In earlier madVR versions you could e.g. have a "24p" in the file name and then madVR treated the file as if it were a 24p file. After adding IVTC recently, I found that this logic didn't make too much sense. Most users who used this feature wanted to define the display mode, not redefine the movie frame rate. Anyway, v0.81 now allows you to do both. If you have e.g. "24fps" or "24 fps" or "24.000 fps" in the file name, madVR will ignore the framerate information coming from DirectShow and instead use your tag information. If you add "24p" or "24Hz" or "24 Hz" or "24.000 Hz" to the file name, madVR will use that only for the display mode changer. So the tag "24p" will no longer disable deinterlacing, as it did in older versions.

(5) In older versions the behaviour of the keyboard shortcuts was kind of a mess. Some shortcuts (e.g. source levels, primaries, decoding matrix etc) changed settings only for the duration of the madVR session, while other shortcuts (e.g. display levels, gamma settings etc) always changed the global settings. Now in v0.81 all changes you do via keyboard shortcuts will by default only affect the current madVR session. Your changes will be lost/reset when you load a new video file. Some changes can be stored globally, though, if you want that, by pressing F2 after having changed the settings by using the usual keyboard shortcuts. The madVR OSD will show an "F2 Save" message when you can store a changed setting. New things you can store are the source levels (PC vs video) and the content type (film vs video).

A note to media player developers:

In order to get really good freeze reports it would help if madVR could get access to some of the media player's debug information. Specifically the function names would be useful, so that the stack traces in the freeze reports can be better readable and understandable. There are 2 ways to provide madVR with this debug information:

(a) You can distribute a "yourMediaPlayer.pdb" file with your media player (the pdb file must be in the same folder as the exe file). madVR will then automatically read the pdb file and extract the necessary information. You can configure MSVC++ to create a pdb file even in release mode. Of course the pdb files are rather large, which is not so nice for distribution. So instead of shipping the pdb file:

(b) You can use my tool "madExceptPatch" to inject a compressed subset of your pdb file into your exe's resource section. Doing so will increase the size of your exe file by about 5-10%. The advantage of doing this is that your distribution package will now grow that much in size, and you also don't need to distribute any more files. The tool "madExceptPatch" is freely available here:

http://madshi.net/madExceptPatch.rar

(Delphi devs, who are interested in providing debug info, please contact me for an alternative solution.)

[WontonNoodle]

With automatic queue size adjustment I got super smooth playback, but now it is choppy for high bitrate videos. Should I set the queue sizes higher or lower?

[robpdotcom]

Big thanks for these:

* added option to switch display mode only when media player is in fullscreen
* added option to restore display mode when media player leaves fullscreen
* deint + ivtc is now auto activated for 60i movies which are tagged to 24 Hz
* ivtc is now auto activated for 50i movies which are tagged to 25 or 24 Hz
* deinterlacing is only turned on for 50i and 60i movies now

I can now sit back with my remote to watch telecined content - no more need for the keyboard :D

Can you explain what this does, or, how it is different from before?:
* display mode changer now reacts on film (24p) vs. video (60p) content type

Does it just mean that the display mode will change when you press CTRL+ALT+SHIFT+T?

Again, big thanks, and awesome work as always.

[Hypernova]

Thanks for the new version madshi.

With Windows 8 Consumer Preview (Beta) releasing this Wednesday, would you be interested in making exclusive mode working on it? Right now it doesn't work at all in the developer preview. I think it's because the present of this four "windows"

Immersive Background
Desktop
Charm Bar
Search Pane
Start Menu

I got these Window names from a multi-desktop program called VirtuaWin. I need to put these four, plus Class name Windows.UI.Core.CoreWindow into exception list to make it works correctly. I'm guessing that you can make madVR exclusive mode works again by just ignoring these windows. Right now it simply won't activated. Note that I'm not using exclusive mode myself, so I'm not really too bothered by this.

Off-topic: I use this VirtuaWin to keep all windows' position and size when I switch from my monitor to projector. This program hide the windows in such a way that even if your projector has a different (i.e. smaller) resolution, your windows' position won't get messed up. With this I don't have to deal with the headache of multi-monitor setup.

