Go Back   nV News Forums > Linux Support Forums > NVIDIA Linux

Newegg Daily Deals

Reply
 
Thread Tools
Old 01-27-09, 01:38 PM   #637
Stephen Warren
Moderator
 
Stephen Warren's Avatar
 
Join Date: Aug 2005
Posts: 1,327
Default Re: mplayer vdpau

Quote:
Originally Posted by quadtree View Post
Stephen,

I installed both version 180.25 of the driver and the latest mplayer vdpau patches from today (1-26). With only the driver changed, I had the same problem. Now that I have installed the lastest mplayer patches as well, things have improved a little.
I have a very hard time believing that either the driver update did not affect the hang, or that the MPlayer patches did. Are you sure that you were actually using the new driver once you installed it and before you upgraded MPlayer?

Quote:
Originally Posted by quadtree View Post
I no longer get the hang or the backtraces after vdp_presentation_queue_create(). Instead I get two error messages each time:

VDPAU nvidia: Error detected 1002 32 12
VDPAU nvidia: Error detected 2 183

Play then continues without hanging. However, I still see tearing as if it isn't really using overlays or the vsync is just off. Again, this only happens with the tv-out port. Using the analog VGA connector, everything works fine.

Unfortunately I had already installed the new driver and mplayer before trying to run it in gdb. If a backtrace from gdb would still be helpful let me know and I'll revert to an older mplayer.
OK. We did add some more error checking, plus a fix for a hang. Given that you're not seeing a hang any more, I believe (as I expected) that the hang you were seeing was the same one we fixed. Unless you see that hang again (with 180.25 or later), there's no need to try to reproduce it and get a backtrace.

So, your current problem is only that when using component out, you get tearing i.e. the blit-based instead of overlay-based presentation queue implementation. That issue is being tracked in this thread:

http://www.nvnews.net/vbulletin/show....php?p=1908816

I haven't had a chance to investigate that yet. Hopefully soon.
Stephen Warren is offline   Reply With Quote
Old 01-27-09, 01:40 PM   #638
Stephen Warren
Moderator
 
Stephen Warren's Avatar
 
Join Date: Aug 2005
Posts: 1,327
Default Re: mplayer vdpau

Quote:
Originally Posted by motd2k View Post
Small bug on vo_vdpau.c line 780 i think....


PHP Code:
max_height = FFMAX(max_height, d_width);
Oops. I thought I fixed that in the latest patch release. It is definitely fixed now! Thanks for pointing it out.
Stephen Warren is offline   Reply With Quote
Old 01-27-09, 01:40 PM   #639
ttabbal
Registered User
 
Join Date: Dec 2008
Posts: 14
Default Re: mplayer vdpau

Quote:
Originally Posted by jyavenard View Post
Hi

I thought VDPAU VC-1 decoding supports was only found on 9300/9400 kind of chipset?

An Asus M3n78-em (GeForce 8300 IGP) reports the following capabilities:

Decoder capabilities:

name level macbs width height

------------------------------------
MPEG1 0 8191 2032 2032
MPEG2_SIMPLE 3 8191 2032 2032
MPEG2_MAIN 3 8191 2032 2032
H264_MAIN 41 8191 2032 2032
H264_HIGH 41 8191 2032 2032
VC1_SIMPLE 1 8191 2032 2032
VC1_MAIN 2 8191 2032 2032
VC1_ADVANCED 4 8191 2032 2032

So what's the go ?

Another datapoint, from an M3N78-VM (GF8200).

Code:
Decoder capabilities:

name          level macbs width height
------------------------------------
MPEG1             0     2  4096  4096
MPEG2_SIMPLE      3     2  4096  4096
MPEG2_MAIN        3     2  4096  4096
H264_MAIN        41    16  4096  4096
H264_HIGH        41    16  4096  4096
VC1_SIMPLE        1     2  4096  4096
VC1_MAIN          2     2  4096  4096
VC1_ADVANCED      4     2  4096  4096
Not only does it report, I download the VC-1 sample from the first page of this thread, and mplayer says it's playing it. I'm remote using SSH right now, so I can't verify for sure. I'll edit this post tonight with the visual results. Interesting anyway. I expected mplayer would error out on me, it proceeds to try to play the video. The WMA audio codecs aren't compiled in, so it does give me a "Audio: no sound". But that's not a big deal.

