On 6/9/2022 2:59 PM, Bo Berglund wrote:
So given my original command in the script:
CMD="ffmpeg -hide_banner ${MODE} -i \"${M3U8URL}\" -vf scale=w=-4:h=480 -c:v
libx264 -preset fast -crf 26 -c:a aac -t ${CAPTURETIME} ${TARGETFILE}"
should I replace -c:v libx264 -preset fast -crf 26 with something else or simply
remove it and rely on ffmpeg to understand how to do based on me setting the
ouptput file extension to ts?
I'd do the latter unless you need other specific parameters. Actually, you
could just copy the streams into the output container if you want:
ffmpeg -hide_banner ${MODE} -i \"${M3U8URL}\" -c copy -t ${CAPTURETIME}
${TARGETFILE}
where TARGETFILE ends with ".ts" -assuming- that the stream is a mpeg-ts.
Or force the output format to match the input. The point of copying is to
avoid inserting another code/decode step. (What -is- the stream format? No
output was included.)
Or do I have to specify an encoding library?
That info is in the ffmpeg doc....
Notice that early on I just downloaded the video and then applied the geometry
change and recoding to mp4 as a second step but it took a long time, almost half
the playing time of the video.
That's a different issue. You are hiding that processing time behind the
download time.
Another option as mentioned before is to do all the cropping/etc on intake
and use an output codec/muxer that doesn't require the finalize step, but I
think that precludes x264 encoding. If you need x264, there's another step.
There are multiple paths to the same goal.
z!
_______________________________________________
ffmpeg-user mailing list
[email protected]
https://ffmpeg.org/mailman/listinfo/ffmpeg-user
To unsubscribe, visit link above, or email
[email protected] with subject "unsubscribe".