[madshi]

With automatic queue size adjustment I got super smooth playback, but now it is choppy for high bitrate videos. Should I set the queue sizes higher or lower?
I would recommend to temporarily install v0.80 again (download link for all older versions see first post in this thread), then look at the OSD (Ctrl+J) to see which queue sizes v0.80 was using. Then you can manually set the v0.81 queue sizes to the same value.

Can you explain what this does, or, how it is different from before?:

Does it just mean that the display mode will change when you press CTRL+ALT+SHIFT+T?
That, too. Basically the display mode changer understands now which effect the content type (film vs video) has on the needed display refresh rate. v0.80 didn't understand that.

With Windows 8 Consumer Preview (Beta) releasing this Wednesday, would you be interested in making exclusive mode working on it?
Not really.

[JarrettH]

Just a quick question...

If I have a DVD that obviously does not need deinterlacing (I trust my eyes if it looks better with IVTC) is there something I can report? It is this DVD...

Thanks for allowing screen switching to only happen in full screen and for allowing a default selection for film or video. Great changes :)

http://www.thedigitalbits.com/reviews/heat.html

[ryrynz]

Small typo in changelog.

* used GPU RAM is now not shown in OSD, anymore

Madshi, When I have ffdshow RAW enabled and Avisynth active with buffered ahead frames I get a previously rendered frame displayed on the screen just as file playback stops and transitions to the next file in the playlist.
This has been mentioned before but I've just now discovered what causes it.

Also I've noticed a couple of FSE mode initialization problems as I've been minimizing MPC thus switching from windowed mode to fullscreen a number of times on my i7 2600K.
When it comes up that initialization has failed it never appears to reattempt it until the player is restarted, I stay in windowed mode whilst in fullscreen even though FSE is enabled.
I never saw FSE issues with 0.80 but I'm having problems trying to reproduce it so I guess no big deal for now.

Thanks for the new build. :)

[Hypernova]

Not really.

OK. One more question and I will stop asking about this. Do you plan to support it when Windows 8 reach RTM then?

[JarrettH]

Since 0.80 there has been a problem of when I return from full screen to windowed mode that the window in MPC stays maximized

More to add:

Display rate shows 59.93hz
Composition rate shows 50hz
Movie fps 25.000 (says source filter)

Playing back the same DVD (Heat) with IVTC and no deinterlacing

Just saying that I expected the display rate to be my LCD refresh rate unless that is what composition rate is.

+

By letting madvr switch refresh rates when entering full screen this much lessens the chance of the Chrome min/windowed/close buttons disappearing. I think others mentioned this happening with Firefox. Also, when the refresh rate changes now it happens at a more appropriate time (in full screen) and allows clips you'd play for quick viewing to load fast with no refresh rate changing.

+

I haven't tested many titles, but this behaviour seems to be fixed now...MPC/madvr would not switch to windowed mode faithfully from full screen and instead give a macrovision error + closing the DVD playback. The first time is ok, but by the second or third there would be a good chance of the movie closing.

Just reporting good findings to you :)

[Andy o]

madshi, just to double check, is this working for you?
* deint + ivtc is now auto activated for 60i movies which are tagged to 24 Hz
When I tag an ATSC .wtv 60i file with "23p" or "23Hz", interlacing is on and display switches to 23p, but it's not switched automatically to film mode.

EDIT: It seems I have to disable automatic source type detection. btw, is auto working yet?

[nevcairiel]

* deint + ivtc is now auto activated for 60i movies which are tagged to 24 Hz
When I tag an ATSC .wtv 60i file with "23p" or "23Hz", interlacing is on and display switches to 23p, but it's not switched automatically to film mode.

The message says 24, did you try that as well, or only 23? :p

[Andy o]

Tried both, but see my edit above.

[robpdotcom]

Are you using LAVSplitter? If I use StreamBufferSource, madVR always switches to 59Hz for 59i material.

[Andy o]

You mean even when tagged to 23p? Otherwise, yes, it's the same with LAV Splitter as far as I can tell, as it should be.