Announcement: libvdpau is moving to its own package

[AaronP]

I just made the first official release of libvdpau as its own standalone source package:
http://lists.freedesktop.org/archive...er/001036.html

[EDIT: A power failure during the X Developers Conference took down people.freedesktop.org, cgit.freedesktop.org, and anongit.freedesktop.org. These packages will be re-uploaded when freedesktop.org service is fully restored]
[EDIT 2: freedesktop.org service has been restored]

For users currently using VDPAU-based applications, we will continue to ship libvdpau and libvdpau_trace as part of the .run installer while distributions start to build their own packages for VDPAU. Once it is established in the distributions, we will stop including them in the .run package and users will need to start using the distro packages or building the library from source.


I also released an updated version of Wladimir J. van der Laan's vdpinfo tool:
http://lists.freedesktop.org/archive...er/001037.html

This version adds support for querying the new high-quality scaling modes, MPEG-4 part 2, and DivX.

[Stephen Warren]

As a follow-up to this, I'd like to hear from anyone who is using VDPAU packages supplied by their distribution, or possibly even third-party package repositories, in cases where their packaging doesn't support the correct split. the question is:

Are libvdpau.so* and libvdpau_trace.so* and vpdau*.h package in the same, or in a different package? If they are already packaged separately, and the wrapper/trace/header packages can be installed without installing the NVIDIA driver package, then there is no issue.

If there is an issue, please supply the following information:
* Main URL for the distribution's website
* Main URL of distribution's bug-tracker, or bug-reporting instructions
* Source and binary package names for the VDPAU-related packages
* A list of the files included in each VDPAU-related package (both the package used to run apps, and any separate -dev or -devel package)

Alternatively, if you're involved in other distros, feel free to report the issue directly to them. See the Ubuntu bug for an example: https://bugs.launchpad.net/ubuntu/+s...80/+bug/432172

I believe I have already fully covered the following distributions, in no particular order:
* Fedora
* Ubuntu
* Debian
* Mandriva
* Suse doesn't appear to provide VDPAU packages
* Slackware doesn't appear to provide VDPAU packages.

Thanks!

[Stephen Warren]

We have released a new version of our MPlayer patches that use pkg-config to locate libvdpau and vdpau*.h. See ftp://download.nvidia.com/XFree86/vd...728141.tar.bz2.

[deadlycheese]

When I run the checkout-patch-build.sh script, I get the following error in the 'configure' step:


Checking for DPMS ... yes (using Xdpms 4)
Checking for Xv ... yes
Checking for XvMC ... no
Checking for VDPAU ... ./configure: line 4289: syntax error near unexpected token `fi'
./configure: line 4289: ` fi'


I was able to workaround this error by adding the following line between lines 4289 and 4290:
noop=noop

I'm sure there is a more elegant patch but I don't know why it is breaking in the first place.

In any case, after making that fix and then rerunning './configure --disable-x264-lavc --disable-x264', I was able to build the patched mplayer successfully.


Just thought someone might care to know.

My system:

Fedora 11, x86_64

gcc --version
gcc (GCC) 4.4.1 20090725 (Red Hat 4.4.1-2)

bash-4.0-8.fc11.x86_64 (bash provides /bin/sh on Fedora)

[Stephen Warren]

If you replace this:

      if cc_check $_inc_tmp $_ld_tmp ; then
        libs_mplayer="$libs_mplayer $_ld_tmp"
        extra_cflags="$extra_cflags $_inc_tmp"
        _vdpau=yes
      else
      fi

with this:

      if cc_check $_inc_tmp $_ld_tmp ; then
        libs_mplayer="$libs_mplayer $_ld_tmp"
        extra_cflags="$extra_cflags $_inc_tmp"
        _vdpau=yes
      fi

Does that also solve it?

Thanks.

[deadlycheese]

Yeah, that fixes it as well, I should have just done that.


