Robert Krüger <[email protected]> added the comment: Thanks for taking the time to explain! I was afraid that might be the outcome. So, just for my understanding: If the P-frames/slices after the I-frames/slices do not use material from before the last I-frame/slice it is simply sloppy by the chip in the camera not output IDR pictures for those I-frames or at least set recovery point SEIs?
My use case is simply to allow seeking within such files (from that camera, assuming they have the same characteristics) in a player without decoding unnecessary (matter of definition) frames. So currently the only possibility to build an index to have libav seek directly to those I-frames for which no following p-slices exist that reference slices from before the last I-frame would be to decode the entire file and do the slice/frame bookkeeping and decide then which positions might serve as seek points? ________________________________________________ FFmpeg issue tracker <[email protected]> <https://roundup.ffmpeg.org/issue2482> ________________________________________________
