2011/3/26 Alex Deucher <alexdeuc...@gmail.com>: > > WebM would definitely be cool. >
Ok I give this idea some thoughts and I think I am gonna go with a hardware accelerated vp8 decoder into a vdpau state tracker. On the plus side : - VP8 has a way more simple design than h264, and that should cut the development time, to keep the project doable. - I saw on this maillist talks about the S3TC patent situation. Why nobody express similar concerns about an h264 implementation living in mesa ? Because that would mean violate much more patents, without paying royalties and everything... - The reference implementation is good, the ffmpeg implementation is even better : a lot of inspiration. - Some work done on VP8 could be used later for h264 (or not, regarding patent problems ?) So two major steps : - Get a pure software implementation of vp8 to run into a proper gallium state tracker, and use it to play videos through various "vdpau ready" softwares. That step include readying the vdpau state tracker to include more video formats to be included. - Start implementing shader based optimizations, where they can be useful. I did my previous internship in a french company, developing an h264 decoder from scratch. Took me 6 months and although functional, the result is not bulletproof (but h264 has so many features...). I learn a lot of thing, including the fact that writing a big piece of software definitively takes a lot of time :-) * So my main question is : will that be ok to pull an existing vp8 decoder into the state tracker to begin with, and concentrate on the gpu side of the project ? The gsoc applications are now open, the deadline is in 10 days. If anyone have any further comments on this idea that would be great. Thanks, Emeric _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