However, my built executable isn't able to play the "lys-20031106.avi" if I specify a vdpau codec. That is, this command line:

./mplayer -vo vdpau -vc ffodivxvdpau lys-20031106.avi

gives this output:

MPlayer SVN-r29354-4.4.1 (C) 2000-2009 MPlayer Team

Playing lys-20031106.avi.
AVI file format detected.
[aviheader] Video stream found, -vid 0
AVI: No audio stream found -> no sound.
VIDEO:  [DIVX]  1920x1080  24bpp  25.000 fps  8290.9 kbps (1012.1 kbyte/s)
Clip info:
 Software: MEncoder 0.90rc5-3.2.2 
==========================================================================
Forced video codec: ffodivxvdpau
Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
Unsupported PixelFormat -1
Unsupported PixelFormat -1
Selected video codec: [ffodivxvdpau] vfm: ffmpeg (FFmpeg MPEG-4,DIVX-4/5 (VDPAU))
==========================================================================
Audio: no sound
Starting playback...
[mpeg4_vdpau @ 0xcc12c0]decoding to PIX_FMT_NONE is not supported.
[VD_FFMPEG] Trying pixfmt=0.
Unsupported PixelFormat -1
VDec: vo config request - 1920 x 1080 (preferred colorspace: MPEG-4 Part 2 VDPAU acceleration)
VDec: using MPEG-4 Part 2 VDPAU acceleration as output csp (no 0)
Movie-Aspect is 1.78:1 - prescaling to correct movie aspect.
VO: [vdpau] 1920x1080 => 1920x1080 MPEG-4 Part 2 VDPAU acceleration 
[vdpau] Failed creating VDPAU decoder: An invalid/unsupported VdpDecoderProfile value was supplied.
FATAL: Cannot initialize video driver.
Unsupported PixelFormat -1
[mpeg4_vdpau @ 0xcc12c0]VDPAU decoding to PIX_FMT_NONE is not supported.
Error while decoding frame!

FATAL: Could not initialize video filters (-vf) or video output (-vo).


Exiting... (End of file)

I can play the file if I don't specify the codec, but then it uses nearly 100% CPU, so I know it isn't accelerated. I've tried several other formats, and it appears that I am only able to get H264 content to work with vdpau.

./mplayer -vc help | grep vdpau

ffmpeg12vdpau ffmpeg    working   FFmpeg MPEG-1/2 (VDPAU)  [mpegvideo_vdpau]
ffwmv3vdpau ffmpeg    problems  FFmpeg WMV3/WMV9 (VDPAU)  [wmv3_vdpau]
ffvc1vdpau  ffmpeg    problems  FFmpeg WVC1 (VDPAU)  [vc1_vdpau]
ffh264vdpau ffmpeg    working   FFmpeg H.264 (VDPAU)  [h264_vdpau]
ffodivxvdpau ffmpeg    working   FFmpeg MPEG-4,DIVX-4/5 (VDPAU)  [mpeg4_vdpau]

Any suggestions?

[cehoyos]

Originally Posted by deadlycheese

[vdpau] Failed creating VDPAU decoder: An invalid/unsupported VdpDecoderProfile value was supplied.

It is statistically unlikely that you own hardware that supports hardware accelerated decoding of MPEG-4:
http://en.wikipedia.org/wiki/VDPAU#Table_of_NVIDIA_GPUs
OTOH, not many people were able to test the code so far...

Carl Eugen

[Stephen Warren]

We have released a new version of our MPlayer patches that fix the problem deadlycheese saw. See ftp://download.nvidia.com/XFree86/vd...753835.tar.bz2.

[deadlycheese]

> It is statistically unlikely ...

heh. Well said. I was doing this testing on my laptop, with a Quadro NVS 140M chipset, and the wikipedia page you linked supports your assertion (VP2, Feature Set A: Partial acceleration for MPEG-1, MPEG-2, VC-1/WMV9, Complete acceleration for H.264).

Sadly, my newest GPU at home is a mere 9800GT, which would appear to have the same VDPAU capabilities, so I don't have anything more to offer as far as testing.

I do have a machine with a GeForce 8100 on the motherboard ... but, unfortunately, the wikipedia page has no reference to that particular chipset, so I don't know what I've got with it (as far as VDPAU).

Anyway, thanks for the help, and for the work on this project.

[cehoyos]

Originally Posted by deadlycheese

I do have a machine with a GeForce 8100 on the motherboard ... but, unfortunately, the wikipedia page has no reference to that particular chipset, so I don't know what I've got with it (as far as VDPAU).

