Announcement

Collapse
No announcement yet.

ITU Approves H.265 / HEVC Video Codec

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

  • ITU Approves H.265 / HEVC Video Codec

    Phoronix: ITU Approves H.265 / HEVC Video Codec

    The International Telecommunication Union (ITU) on Friday granted first-stage approval for H.265, a.k.a. High Efficiency Video Coding (HEVC), as the successor to the widely-used H.264 / MPEG-4 AVC video codec...

    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
    MPEG LA is going to work with HEVC, so you can expect the same legal situation here as with H264.

    Meanwhile Daala codec is of same quality and free.

    Please go with open technologies and drop this!
    Film Studios should pay for codec technologies, but it should always be free for private and non-commercial entities. Something that MPEG LA never gets.
    Last edited by brosis; 26 January 2013, 11:32 AM.

    Comment


    • #3
      There is also VP9 coming, see here, as well as the works for a IETF (opus like) supported codec, see here and here.
      Last edited by oibaf; 26 January 2013, 11:41 AM.

      Comment


      • #4
        Is there any recent SSIM benchmarks of daala and/or VP9 vs h.265? I'm curious how they compare to each other at this point.

        Comment


        • #5
          Originally posted by Prescience500 View Post
          Is there any recent SSIM benchmarks of daala and/or VP9 vs h.265? I'm curious how they compare to each other at this point.
          VP9 is still WIP, but see here (post 42).

          Comment


          • #6
            I don't fully understand the numbers, but if I read it correctly, they are really catching up. Super high quality would need work, but I doubt that would be used much outside of blu-ray movies. It would be awesome if VP9 was good enough to become the next html5 video standard. Even better if it could be the new standard for professional hi-def and video recorders, but we'll probably have to wait for daala or VP10 or better for that as that's probably not Google's main priority.

            Comment


            • #7
              50% is best case.

              Qualcom showed their early H265 implementation (on CPU), it was 50% less than H264. But Qualcom engineer stated that 40-45% should be more acureate for most of vids.

              On the happy note:
              VP9 is "only" bigger than H265 by 7% currently. And google aim at 20% less.

              That is very good!

              Now if they could get MIME for WebP and finish its specification......

              Comment


              • #8
                But without Apple and Microsoft support, VP9 is going nowhere as an standard. It's the same situation we have now. In fact, Apple alone can kill any de facto standard just like they did with Flash support in iOS. Not even Google has been able to force with Youtube and Chrome any free standard codec.

                Comment


                • #9
                  Originally posted by newwen View Post
                  Not even Google has been able to force with Youtube and Chrome any free standard codec.
                  I think Google could have, if YouTube would be VP8-only.
                  I really don't understand why they didn't force all YouTube-videos to be WebM if they want it to be the Internet-standard.

                  Comment


                  • #10
                    Originally posted by newwen View Post
                    But without Apple and Microsoft support, VP9 is going nowhere as an standard.
                    See the links I posted before, there is an ongoing work for a IETF approved patent-free video codec standard, similar to the audio codec opus. See also the video codec mailing list for updates. This is however a longer term work then VP9 and Daala. VP9 should be ready soon, Daala will proably converge in the IETF.

                    Comment

                    Working...
                    X