Thanks Ralph suggestion. The m-c doesn't have the vp9 support right now, but we can port the libvpx and use the encoder part into m-c if we have no concern about this. My plain would likely to implement the WebM 1 first, fine-tune the architecture to be stable. Then we can move to WebM 2 if the spec is ready to implement. Does it make sense? ------------------------------------- -rlin
----- Original Message ----- From: "Ralph Giles" <[email protected]> To: [email protected] Sent: Thursday, September 5, 2013 7:57:00 AM Subject: Re: MediaRecorder video container format On 13-09-04 1:15 AM, Randy Lin wrote: > Hi all, > I am working on video recording container stuff. > Right now we have the vp8/vp9 video encoder and opus audio encoder, Do we have vp9? I thought the libvpx in m-c only did vp8. > But the opus in WebM 2.0 spec seems have not defined well already. I'm not happy with the state of the spec, but it is ready for implementation feedback. It would be valuable to have experience with the issues there before WebM 2 gets too far along in other implementations. > If we use the WebM 1.0, we may need to implement vorbis and mux the > vp8+vobris in WebM 1.0. Yes, we'd need to enable the vorbis encoder and add muxing support for vp8+vorbis. My instinct would be to do WebM 1 first, as it's lower risk. Software VP8 encoding uses less cpu and hardware support is better deployed. However it's a shame to not use the Opus encoder we've already implemented, and I intend to have vp9 playback support working for ff28. I'd say it depends on how much do you want to work on the experimental Opus muxing. -r _______________________________________________ dev-media mailing list [email protected] https://lists.mozilla.org/listinfo/dev-media _______________________________________________ dev-media mailing list [email protected] https://lists.mozilla.org/listinfo/dev-media
