A utility to analyse sound files and print out the volume profile of the start and end. The idea is to embed those profiles in metadata so players can do intelligent mixing between tracks. Supports FLAC and MP3. Implemented in http://git.musicpd.org
Great tool, works excellent with mpd! But to compile it, edit "Makefile", and remove the "-Werror" argument from the second line. (You'll also need to install libaudiofile-dev to compile.) Also, the bash script "mixramp-tag" is not ideal (as it uses the problematic eyed3). Replace mixramp-tag with the version in the code section (it uses mid3v2): sourceforge.net/p/mixramp/code/HEAD/tree/mixramp-tag, "Download this file". For this, you need version >= 1.3 of mid3v2, included in mutagen >=1.21, which is in the repositories of Debian Jessie, Ubuntu 13.10 or newer. Or else from launchpad.net/~lazka/+archive/ppa . Once compiled, put "mixramp" and "mixramp-tag" in your path, e.g. /usr/local/bin/, chown to root and chmod to 755, and you're ready to go. Usage to prepare songs: mixramp-tag <audio_file> ... <audio_file>s should end in .flac or .mp3. To play: in mpd (using mpc), set mixrampdb [<db>] to e.g. -6, and mixrampdelay [<seconds>] to e.g. 0.5, for a great sounding mix. Cheers, Karl
mid3v2 do support TXXX frames now!