michaelni changed the topic of #ffmpeg-devel to: Welcome to the FFmpeg development channel | Questions about using FFmpeg or developing with libav* libs should be asked in #ffmpeg | This channel is publicly logged | FFmpeg 7.0.1 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
<aaabbb_>
has there been any work or thought on using constant folding as an optimization for ffmpeg's eval?
thilo has quit [Ping timeout: 246 seconds]
tufei__ has quit [Quit: Leaving]
thilo has joined #ffmpeg-devel
thilo has quit [Changing host]
thilo has joined #ffmpeg-devel
<aaabbb_>
there is a significant performance increase in some situations, where eval.c has to compute several multiplications, powers, and sine/cosine for every audio sample when everything in the expression is constant and could have been folded into one constant value
Teukka has quit [Read error: Connection reset by peer]
Teukka has joined #ffmpeg-devel
Teukka has quit [Changing host]
Teukka has joined #ffmpeg-devel
tufei has joined #ffmpeg-devel
kasper93 has quit [Remote host closed the connection]
Coinflipper has quit [Quit: ]
Coinflipper has joined #ffmpeg-devel
rvalue- has joined #ffmpeg-devel
rvalue has quit [Ping timeout: 246 seconds]
rvalue- is now known as rvalue
<Marth64>
Lynne: Is your use case playback? e.g. via mpv
<Lynne>
yup
<Marth64>
Great. This should work well then. I'm limited by what libdvdnav lets me do in terms of precision but its buttery smooth.
HarshK23 has quit [Quit: Connection closed for inactivity]
aaabbb_ has joined #ffmpeg-devel
jamrial has joined #ffmpeg-devel
mkver has joined #ffmpeg-devel
Krowl has joined #ffmpeg-devel
aaabbb_ has quit [Quit: Lost terminal]
Krowl has quit [Read error: Connection reset by peer]
HarshK23 has joined #ffmpeg-devel
cone-105 has quit [Quit: transmission timeout]
awehr_ has quit [Ping timeout: 252 seconds]
jamrial has quit [Read error: Connection reset by peer]
feiw1 has joined #ffmpeg-devel
feiw has quit [Read error: Connection reset by peer]
Guest10 has quit [Ping timeout: 264 seconds]
Guest10 has joined #ffmpeg-devel
jamrial_ has joined #ffmpeg-devel
awehr_ has joined #ffmpeg-devel
cone-488 has joined #ffmpeg-devel
<cone-488>
ffmpeg James Almer master:ab7893a7b114: avformat/hevc: don't write the same array values per nal addition
<cone-488>
ffmpeg Derek Buitenhuis master:46f7ea445637: avformat: Add a new stream disposition for multilayer video
<cone-488>
ffmpeg James Almer master:a696b288861a: avformat/hevc: don't write NALUs with nuh_layer_id > 0 in hvcC boxes
<cone-488>
ffmpeg James Almer master:cee43e512de9: avformat/hevc: store parameter set and layer IDs in HVCCNALUnit
<cone-488>
ffmpeg James Almer master:361b01329f27: avformat/hevc: use a single array for per-PS NALUs
<cone-488>
ffmpeg James Almer master:cbfbacff20a8: avformat/hevc: add a function to write a lhvC box
<cone-488>
ffmpeg James Almer master:c657c694e30f: avformat/mov: add support for lhvC box parsing
<cone-488>
ffmpeg Derek Buitenhuis master:c4ebdd8d2d7b: avformat/mov: Mark streams with a layered HEVC box as multilayer
<cone-488>
ffmpeg James Almer master:d4e2af414437: avformat/movenc: add support for writing lhvC boxes
<cone-488>
ffmpeg James Almer master:6a428876fc4f: avformat/movenc: add support for writting vexu boxes
<cone-488>
ffmpeg James Almer master:aa8230c7baad: avformat/movenc: add support for writting hfov boxes
<jamrial_>
mkver: so the spec says "When the decoder configuration record defined in subclause 8.3.2.1 is used for a stream that can be interpreted as either an L-HEVC or HEVC stream, the HEVC decoder configuration record shall apply to the HEVC compatible base layer, and should contain only parameter sets needed for decoding the HEVC base layer"
<jamrial_>
meaning what we were doing was not wrong, just not ideal/recommended
awehr_ has quit [Ping timeout: 268 seconds]
awehr_ has joined #ffmpeg-devel
<cone-488>
ffmpeg James Almer master:33d6e14d6fa9: avformat/hevc: reindent after previous commits