I'm playing around with FFmpeg options for large and small videos of different container formats. I'm wanting to trim the video in a performant way that is also accurate.
I've read a lot on this topic as well as this page: https://trac.ffmpeg.org/wiki/Seeking#Cuttingsmallsections
Input seeking has actually, so far, seemed pretty good (and fast!).
ffmpeg -ss 09:25:03.0000 -to 09:56:15.0000 -i timer.mp4 -codec copy timer_modified.mp4
What are the gotchas here that I'm missing? From the docs, it seems like this is the best approach and with newer updates, the more accurate as well.
Output seeking has surprisingly given me black frames, which is strange. I initially had a thought that I'd trim to the approximate location with input seeking and then re-encode that output back into ffmpeg to get a more precise cut, but looking for feedback.