Announcement

Collapse
No announcement yet.

AOMedia Forms AV1 Software Working Group Using Intel's SVT-AV1

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

  • AOMedia Forms AV1 Software Working Group Using Intel's SVT-AV1

    Phoronix: AOMedia Forms AV1 Software Working Group Using Intel's SVT-AV1

    The Alliance for Open Media (AOMedia) today announced the formation of a Software Implementation Working Group (SIWG) to bring AV1 video support to more platforms by leveraging Intel's open-source SVT-AV1 implementation...

    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
    Typo:

    Originally posted by phoronix View Post
    AOMedia Software Implementation Workign Group

    Comment


    • #3
      They can start by making the noise encoder actually do something.

      Comment


      • #4
        What about us using AMD GPUs ?
        Intel is nice, but their GPUs are very weak and their CPUs are riddled with security vulnerabilities.

        Comment


        • #5
          Originally posted by Danny3 View Post
          What about us using AMD GPUs ?
          Intel is nice, but their GPUs are very weak and their CPUs are riddled with security vulnerabilities.
          This has nothing to do with GPU's. These are all software encoders running on CPU's. There's nothing that stops you from running svt-av1 on an AMD CPU.

          Comment


          • #6
            Originally posted by Brisse View Post

            This has nothing to do with GPU's. These are all software encoders running on CPU's. There's nothing that stops you from running svt-av1 on an AMD CPU.
            On top of all that, there's even a generic-C codepath that can compile to ARM/Risc-V/whatever for non-x86 chips: https://www.phoronix.com/scan.php?pa...0.8.2-Released

            Comment


            • #7
              Why now? They already developed the reference codec and had dav1d as a seperate effort implementing the standard. They could have agreed upon to produce a production ready implementation right at the start of their efforts. My gut feeling tells me that some companies were not pleased with the progress of the software implementations so far and want to accelerate its development. As AV1 is still not used as commonly today as I would have thought two years ago, this makes some sense albeit having sooner hardware encode/decode support would be more effective to further its adoption.

              Let's wait and see if RDNA2 and Ampere will get AV1 support in hardware finally. But this effort could also mean that hardware support could be way off still...
              Last edited by ms178; 20 August 2020, 03:39 PM.

              Comment


              • #8
                Originally posted by ms178 View Post
                Why now? They already developed the reference codec and had dav1d as a seperate effort implementing the standard. They could have agreed upon to produce a production ready implementation right at the start of their efforts. My gut feeling tells me that some companies were not pleased with the progress of the software implementations so far and want to accelerate its development. As AV1 is still not used as commonly today as I would have thought two years ago, this makes some sense albeit having sooner hardware encode/decode support would be more effective to further its adoption.

                Let's wait and see if RDNA2 and Ampere will get AV1 support in hardware finally. But this effort could also mean that hardware support could be way off still...
                libaom is more research-based. dav1d is a decoder.

                Between libaom and SVT-AV1, the latter is actually useable. I have a 3950X and SVT-AV1 helps itself to all of it. Meanwhile, libaom is unbearably slow.

                Comment


                • #9
                  Originally posted by ms178 View Post
                  Why now? They already developed the reference codec and had dav1d as a seperate effort implementing the standard. They could have agreed upon to produce a production ready implementation right at the start of their efforts. My gut feeling tells me that some companies were not pleased with the progress of the software implementations so far and want to accelerate its development. As AV1 is still not used as commonly today as I would have thought two years ago, this makes some sense albeit having sooner hardware encode/decode support would be more effective to further its adoption.

                  Let's wait and see if RDNA2 and Ampere will get AV1 support in hardware finally. But this effort could also mean that hardware support could be way off still...
                  dav1d is the decoder, there also is rav1e for encoding.

                  Comment


                  • #10
                    Originally posted by ms178 View Post
                    Let's wait and see if RDNA2 and Ampere will get AV1 support in hardware finally. But this effort could also mean that hardware support could be way off still...
                    Considering 2020 TVs are getting AV1 support (via Broadcom), I'd be surprised if they didn't.

                    Comment

                    Working...
                    X