You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
VideoCodecId::ProRes exists in the API and returns a matchable VideoError::UnsupportedCodec. Nothing decodes it, and no container currently surfaces a ProRes track.
Two parts:
Container — rawshift-video-isobmff maps sample entries through mp4_atom::Codec, which does not model the ProRes 4CCs (ap4h, apch, apcn, apcs, apco), so a ProRes track currently falls through to Track::Other. Either upstream them to mp4-atom or handle the raw 4CCs directly.
Decode — unlike H.264 and HEVC, ProRes is a plausible candidate for a clean-slate safe-Rust decoder: the bitstream is publicly specified (SMPTE RDD 36) and intra-only, so docs/SUPPORT.md records it as undecided rather than never. Hardware ProRes decode also exists on Apple silicon via VideoToolbox, which would come free with Hardware decode: VideoToolbox backend (macOS/iOS, HEVC + AV1) #28.
ProRes is on the roadmap for iPhone Pro and professional editing workflows, so this is worth doing — but it is a genuine decision about whether rawshift wants to own a codec implementation.
VideoCodecId::ProResexists in the API and returns a matchableVideoError::UnsupportedCodec. Nothing decodes it, and no container currently surfaces a ProRes track.Two parts:
rawshift-video-isobmffmaps sample entries throughmp4_atom::Codec, which does not model the ProRes 4CCs (ap4h,apch,apcn,apcs,apco), so a ProRes track currently falls through toTrack::Other. Either upstream them tomp4-atomor handle the raw 4CCs directly.docs/SUPPORT.mdrecords it as undecided rather than never. Hardware ProRes decode also exists on Apple silicon via VideoToolbox, which would come free with Hardware decode: VideoToolbox backend (macOS/iOS, HEVC + AV1) #28.ProRes is on the roadmap for iPhone Pro and professional editing workflows, so this is worth doing — but it is a genuine decision about whether rawshift wants to own a codec implementation.
Follow-up to #39.