Our organisation has a very large, nested, directory of music. However most of the songs have been recorded at different volumes.
This means that when playing songs (from different albums etc), you might have the speakers turned up load to hear a quiet song, then the next song will be loud and so with the speakers turned up, will be deafening.
Thus I am looking for a program to scale the average volume of all my songs to be all equal to each other. Obviously within the songs, I want to preferve the volume differences, (eg if there is a quiet intro, before the "beat is dropped", I don't want that ruined.)
- Must work on Debian Wheezy
- Must support
.ogg
and.mp3
. But the more other formats the better - Ideally would be installable from a debian repository, either main, non-free (being not open source, backports or through something like pip.
- Must support nested directories and normalise across all of them.
- Must run on the command-line
- Ideally would have some ability to repeat as new songs are added. (eg if it records the normal volume somewhere.
- Should not freak out if the songs are accessed when it is trying to convert them. If a song it is about to edit is suddenly played it must handle this gracefully. Ideally by moving that song to latter in the queue of songs to be adjusted.
- Should take advantage of multiple cores/CPUs
mpd
(music player daemon). I'm not sure what replaygain/tag based normalisation storage means. Looking it up