Page 1 of 2 12 LastLast
Results 1 to 10 of 18

Thread: FFmpeg's VP8 Decoder Blasts Google's Decoder

  1. #1
    Join Date
    Jan 2007
    Posts
    14,643

    Default FFmpeg's VP8 Decoder Blasts Google's Decoder

    Phoronix: FFmpeg's VP8 Decoder Blasts Google's Decoder

    It was just back in May that Google opened up the VP8 video format that they got their hands on through the acquisition of On2 and at the same time they created the WebM container format. VP8 has already received a lot of love by the open-source community -- both developers and end-users -- and support for it has already worked its way into FFmpeg, GStreamer, and other multimedia projects. Google released the libvpx library as their official VP8 decoder library, but now the FFmpeg developers have created their own decoder and it's shockingly faster than that of Google's own open-source library...

    http://www.phoronix.com/vr.php?view=ODQ0NQ

  2. #2
    Join Date
    Aug 2007
    Location
    Poland
    Posts
    215

    Default

    I wonder what's the performance of libvpx compared to ffvpx on P4 32bit and simmilar machines (32bit, only one core (or HT) and high CPU clock (3GHz).

  3. #3
    Join Date
    Sep 2007
    Location
    Connecticut,USA
    Posts
    956

    Default

    Quote Originally Posted by xeros View Post
    I wonder what's the performance of libvpx compared to ffvpx on P4 32bit and simmilar machines (32bit, only one core (or HT) and high CPU clock (3GHz).
    I'm sure it'll do quite well even on a uniprocessor (or hyperthreading) P4 if the ffvp8 library is compiled with the right compiler flags.

  4. #4
    Join Date
    Oct 2009
    Posts
    59

    Default

    Quote Originally Posted by DeepDayze View Post
    I'm sure it'll do quite well even on a uniprocessor (or hyperthreading) P4 if the ffvp8 library is compiled with the right compiler flags.
    Yep, FFmpeg devs tend to make sure that things work well on all numbers of cores. so far as to say that any change for multi-core systems should not degrade single core performance. At least for h264, I would expect the same treatment for webM/vp8

  5. #5
    Join Date
    Jan 2007
    Location
    Germany
    Posts
    2,124

    Default

    It is said to do 26fps with 1080p even on the Intel Atom which is excellent.

  6. #6
    Join Date
    Sep 2007
    Location
    Connecticut,USA
    Posts
    956

    Default

    I'd love to be able to watch quality 1080p content even on my aging Thinkpad T42

  7. #7
    Join Date
    Aug 2007
    Posts
    6,615

    Default

    Well did you see that the bitrate is only 4.5 mbit for those videos? Of course it is 1080p, but what matters is the bitrate. Well youtube will most likely not use much more, so it should be possible to decode it. But html5 still misses a good full screen playback.

  8. #8
    Join Date
    Sep 2007
    Location
    Connecticut,USA
    Posts
    956

    Default

    Quote Originally Posted by Kano View Post
    Well did you see that the bitrate is only 4.5 mbit for those videos? Of course it is 1080p, but what matters is the bitrate. Well youtube will most likely not use much more, so it should be possible to decode it. But html5 still misses a good full screen playback.
    I'm sure that shortcoming will be fixed eventually once VP8 encoded video hits the mainstream

  9. #9
    Join Date
    Feb 2009
    Location
    Belgium
    Posts
    115

    Default

    Hopefully it won't take too long before graphic chipset producers add VP8 decoding support to their GPUs.

    I'd really like to see how much VP8 differs from H.264 in file size and quality for 720p and 1080p videos.

  10. #10
    Join Date
    Oct 2008
    Posts
    143

    Default

    That blog post has something ive wanted to see in benchmarks for a loooong long time: standard deviation. Without it you can never tell if x was slightly faster than y in z test...

Posting Permissions

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