Could you test it? I would have thought it supports feature set B, but this page suggests it supports no hardware accelerated decoding at all:
ftp://download.nvidia.com/XFree86/Li...ppendix-a.html

Carl Eugen

[Stephen Warren]

The GeForce 8100 is marketed without video features.

[deadlycheese]

my 8100 chipset is on a cheap Jetway mobo (PN78VM2-LF) that I'm using in a server, running headless ... not too easy to test out, sorry.

Interestingly, the box for the motherboard markets the 8100 as "For Microsoft DX10.0", "Support PureVideo technology" and "Support Blu-ray/Full HD DVD playback". The motherboard also has an HDMI port.

Hey, I just realized that I could just run the vdpauinfo utility on that box ... hang on ...
<couple hours of installing nvidia drivers, vnc twiddling, etc, before finally realizing that all I had to do was "DISPLAY=:0 vdpauinfo > vdpauinfo.out" ...>

OK, finally, here is the output of vdpauinfo on my machine with the 8100 chipset:

display: :0.0   screen: 0
API version: 0
Information string: NVIDIA VDPAU Driver Shared Library  190.36  Wed Sep 23 06:55:52 PDT 2009

Video surface:

name   width height types
-------------------------------------------
420     4096  4096  NV12 YV12 
422     4096  4096  UYVY YUYV 

Decoder capabilities:

name               level macbs width height
-------------------------------------------
MPEG1                 0  8192  2048  2048
MPEG2_SIMPLE          3  8192  2048  2048
MPEG2_MAIN            3  8192  2048  2048
H264_MAIN            41  8190  2032  2048
H264_HIGH            41  8190  2032  2048
VC1_SIMPLE            1  8190  2048  2048
VC1_MAIN              2  8190  2048  2048
VC1_ADVANCED          4  8190  2048  2048

Output surface:

name              width height nat types
----------------------------------------------------
B8G8R8A8          8192  8192    y  Y8U8V8A8 V8U8Y8A8 
R10G10B10A2       8192  8192    y  Y8U8V8A8 V8U8Y8A8 

Bitmap surface:

name              width height
------------------------------
B8G8R8A8          8192  8192
R8G8B8A8          8192  8192
R10G10B10A2       8192  8192
B10G10R10A2       8192  8192
A8                8192  8192

Video mixer:

feature name                    sup
------------------------------------
DEINTERLACE_TEMPORAL             y
DEINTERLACE_TEMPORAL_SPATIAL     y
INVERSE_TELECINE                 y
NOISE_REDUCTION                  y
SHARPNESS                        y
LUMA_KEY                         y
HIGH QUALITY SCALING - L1        -
HIGH QUALITY SCALING - L2        -
HIGH QUALITY SCALING - L3        -
HIGH QUALITY SCALING - L4        -
HIGH QUALITY SCALING - L5        -
HIGH QUALITY SCALING - L6        -
HIGH QUALITY SCALING - L7        -
HIGH QUALITY SCALING - L8        -
HIGH QUALITY SCALING - L9        -

parameter name                  sup      min      max
-----------------------------------------------------
VIDEO_SURFACE_WIDTH              y         1     4096
VIDEO_SURFACE_HEIGHT             y         1     4096
CHROMA_TYPE                      y  
LAYERS                           y         0        4

attribute name                  sup      min      max
-----------------------------------------------------
BACKGROUND_COLOR                 y  
CSC_MATRIX                       y  
NOISE_REDUCTION_LEVEL            y      0.00     1.00
SHARPNESS_LEVEL                  y     -1.00     1.00
LUMA_KEY_MIN_LUMA                y  
LUMA_KEY_MAX_LUMA                y

This is almost identical to the output from my NVS 140M chipset:

$ diff -b vdpauinfo.nvs140M.out vdpauinfo.8100.out 
19,20c19,20
< H264_MAIN            41  8192  2048  2048
< H264_HIGH            41  8192  2048  2048
---
> H264_MAIN            41  8190  2032  2048
> H264_HIGH            41  8190  2032  2048

So, if we can trust the vdpauinfo information, I would expect the 8100 to have the same capabilities as the NVS 140M chipset in my laptop.