volume normalization

For people who listen to music on their computer, a big problem is the varying volume which the sound comes out of the speakers. This problem is especially bad with headphones. As an example, someone is listening to Nirvana and they have adjusted the volume to their liking. Nirvana's song ends and Soundgarden comes on. Soundgarden's music file may specify a different volume, making the sound very loud or very quiet.

Volume normalization attempts to take care of this. By analyzing the volume on each file, it can try to compensate for the ones which are louder or softer, making them come out of the speakers all at about the same volume.

Volume normalization is a part of sound engineering which should be applied on the recorded material during the post-production.

If this is done digitally on a computer, it can be done easily by an automatic process that increases the volume on the entire recording so that the highest peaks are as close to 0db as possible (but never over). When editing it manually, for instance with an analogue mixer, the recorded piece's volume is also increased, but with analogue mixing, the signal is pitched up so that the average volume is at 0db.

There are also other ways of normalizing. For instance, if the recorded piece contains no peaks that are considerably at higher volume than the average (happens frequently when you are recording a choir), the recording is pitched up (or down) so that the average volume is around -14db (digital) or 0db (analogue).