Announcement

Collapse
No announcement yet.

Google Starts Pushing Out VP10 Open-Source Code Into Libvpx

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Google Starts Pushing Out VP10 Open-Source Code Into Libvpx

    Phoronix: Google Starts Pushing Out VP10 Open-Source Code Into Libvpx

    Google has begun committing open-source code to the libvpx repository for supporting their next-generation VP10 video format...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Why even bother with VP10? Why not just start backing Daala or Thor, considering they're doing the same exact thing as VP10 is (in the same generation) and both are farther ahead.
    Is Google just suffering from a need to control everything it uses?

    Comment


    • #3
      Originally posted by Daktyl198 View Post
      Why even bother with VP10? Why not just start backing Daala or Thor, considering they're doing the same exact thing as VP10 is (in the same generation) and both are farther ahead.
      Is Google just suffering from a need to control everything it uses?
      How do you know how far along vp10 is?

      Comment


      • #4
        It's partly releasing updates based on their own research. Partly potentially it will be posturing for the "netvc" codec, which currently will have submissions from Xiph and cisco - the best of the submitted codecs would get combined into a new codec, a bit like Opus for audio.

        Comment


        • #5
          Originally posted by Daktyl198 View Post
          Why even bother with VP10? Why not just start backing Daala or Thor, considering they're doing the same exact thing as VP10 is (in the same generation) and both are farther ahead.
          Is Google just suffering from a need to control everything it uses?


          From the Cisco blog last time:

          Having Google submit VP9 as a contribution to the NetVC would actually be very welcome. The more technology we have to work with the better. The final NetVC codec will be neither Thor, nor Daala. It will be some kind of mix of the various contributions received. (disclosure: I'm in the Daala team at Mozilla)

          So even a Daala guy seems to think that more options/ideas is a good thing..

          Comment


          • #6
            Originally posted by geearf View Post



            From the Cisco blog last time:


            So even a Daala guy seems to think that more options/ideas is a good thing..[/FONT][/COLOR]
            Not to mention the more of these codecs exist, the more protection these codecs have.

            Comment


            • #7
              Originally posted by Daktyl198 View Post
              Why even bother with VP10? Why not just start backing Daala or Thor, considering they're doing the same exact thing as VP10 is (in the same generation) and both are farther ahead.
              VP10 by itself isn't so bad. What's bad is that Google is again doing it alone. What they should do is submit it to the IETF and work together with them, so that NetVC will in the end be a mashup of all three codecs - Daala, Thor *and* VP10. The VP family of codecs would benefit a lot from the tech Thor provides. I explained this a bit when I posted a low-bitrate test of x264 and vp9 here.

              Originally posted by Daktyl198 View Post
              Is Google just suffering from a need to control everything it uses?
              Seems so. And that means VP10 will be the same as 8/9 - there will be implementation bugs that will become part of the format, there won't be any spec, the only user of the codec will be Youtube, there won't be enough focus on rate control and psy optimizations. So in the end, I doubt VP10 will beat even x264.

              @My8th: VP10 as it exists currently in the main libvpx tree is the same as VP9, there's nothing new. There's an experimental branch that contains some, well, experiments. But that's that.


              All this new codec stuff is just meh anyway, since a few days my favorite codec has become... Cinepak! It's awesome. Too bad ffmpeg's encoder is crazy slow, because it's tons better than the encoder Windows ships with. That post I linked to above contains a demonstration of the awesome power of Cinepak

              Comment


              • #8
                Hmm... Anti-VP8/9/10 trolls are out in force these days. To be clear: Yes, Google control VP9/10... but Cisco controls NetVC (look at the members).

                Also, don't forget Cisco blocked VP8 in WebRTC, forcing everyone to use H264. I'm not a conspiracy theorist, but I do want to know what is motivating all these developments.

                In any case, as I have said before: Stability of video codec format is more important than the 1.2% performance improvement. Google - stop writing more versions, start getting your existing VP9 support into hardware. Then let's do the next iteration in 10 years' time...

                Comment


                • #9
                  Originally posted by OneTimeShot View Post
                  Hmm... Anti-VP8/9/10 trolls are out in force these days.
                  Rather than troll accusations, can you actually refute the points I make here and in that other thread I linked to? Can you provide your own encoder tests, similar to how I did? This is the thread: http://www.phoronix.com/forums/forum...ge4#post814422

                  Also, no, Cisco doesn't control NetVC, you're forgetting Mozilla and Xiph. Stuff from Thor is flowing into Daala and the other way around. VP10 could be in the mix as well. But that's up to Google.

                  Comment


                  • #10
                    Maybe troll is not the right word. Nay sayers? People who seem to be against VP* for some reason or other?

                    I am assuming that you want a royalty free codec, so your choices are: Theora, VP8, VP9, VP10, Thor, Daala. Comparisons against pay-for H264 and H265 are interesting, but IMO only important if you show 25% or better levels of compression.

                    What is your conclusion by the way? If my H264 file is 1Gb, what VP9 size do I need for the same quality? 2Gb, 3Gb? This is genuine interest, not trolling - I have seen lots of "better than" comparisons, but no actual metrics for "how much bandwidth do I save/lose for the same quality?".

                    There isn't much to refute in your testing, other than a minor quibble that BluRays are encoded with H264 to start with which will unfairly benefit H264 encoders.

                    My view is that I don't care (again maybe wrong word) about codecs that will be available in the future. They could be fantastic, or H268 could blow them out of the water. IMO, we need to concentrate on pushing one (I'm not sure that it even matters which one!) of these codecs into general use and all widely used hardware...

                    Comment

                    Working...
                    X