Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > Hardware & Software > Software players

Reply
 
Thread Tools Search this Thread Display Modes
Old 25th September 2012, 15:31   #14161  |  Link
SamKook
Registered User
 
Join Date: Mar 2011
Posts: 216
Quote:
Originally Posted by madshi View Post
Can I have a debug log with v0.83.4, please?
It took me a while to figure out that it was created on the desktop instead of the video folder or MadVR install folder(a message that say's where the log is created when you activate debug mode would be nice) but here it is: http://www.mediafire.com/file/6pe56fufjr6i2c6/madVR_log_SamKook_2012-09-25.rar
SamKook is offline   Reply With Quote
Old 25th September 2012, 16:05   #14162  |  Link
cyberbeing
Broadband Junkie
 
Join Date: Oct 2005
Posts: 1,859
Crash when activating FFDShow's resize filter during playback with horizontal size set to 4096.

Crash & Debug logs
cyberbeing is offline   Reply With Quote
Old 25th September 2012, 16:23   #14163  |  Link
oddball
Registered User
 
Join Date: Jan 2002
Posts: 1,264
I think I figured it out. Needed to add 1080p29
oddball is offline   Reply With Quote
Old 25th September 2012, 16:37   #14164  |  Link
CiNcH
Registered User
 
CiNcH's Avatar
 
Join Date: Jan 2004
Posts: 567
I just checked clock deviation with my newly built HTPC which is based on a Core i3-3225 with HD 4000 graphics. HDMI Audio is supplied by the OnBoard RealTek ALC898 (MSI Z77A-GD65). I tried 576i25 (deinterlaced to 50fps @ 50Hz) and 1080p (@23Hz) source material.

Here is what madVR displayed after about 1 hour of playback...

576i25:
display 50.00090Hz [diff: +0.0009]
clock deviation 0.00180%
A frame drop is supposed to occur every few days, number heavily fluctuates

1080p:
display 23.97227 [diff: -0,00373]
clock deviation 0.00090%
1 frame drop every 4.18 minutes


At 50Hz, the actual refresh rate is much closer to the target. Still the calculated clock deviation is higher compared to the 23Hz mode. The clock deviation should be much smaller for the more accurate 50Hz mode, shouldn't it? Don't you derive the VSync clock from the D3D raster status? Drop rate seems to more closely reflect the refresh rate accuracy.
__________________
Bye
CiNcH is offline   Reply With Quote
Old 25th September 2012, 16:56   #14165  |  Link
cyberbeing
Broadband Junkie
 
Join Date: Oct 2005
Posts: 1,859
Quote:
Originally Posted by CiNcH View Post
576i25:
display 50.00090Hz [diff: +0.0009]
clock deviation 0.00180%
A frame drop is supposed to occur every few days, number heavily fluctuates
Optimal Refresh rate: 50.00090Hz (diff: 0.00000)

Quote:
Originally Posted by CiNcH View Post
1080p:
display 23.97227 [diff: -0,00373]
clock deviation 0.00090%
1 frame drop every 4.18 minutes
Optimal Refresh rate: 23.97624Hz (diff: 0.00397)

Quote:
Originally Posted by CiNcH View Post
The clock deviation should be much smaller for the more accurate 50Hz mode, shouldn't it?
Not necessarily, since that's not what clock deviation represents.

Last edited by cyberbeing; 25th September 2012 at 16:59.
cyberbeing is offline   Reply With Quote
Old 25th September 2012, 17:11   #14166  |  Link
CiNcH
Registered User
 
CiNcH's Avatar
 
Join Date: Jan 2004
Posts: 567
Quote:
Not necessarily, since that's not what clock deviation represents.
But a higher clock deviation should result in a higher drop rate. Above it is the other way around. Maybe madVR does not handle a negative deviation properly? In this case, no frame should be dropped but repeated..
__________________
Bye
CiNcH is offline   Reply With Quote
Old 25th September 2012, 17:16   #14167  |  Link
cyberbeing
Broadband Junkie
 
Join Date: Oct 2005
Posts: 1,859
Quote:
Originally Posted by CiNcH View Post
But a higher clock deviation should result in a higher drop rate.
That's absolutely false. You could theoretically have a massive clock deviation with a dropped frame interval in days.

In simple terms, you could probably think of clock deviation as a representation of how much faster or slower your actual playback rate is compared to your video framerate.
cyberbeing is offline   Reply With Quote
Old 25th September 2012, 17:21   #14168  |  Link
CiNcH
Registered User
 
CiNcH's Avatar
 
Join Date: Jan 2004
Posts: 567
Quote:
In simple terms, you could probably think of clock deviation as a representation of how much faster or slower your actual playback rate is compared to your video framerate.
I actually think of it as the speed difference between video playback (based on graph reference clock which is the audio clock) and video output (VSync based on GPU clock). To compensate for the two rates, frames have to either be dropped or repeated, depending on the drift direction. The higher the deviation between the two, the more frames have to be repeated/dropped.
__________________
Bye

Last edited by CiNcH; 25th September 2012 at 17:23.
CiNcH is offline   Reply With Quote
Old 25th September 2012, 17:35   #14169  |  Link
cyberbeing
Broadband Junkie
 
Join Date: Oct 2005
Posts: 1,859
Quote:
Originally Posted by CiNcH View Post
I actually think of it as the speed difference between video playback (based on graph reference clock which is the audio clock) and video output (VSync based on GPU clock). To compensate for the two rates, frames have to either be dropped or repeated, depending on the drift direction. The higher the deviation between the two, the more frames have to be repeated/dropped.
That's all well and good, but that's completely different from what madVR's clock deviation calculation represents.
cyberbeing is offline   Reply With Quote
Old 25th September 2012, 18:21   #14170  |  Link
fairchild
Registered User
 
Join Date: Sep 2010
Posts: 321
I just watched The Avenger's Blu-ray with v0.83.4 and not a single drop frame or glitch while using FSE (while using Lav filters + reclock), great job madshi!
__________________
MPC-HC/MPC-BE, Lav Filters, MadVR
CPU: AMD Ryzen 5 1600, Video: AMD Radeon RX Vega 56 -> TCL S405 55", Audio: Audio-Technica M50S
fairchild is offline   Reply With Quote
Old 25th September 2012, 18:52   #14171  |  Link
mark0077
Registered User
 
Join Date: Apr 2008
Posts: 1,106
Quote:
Originally Posted by madshi View Post
Ok, sorry, please try with v0.83.4.
Debug log attached from 0.83.4 after seeing the Direct3D error. With mpc-hc closed, I opened a file directly in FSE, and then double clicked mpc-hc's window to get to overlay mode and saw the Direct3D error appear, mpc-hc also paused itself after the error.

http://www.sendspace.com/file/ojmmei

I also have an issue, I'm not sure if its expected. Resizing the windowed overlay mode shows strange looking effects inside the window, like a stair step scaling effect, with a dark red background.

Last edited by mark0077; 25th September 2012 at 19:02.
mark0077 is offline   Reply With Quote
Old 25th September 2012, 20:45   #14172  |  Link
Pat357
Registered User
 
Join Date: Jun 2006
Posts: 452
Quote:
Originally Posted by mark0077 View Post
Debug log attached from 0.83.4 after seeing the Direct3D error. With mpc-hc closed, I opened a file directly in FSE, and then double clicked mpc-hc's window to get to overlay mode and saw the Direct3D error appear, mpc-hc also paused itself after the error.
I had the same error a few times : always when switching from FSE to overlay.
I think madshi could try to speed up the destruction from the D3D-surface that is left from FSE on the moment you switch to overlay.
Because FSE "holds" the D3D until it's destructed, Overlay can't create a new one.
Another option could be not to destroy the D3D from FSE, but.., if this is somehow feasible, re-use it instead for Overlay . The minimum time gain would be the time needed for destruction + creation from a new D3D-surface.

