Page 2 of 3 FirstFirst 123 LastLast
Results 11 to 20 of 23

Thread: Which power must be a cpu for HD 1080p?

  1. #11
    Join Date
    Apr 2010
    Location
    Italy
    Posts
    92

    Default

    I don't find ffplay on my repository.

    I use the commands:
    $ mplayer 00040.MTS
    $ mplayer -lavdopts threads=2 00040.MTS
    $ mplayer-mt 00040.MTS
    but the final result regarding the use of the cpu are always the same.

    I use the packages available on Debian.

    Note:
    It's a free video at 17 mbps available fontana di sanremo (notte) 17mbps.
    It's only a test video about an owner of a Sony HDR-CX550VE.
    [Sony HDR-CX550VE e altre VCR] Video Sample e Opinioni

  2. #12
    Join Date
    Apr 2010
    Location
    Italy
    Posts
    92

    Default

    Using
    mplayer-mt -lavdopts threads=2 00040.MTS
    i obtain:
    75 %
    30 %
    a little better then before.

  3. #13
    Join Date
    Dec 2009
    Location
    Italy
    Posts
    176

    Default

    git clone git://gitorious.org/~astrange/ffmpeg/ffmpeg-mt.git

    Here you can download ffmpeg-mt, which contains ffplay (you will need to build it from source). Cpu usage is still strange, have you tried with other videos?

  4. #14
    Join Date
    Dec 2009
    Location
    Italy
    Posts
    176

    Default

    Ok, I've built especially for you ffmpeg-mt. On the systems I tried it works as expected (load balanced between the cores).
    You can get it at http://www.mattea.info/ffmpeg/ffmpeg...21-1_amd64.deb
    Everything is installed in /usr/local so it doesn't conflict with system packages. After installation, you should run "ldconfig" as root to detect the new libraries. To play a file, use ffplay -threads N filename.

  5. #15
    Join Date
    Apr 2010
    Location
    Italy
    Posts
    92

    Default

    Thanks for the deb.

    Unfortunately the problem remain the same, also ffmpeg work worst respect mplayer and mplayer-mt.
    During the start, the video don't run immediately (as happen instead using both versions of mplayer).
    Also the video result a lot less fluent respect mplayer.

  6. #16
    Join Date
    Dec 2009
    Location
    Italy
    Posts
    176

    Default

    You're not lucky
    Here (http://www.mattea.info/mplayer/mplay...22-1_amd64.deb) you can download a deb of mplayer-mt (latest svn). As always, it installs everything in /usr/local, so you'll need to call it using /usr/local/bin/mplayer. Could you try using it with these options:
    -vc ffh264 -vo null -benchmark -nosound -lavdopts threads=N
    and use system monitor to see if all your cores are used? On my system it happened that since one core was more than enough to play the file (Phenom II 940 BE) not all the cores were equally used with normal playback even if I set threads to 4, but with the benchmark option every core jumped at 90%.

  7. #17
    Join Date
    Apr 2010
    Location
    Italy
    Posts
    92

    Default

    Thanks.
    $ /usr/local/bin/mplayer -vc ffh264 -vo null -benchmark -nosound -lavdopts threads=2 00040.MTS
    MPlayer SVN-r31054-4.4.1 (C) 2000-2010 MPlayer Team

    Playing 00040.MTS.
    TS file format detected.
    VIDEO H264(pid=4113) NO AUDIO! NO SUBS (yet)! PROGRAM N. 1
    FPS seems to be: 25.000000
    ================================================== ========================
    Forced video codec: ffh264
    Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
    Selected video codec: [ffh264] vfm: ffmpeg (FFmpeg H.264)
    ================================================== ========================
    Audio: no sound
    Starting playback...
    Movie-Aspect is 1.78:1 - prescaling to correct movie aspect.
    VO: [null] 1920x1080 => 1920x1080 Planar YV12
    V: 23.3 1117/1117 39% 0% 0.0% 0 0

    BENCHMARKs: VC: 17.519s VO: 0.002s A: 0.000s Sys: 0.360s = 17.882s
    BENCHMARK%: VC: 97.9735% VO: 0.0129% A: 0.0000% Sys: 2.0136% = 100.0000%

    Exiting... (End of file)
    I obtain 175 %.