Code:
VDec: vo config request - 1280 x 720 (preferred colorspace: VC1 ADVANCED VDPAU acceleration)
VDec: using VC1 ADVANCED VDPAU acceleration as output csp (no 0)
Movie-Aspect is undefined - no prescaling applied.
VO: [vdpau] 1280x720 => 1280x720 VC1 ADVANCED VDPAU acceleration
ttabbal is offline   Reply With Quote
Old 01-27-09, 01:44 PM   #640
Stephen Warren
Moderator
 
Stephen Warren's Avatar
 
Join Date: Aug 2005
Posts: 1,327
Default Re: mplayer vdpau

Yes, any card/motherboard using G98 or MCP 77/79 supports VC-1 with VDPAU. Describing exactly which cards/chipsets that refers to using marketing names is challenging. Hence, that's why the latest README uses chips names not marketing names when describing which GPUs are supported by VDPAU.
Stephen Warren is offline   Reply With Quote
Old 01-27-09, 01:49 PM   #641
quadtree
Registered User
 
Join Date: Jan 2009
Posts: 34
Default Re: mplayer vdpau

Quote:
Originally Posted by Stephen Warren View Post
I have a very hard time believing that either the driver update did not affect the hang, or that the MPlayer patches did. Are you sure that you were actually using the new driver once you installed it and before you upgraded MPlayer?
Your disbelief is well-founded, because I mispoke earlier. I did not see the hang after installing the new drivers. I still had the backtraces with VDPAU_TRACE=1, but no hang. After installing the new mplayer, the backtraces were replaced with the error messages. Sorry about the misinformation.
quadtree is offline   Reply With Quote
Old 01-27-09, 01:58 PM   #642
ttabbal
Registered User
 
Join Date: Dec 2008
Posts: 14
Default Re: mplayer vdpau

Quote:
Originally Posted by Stephen Warren View Post
Yes, any card/motherboard using G98 or MCP 77/79 supports VC-1 with VDPAU. Describing exactly which cards/chipsets that refers to using marketing names is challenging. Hence, that's why the latest README uses chips names not marketing names when describing which GPUs are supported by VDPAU.
I don't suppose there is any sort of list detailing what chips match what marketing names / product names? That was the biggest reason I posted. Someone on the mythTV mailing list asked me to run the vdpinfo utility because he wanted to know if that motherboard supported VC-1. I was convinced it didn't until I saw those results.

Anyone know what GPU the BFG 9800GT/OC has under the heatsink? I've got a brand new one in the box just sitting around. I got it as part of the AGP card upgrade program they did a while back.
ttabbal is offline   Reply With Quote
Old 01-27-09, 02:07 PM   #643
sl1pkn07
Registered User
 
Join Date: Nov 2006
Location: Spanishtan
Posts: 103
Default Re: mplayer vdpau

Stephen. any notice work vdpau cards with turbocache? (9300GS G98)
sl1pkn07 is offline   Reply With Quote
Old 01-27-09, 02:11 PM   #644
sl1pkn07
Registered User
 
Join Date: Nov 2006
Location: Spanishtan
Posts: 103
Default Re: mplayer vdpau

Sorry doble post


Quote:
Originally Posted by John Teg View Post
This code return 16 for all resolution (except 1920*1080 where it return 15) it's probably better to replace all of that by simply :

max_references = 16;

The code I use is :

Code:
if(vid_height > 720)
	max_references = 8;
else
	max_references = 16;
But it's only because my card seems to not be able to handle more than 8 references frames at 1080p.

And I have initialize NUM_VIDEO_SURFACES_H264 at 18 else some video didn't play (all the one using 16 references frames + b_pyramide).

My card is a 8500GT 512Mo (G86)
change to:

uint32_t surf_size = (round_width * round_height * 3) / 2;
max_references = 16;
if (max_references > 16) {
max_references = 16;
}

the video not working

vo_vdpau: get_image failed000 ct: -0.062 0/ 0 13% 10% 3.9% 10 0
[VD_FFMPEG] Only buffers allocated by vo_vdpau allowed.
mplayer: libmpcodecs/vd_ffmpeg.c:1198: vdpau_get_buffer: Assertion `0' failed.


the patch proposed by jyavenard not working.

VO: [vdpau] 1920x1078 => 1920x1078 H.264 HIGH VDPAU acceleration
Error 25 at libvo/vo_vdpau.c:990: -0.062 0/ 0 10% 8% 3.9% 7 0
sl1pkn07 is offline   Reply With Quote

Old 01-27-09, 04:26 PM   #645
The_Rebel
Registered User
 
Join Date: Jan 2009
Posts: 9
Default Re: mplayer vdpau

Quote:
Originally Posted by Stephen Warren View Post
On the screen with compiz, you're either just getting lucky and no seeing the tearing, or compiz itself is syncing to VBLANK when it renders (or perhaps you've turned that on in the OpenGL settings).

On the other screen, you have the composite extension enabled, so it uses the blit-based presentation queue implementation, which doesn't sync to VBLANK, and so tears.
I am using VBLANK syncing with opengl, but if i run compiz on both screens (separate X sessions) i get tearing on the secondary display.

Is there a way to disable compositing on a per screen basis?
The_Rebel is offline   Reply With Quote
Old 01-27-09, 04:54 PM   #646
Stephen Warren
Moderator
 
Stephen Warren's Avatar
 
Join Date: Aug 2005
Posts: 1,327
Default Re: mplayer vdpau

Quote:
Originally Posted by The_Rebel View Post
I am using VBLANK syncing with opengl, but if i run compiz on both screens (separate X sessions) i get tearing on the secondary display.
That makes sense; I believe OpenGL sync to VBLANK only syncs to a single display device's VBLANK, even when rendering to a region of the desktop/X display/screen/... scanned out by a different display device. I could be wrong though.

Quote:
Originally Posted by The_Rebel View Post
Is there a way to disable compositing on a per screen basis?
I don't think so; extensions are advertized on a per-X-display not per-screen basis.
Stephen Warren is offline   Reply With Quote
Old 01-27-09, 05:06 PM   #647
The_Rebel
Registered User
 
Join Date: Jan 2009
Posts: 9
Default Re: mplayer vdpau

Quote:
Originally Posted by Stephen Warren View Post
That makes sense; I believe OpenGL sync to VBLANK only syncs to a single display device's VBLANK, even when rendering to a region of the desktop/X display/screen/... scanned out by a different display device. I could be wrong though.



I don't think so; extensions are advertized on a per-X-display not per-screen basis.
Okay thanks for answering my questions, i'll wait for you guys to fix up the vsync'ing with
VDPAU.
The_Rebel is offline   Reply With Quote
Old 01-27-09, 05:31 PM   #648
mooninite
Registered User
 
Join Date: May 2006
Posts: 477
Default Re: mplayer vdpau

Quote:
Originally Posted by The_Rebel View Post
Okay thanks for answering my questions, i'll wait for you guys to fix up the vsync'ing with
VDPAU.
I don't think there is a fix planned?

It'd be nice if they [nVidia] would fix their strict-binding performance. Then you could run compiz for both screens and have no tearing.
mooninite is offline   Reply With Quote
Reply


Thread Tools

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


Similar Threads
Thread Thread Starter Forum Replies Last Post
xorg locks-up with newest nvidia drivers w/ vdpau. theroot NVIDIA Linux 1 06-24-12 11:04 AM
NVIDIA VDPAU Reference Benchmarks News Archived News Items 0 06-11-12 05:30 AM
VDPAU and KDE 4.8 compositing = slow Pie-rate NVIDIA Linux 30 05-23-12 07:07 AM
VDPAU testing tool crisalide NVIDIA Linux 392 04-29-12 06:01 PM
mplayer & xmms problems! replys2me NVIDIA Linux 5 09-06-02 02:34 PM

All times are GMT -5. The time now is 08:47 PM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright 1998 - 2014, nV News.