Quote:
I also have an issue, I'm not sure if its expected. Resizing the windowed overlay mode shows strange looking effects inside the window, like a stair step scaling effect, with a dark red background.
Are you resizing using the mouse ?
For me, If I use the keyboard (ctrl +1, ...) to resize to a predefined size, the process is smooth.
But if I use the mouse to resize, I see indeed strange things, because I'm in fact resizing to a changing target-size in every step, until I finally have the size that I wanted.

Last edited by Pat357; 26th September 2012 at 00:20. Reason: typo's
Pat357 is offline   Reply With Quote
Old 25th September 2012, 22:04   #14173  |  Link
yok833
Registered User
 
Join Date: Aug 2012
Posts: 73
I run the last version of Madvr + AMD 6850.
When I use FSE, I have dropped frame but if I stay in FS, everything is very smooth...??? Before it was completely the opposite???
Have I made something bad with MPCHC??? I have tried many times but I cannot fix this issue :5
yok833 is offline   Reply With Quote
Old 25th September 2012, 23:46   #14174  |  Link
TheShadowRunner
Registered User
 
TheShadowRunner's Avatar
 
Join Date: Feb 2004
Posts: 399
Hi madshi, could you possibly make the new "slower/delay stopping playback" optional ?
Here it's significantly slower (than 0.82.5) and brings absolutely nothing (madVR has _never_ crashed on stopping playback/exiting player). On XP SP3/ ZoomPlayer..
__________________
XP SP3 / Geforce 8500 / Zoom Player
TheShadowRunner is offline   Reply With Quote
Old 25th September 2012, 23:58   #14175  |  Link
Blight
Software Developer
 
Blight's Avatar
 
Join Date: Oct 2001
Location: Israel
Posts: 1,005
Prinz:
Here's a MadExcept enabled version of ZP's latest MAX build:
https://dl.dropbox.com/u/104094858/zplayer_madExcept.rar
__________________
Yaron Gur
Zoom Player . Lead Developer
Blight is offline   Reply With Quote
Old 26th September 2012, 00:03   #14176  |  Link
Blight
Software Developer
 
Blight's Avatar
 
Join Date: Oct 2001
Location: Israel
Posts: 1,005
Madshi:
Yes, enabling overlay showing up a black screen until a new graph is loaded was with Aero Enabled.
I'll do more tests and provide logs once I can free up more time :/
__________________
Yaron Gur
Zoom Player . Lead Developer
Blight is offline   Reply With Quote
Old 26th September 2012, 04:12   #14177  |  Link
andybkma
Registered User
 
Join Date: Sep 2006
Posts: 212
Quote:
Originally Posted by TheShadowRunner View Post
Hi madshi, could you possibly make the new "slower/delay stopping playback" optional ?
Here it's significantly slower (than 0.82.5) and brings absolutely nothing (madVR has _never_ crashed on stopping playback/exiting player). On XP SP3/ ZoomPlayer..
I second this request. Have never experienced an mVR crashing problem upon stopping or exiting player.
andybkma is offline   Reply With Quote
Old 26th September 2012, 09:22   #14178  |  Link
Prinz
Registered User
 
Join Date: Jul 2011
Posts: 83
Quote:
Originally Posted by Blight View Post
Prinz:
Here's a MadExcept enabled version of ZP's latest MAX build:
https://dl.dropbox.com/u/104094858/zplayer_madExcept.rar
Bugreport
Prinz is offline   Reply With Quote
Old 26th September 2012, 09:34   #14179  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,140
Quote:
Originally Posted by SamKook View Post
It took me a while to figure out that it was created on the desktop instead of the video folder or MadVR install folder(a message that say's where the log is created when you activate debug mode would be nice) but here it is: http://www.mediafire.com/file/6pe56fufjr6i2c6/madVR_log_SamKook_2012-09-25.rar
Quote:
Originally Posted by mark0077 View Post
Debug log attached from 0.83.4 after seeing the Direct3D error. With mpc-hc closed, I opened a file directly in FSE, and then double clicked mpc-hc's window to get to overlay mode and saw the Direct3D error appear, mpc-hc also paused itself after the error.

http://www.sendspace.com/file/ojmmei
Thanks, I'll have a look at those logs.