    So i tried even the mplayer (not mt) present on the system:
    mplayer -vc ffh264 -vo null -benchmark -nosound -lavdopts threads=2 00040.MTS
    and i obtain the same result.

    While i run the video the cpu is below 100 % but in both case (mplayer or mplayer-mt) both cpus are use more equally (in precedence i read the wrong line on top).
    I don't see difference with mplayer-mt.

    $ /usr/local/bin/mplayer -lavdopts threads=2 00040.MTS
    MPlayer SVN-r31054-4.4.1 (C) 2000-2010 MPlayer Team

    Playing 00040.MTS.
    TS file format detected.
    VIDEO H264(pid=4113) AUDIO A52(pid=4352) NO SUBS (yet)! PROGRAM N. 1
    FPS seems to be: 25.000000
    [VO_XV] It seems there is no Xvideo support for your video card available.
    [VO_XV] Run 'xvinfo' to verify its Xv support and read
    [VO_XV] DOCS/HTML/en/video.html#xv!
    [VO_XV] See 'mplayer -vo help' for other (non-xv) video out drivers.
    [VO_XV] Try -vo x11.
    ================================================== ========================
    Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
    Selected video codec: [ffh264] vfm: ffmpeg (FFmpeg H.264)
    ================================================== ========================
    ================================================== ========================
    Opening audio decoder: [ffmpeg] FFmpeg/libavcodec audio decoders
    AUDIO: 48000 Hz, 2 ch, s16le, 448.0 kbit/29.17% (ratio: 56000->192000)
    Selected audio codec: [ffac3] afm: ffmpeg (FFmpeg AC-3)
    ================================================== ========================
    AO: [oss] 48000Hz 2ch s16le (2 bytes per sample)
    Starting playback...
    Movie-Aspect is 1.78:1 - prescaling to correct movie aspect.
    VO: [x11] 1920x1080 => 1920x1080 Planar YV12
    [swscaler @ 0xe40aa0]using unscaled yuv420p -> bgra special converter
    A: 23.4 V: 23.3 A-V: 0.088 ct: 4.381 1117/1117 14% 17% 0.8% 12 0

    Exiting... (End of file)
    $ xvinfo
    X-Video Extension version 2.2
    screen #0
    no adaptors present
    Note:
    However kbios i'm interested how you have build that packages, there is some link that could i see?
    The only package where i'm more confortable are archlinux packages ...
    When i build simple packages on debian or ubuntu i use checkconfig.

  8. #18
    Join Date
    Dec 2009
    Location
    Italy
    Posts
    176

    Default

    The basic mplayer has support for multiple threads/cores if the video is encoded using slices, so that might be the reason you don't see a difference: if you try with a video which doesn't use them, mplayer-mt results should be vastly superior.

    About my packages: I have never heard of checkconfig, they were built with checkinstall (./configure; make; sudo checkinstall).

  9. #19
    Join Date
    Apr 2010
    Location
    Italy
    Posts
    92

    Default

    Yes is checkinstall i'm write wrong the command.

    Thanks.

  10. #20
    Join Date
    Aug 2007
    Posts
    6,645

    Default

    Well i would not use checkinstall, especially when i have got so nice scripts for that

    but i also did some benchmarking - git-mt was patched to support vc1 too but it does not seem to do anything.

    maybe more cores should be tested...

    100 mb m2ts h264:

    mplayer -vo null -benchmark -nosound -lavdopts threads=X testfile

    mplayer svn 1 thread: 11.7s
    mplayer svn 2 threads: 7.9s
    mplayer git-mt 1 thread: 11.9s
    mplayer git-mt 2 threads: 7.0s

    100 mb m2ts vc1 (no diff visable):

    mplayer svn 1 thread: 18.2s
    mplayer svn 2 threads: 18.2s
    mplayer git-mt 1 thread: 18.2s
    mplayer git-mt 2 threads: 18.2s

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •