using two features of ffmpeg I can get accurate cue points out of an
audio file based on silences
ffmpeg -hide_banner -i $AUDIO -af silencedetect=noise=-50dB:d=0.898 -f
null - &> somewhatuseful.txt
and employ them to edit the audio
ffmpeg -y -hide_banner -loglevel error -i $AUDIO -ss $HEADNUM -to
$TAILNUM audio-$(date +'%H')_hours.wav
This is far more accurate then my previous attempts have delivered
My thanks to all who contributed to this code.
Robert
--
Robert Jeffares
Communication Consultants
64 Warner Park Avenue
Laingholm
Auckland
New Zealand 0604
+64 9 817 6358
_______________________________________________
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".
