Normalization is simply a way to make the volumes of your music “normal.” You may have noticed that some of your CDs are a lot quieter than others, and this is what normalization attempts to fix. Digital audio is just a big stream of numbers, one after the next. When these numbers are big, the music is loud. When the numbers are small, the music is quiet.
Normalization scans a digital audio file and finds the biggest number in the whole song. If the biggest number is only half as big as it possibly could be, then your song is really only half as loud as it could be, so the normalizer detects that and scales all the numbers up to make the volume “normal.”
The normalization percent that you set is the percent of the maximum that you want the new peak level to be at. If you want all of your music to be as loud as possible without any degradation (clipping), just set the value to 100%. If you want to disable normalization, set this value to 0%.
Normalization can be handy on compilation CDs where you don’t want to have to adjust the volume knob between tracks to maintain the desired volume. On the other hand, normalization isn’t always a good choice on CDs where certain tracks are intended to be a lot loader or quieter than other.
Normalization, like any DSP (Digital Sound Processor), introduces a small amount of noise into the signal. In the case of normalization, the amount of noise is tiny. However, if you’re working with lossless audio, you may not want to normalize.