I did some sound file mangling earlier this week. I’ve been using the Bambuser app for Symbian recently to capture some recordings. If you tell their site your User-Agent is a mobile browser, it’ll work OK with free software – icky but better than many video services today.
Even when I was recording audio only, the site left me with flash video (flv) files which contained both audio (mp3) and video streams. Step one was to remove the blank video and convert to mp3: ffmpeg -i videofile -acodec copy -vn part1.mp3
Mobile service isn’t great in rural Somerset around Weston-super-Mare, so some recordings were split into several parts. Step two is to combine them with cat part1.mp3 part2.mp3 >whole.mp3
That’s OK, but it leaves some painful squawking at the join, where the player hits the header information of part2. If you’re unlucky, it’ll crash some players. So step three is to remove it: ffmpeg -i whole.mp3 -acodec copy whole-clean.mp3
. Often I transcode to ogg vorbis at the same time with ffmpeg -i whole.mp3 whole-clean.ogg
but it depends what I’m doing with the recording.
Are there any other handy ffmpeg tricks I could use here?
2 Responses to Audio Processing Tips for ffmpeg