Slashdot Mirror

← Back to Stories (view on slashdot.org)

Native Sorenson Playback Comes to Linux

Posted by timothy on from the snowball-effect dept.

Pivot writes: "With the release of Xine v0.9.11a, it is now possible to play back Quicktime movies encoded with the Sorenson SVQ1 encoding natively. There are still some minor issues with sound, and still no support for SVQ3 encoding, but overall this is a major achievement. Downloads are at xine.sf.net. I wonder what apple will do about this." Note: you may have to cut and paste that "movies" link into a new tab or browser.

5 of 276 comments (clear)

  1. Licensing? Patents? by JoshuaDFranklin · · Score: 5, Interesting

    Licensing? Patents?

    Someone care to explain what the team did about
    these little problems?

    1. Re:Licensing? Patents? by prockcore · · Score: 5, Interesting

      Yes please tell.. I have a fully working SVQ3 codec that I reverse engineered sitting on my harddrive (note, only video, the audio is QDM1 which I haven't done yet) I haven't released it due to blatent patent infringement :) but if Sorenson isn't going do to anything about it.. I may release it after all.

    2. Re:Licensing? Patents? by G-funk · · Score: 5, Insightful

      Stuff them, release it AC somewhere, post it here, then by the time somebody starts to "cease and desist" it'll be too late.

      --
      Send lawyers, guns, and money!
    3. Re:Licensing? Patents? by HeUnique · · Score: 5, Interesting

      Here's a snippet from an email which was posted on FFMPEG's mailing list:

      From: Arpi
      To: ffmpeg-devel@lists.sourceforge.net

      Date: Yesterday 23:02:26

      Hi,

      I've just examined xine's fresh working SVQ1 decoder. It's implemented in a ~60k .c file, and uses a 90k .h containing the tables.

      Looking at the source, it looks like SVQ1 is a tricky h263 variant - as gerard also noticed some time ago. They crypted (don't worry, just order
      change and some xor) the first 4 bytes of the header, to hide it's a h263 one. Ah, and they replaced the patented DCT by recursive VQ.
      And, they use YVU9 (chrominance 4x4 subsampled) instead of YV12 (2x2 subsampling).

      So, as you can see, the SVQ1 guy who wrote the native decoder, replaced the sorenson patented stuff with something free..

      --
      Hetz (Heunique)
  2. This is NOT clean-room implemented by Anonymous Coward · · Score: 5, Insightful

    If you read the thread you will see that the author looked at Apple's QT binaries for codebooks to decode some of the encodings. I'm sure there are EULAs that prohibit this. This patch is going to have a lot of legal problems. That's a shame because it is a big boost for QT and thus for Apple, but that's the way it is. I grabbed a copy of it so that when they get an injunction from Apple I'll still be able to post it somewhere in the Free World (ie, not in the US).