normalize − adjusts volume levels of audio files.
normalize [ options ] [ -- ] file...
normalize is used to adjust the volume of wav audio files to a standard volume level. This is useful for things like creating mp3 mixes, where different recording levels on different albums can cause the volume to vary greatly from song to song.
normalize operates in two phases. In the first phase, it analyzes the specified files as wav audio files, and computes the volume of each file. In the second phase, it applies a volume adjustment to each file to set each file’s volume to a standard level.
Adjust the RMS volume to the target amplitude AMPLITUDE; must be between 0.0 and 1.0. If a number suffixed by "dB" or "dBFS" is specified, the amplitude is assumed to be in decibels from full scale. The default is -12dBFS.
Enable batch mode: see BATCH MODE, below.
Deprecated. In previous versions, this enabled the limiter, but now the limiter is enabled by default.
Disable the limiter, and just clip any samples that are too large. Same effect as -l 0dBFS.
Display all values as decimal fractions instead of in decibels. By default, volume adjustments are shown in decibels, and volume levels in dBFS, where 0 dBFS is the level of a square wave of maximum amplitude.
Skip the volume computation phase: don’t compute the volume adjustment from the current volumes of the files. Instead, just apply the given gain as a volume adjustment to all files. As a plain number this is just a multiplier applied to all samples, If a number suffixed by "dB" is specified, all volumes are adjusted by that many decibels.
Use this option when adjusting MPEG audio files if your MP3 player does not recognize ID3v2.4 tags. See MPEG AUDIO ADJUSTMENT, below, for details.
Use this option when adjusting MPEG audio files if your MP3 player does not recognize ID3v2 tags and has trouble playing some ID3v2 tagged MP3 files. See MPEG AUDIO ADJUSTMENT, below, for details.
This controls the behavior of the limiter. By default, all samples above -6dBFS (0.5) are limited, but this option sets the limiting level to LEVEL. Setting LEVEL to 1 (or 0dBFS) does no limiting (clipping is done instead); setting LEVEL to 0 does limiting on all samples. The default value is recommended unless you know what you’re doing.
Enable mix mode: see MIX MODE, below. Batch mode and mix mode are mutually exclusive.
Compute and output the volume adjustment that would set the volume to the target, but don’t apply it to any of the files (i.e. skip the second phase). If you use this option, your files will not be altered in any way.
Don’t print any progress information. All other messages are printed as normal according to the verbosity level.
Adjust using peak levels instead of RMS levels. Each file will be adjusted so that its maximum sample is at full scale. This just gives a file the maximum volume possible without clipping; no normalization is done.
Don’t output progress information. Only error messages are printed.
When averaging volume levels for batch mode or mix mode, throw out any volumes that are more than THRESHOLD decibels from the average. A high value here (say, 50) will make sure that the volumes of all files are considered in the average.
If an adjustment to be made to a file is smaller than THRESHOLD decibels, consider the file already normalized and don’t do the adjustment. This is 0.125 by default, or 0 if the -g option is given.
Increase verbosity. This option can be repeated for more messages.
Force output files to have samples that are W bits wide. This option is ignored when adjusting MP3 files.
Display usage information and exit.
Print version information and exit.
Terminate option list.
This mode is made especially for making mixed CD’s and the like. You want every song on the mix to be the same volume, but it doesn’t matter if they are the same volume as the songs on some other mix you made last week. In mix mode, average level of all the files is computed, and each file is separately normalized to this average volume.
When operating on a group of unrelated files, you usually want all of them at the same level, and this is the default behavior. However, a group of music files all from the same album is generally meant to be listened to at the relative volumes they were recorded at. In batch mode, all the specified files are considered to be part of a single album and their relative volumes are preserved. This is done by averaging the volumes of all the files, computing a single adjustment from that, and applying the same adjustment to all the files. Some analysis is also done so that files with volumes that appear to be statistical aberrations are not considered in the average. This is useful if you have albums (like many of the author’s) in which there is one "quiet song" that throws off the average.
MP3 files are "adjusted" by setting a relative volume adjustment frame in their ID3 tags. There is a frame for this, called "RVA2", that does exactly what we want, and is a native frame in ID3v2.4. Unfortunately, many MP3 players do not support v2.4 tags (including xmms, as of this writing), and the RVA2 tag is not native in previous ID3 versions. In fact, adding an RVA2 frame to a v2.3 tag confuses some MP3 players. Therefore, we are left with two choices when trying to add volume adjustment information to an ID3 tag:
Go ahead and upgrade the tag to version 2.4, and use RVA2 tags. This is the default behavior, in the hope that eventually MP3 players will support v2.4 tags and this won’t be a problem anymore.
Upgrade the tag to only version 2.3. Instead of RVA2, use an "XRVA" tag with the same format as an RVA2 tag. This isn’t a native frame, but since it starts with an "X", it’s considered experimental and therefore legal, according to the ID3 spec. The --id3-compat option turns on this behavior.
The disadvantage of the first method is that your MP3 player may no longer read the ID3 tags on your files. Bug the author of your favorite MP3 player to support ID3v2.4 tags!
The disadvantage of the second method is that the XRVA frame is only recognized by the xmms-rva plugin that is packaged with normalize. On the other hand, I don’t know of any MP3 players that recognize the RVA2 frame, either, so it may not make any difference.
The other option related to ID3 tags, --id3-unsync, is only necessary for compatibility with old MP3 players that don’t recognize ID3v2 tags at all. If your MP3 player complains of garbage at the start of tagged files, or is unable to play the files at all, turn this option on. This option should never hurt, but if your MP3 player knows about ID3v2 tags, you don’t need it.
Note that your version of normalize must be compiled with MAD library support to analyze MP3 file volume levels.
Chris Vaill <cvaill [AT] cs.edu>