When trimming packets and later concatenating packets, ensure that no packets that are not trimmed from stream of packets and thus are still part of final file depend on any frame from any dropped/removed stream of packets for correct glitch-free decoding.
Also ensure that there is no dangling packet in final file, dangling frames - the frames that are not key frames and are not used by any other frame for correct glitch-free decoding. Also probably modern codecs can use various techniques to code packets in various subtle incompatible ways for concatenation and trimming, like single packet having multiple or partial frames encoded inside. On Sun, Feb 18, 2024 at 6:04 PM Mark Filipak <markfilipak.i...@gmail.com> wrote: > On 18/02/2024 11.53, Gyan Doshi via ffmpeg-user wrote: > > On 2024-02-18 09:59 pm, Mark Filipak wrote: > >> On 18/02/2024 05.09, Gyan Doshi wrote: > >>> > >>> Add -mpegts_flags +initial_discontinuity to get rid of the corrupt > packet messages. This may or > >>> may not get rid of the glitches. > >>> > >>> Regards, > >>> Gyan > >> > >> Hey, Gyan, > >> > >> No effect whatsoever. Same message, same glitches. > > > > I forgot to mention, it has to be added to the trim commands. > > Do you mean to '-ss' & '-to'? > > Oh, no, no. Even if it worked, what would I do with all the bad PTSs and > audio frames left dangling. > > No, '-ss' & '-to' are disasters. > > Do you have any reaction at all to the news that decimal PTSs & DTSs are > being truncated up? Did you > know that? Why do you think it was done? > > --Mark. > > _______________________________________________ > ffmpeg-user mailing list > ffmpeg-user@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-user > > To unsubscribe, visit link above, or email > ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". > _______________________________________________ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".