Quote:
Originally Posted by mark0077 View Post
I also have an issue, I'm not sure if its expected. Resizing the windowed overlay mode shows strange looking effects inside the window, like a stair step scaling effect, with a dark red background.
I fear there's not much I can do about it. It seems this effect is caused by the Overlay color key not being update quickly enough.

Quote:
Originally Posted by cyberbeing View Post
Crash when activating FFDShow's resize filter during playback with horizontal size set to 4096.

Crash & Debug logs
This looks like a bug in ffdshow to me. Basically ffdshow is sending a buffer which is too small to actually contain a frame in the size it says it's sending.

Quote:
Originally Posted by Pat357 View Post
Another option could be not to destroy the D3D from FSE, but.., if this is somehow feasible, re-use it instead for Overlay .
But that's what I'm already doing!!

Quote:
Originally Posted by yok833 View Post
I run the last version of Madvr + AMD 6850.
When I use FSE, I have dropped frame but if I stay in FS, everything is very smooth...??? Before it was completely the opposite???
Have I made something bad with MPCHC??? I have tried many times but I cannot fix this issue :5
That is weird. In FSE mode, in the madVR debug OSD (Ctrl+J) are all queues full? Or are some of them empty or near empty? Which ones? Maybe you can make a screenshot of the debug OSD in the moment when you get frame drops?

Quote:
Originally Posted by TheShadowRunner View Post
Hi madshi, could you possibly make the new "slower/delay stopping playback" optional ?
Here it's significantly slower (than 0.82.5) and brings absolutely nothing (madVR has _never_ crashed on stopping playback/exiting player). On XP SP3/ ZoomPlayer..
Quote:
Originally Posted by andybkma View Post
I second this request. Have never experienced an mVR crashing problem upon stopping or exiting player.
Oh well, I can try, but it will be difficult. I can't simply go back to the old solution, that's impossible due to some other changes I've done (allowing multiple madVR instances at the same time).

Quote:
Originally Posted by Blight View Post
Yes, enabling overlay showing up a black screen until a new graph is loaded was with Aero Enabled.
I'll do more tests and provide logs once I can free up more time :/
Thanks, but more tests and logs are not necessary for this specific problem. I can reproduce it, and I've found no way to fix it. It seems to be an NVidia driver bug. It would be interesting whether the same problem occurs with Intel, too, though. Maybe it's a bug in D3D and not in the GPU driver? In any case, I don't think I can fix it. I've really tried, even totally destroyed and recreated the D3D devices, but it doesn't help.

Quote:
Originally Posted by Prinz View Post
@Blight, what does ZP do here:

> dwTaskbarThumbnails 331 +7 TdwTaskbarThumbnails.DoUpdate

Could that code line result in an EOleException exception somehow?
madshi is offline   Reply With Quote
Old 26th September 2012, 10:07   #14180  |  Link
cyberbeing
Broadband Junkie
 
Join Date: Oct 2005
Posts: 1,859
Quote:
Originally Posted by madshi View Post
This looks like a bug in ffdshow to me. Basically ffdshow is sending a buffer which is too small to actually contain a frame in the size it says it's sending.
Is there nothing you can do about it then? It seems every video renderer in MPC-HC except madVR is able to handle that case fine.
cyberbeing is offline   Reply With Quote
Reply

Tags
direct compute, dithering, error diffusion, madvr, ngu, nnedi3, quality, renderer, scaling, uhd upscaling, upsampling

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 20:05.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.