"Normalization" causes a lot of confusion and is so controversial because there are some music programs that abuse the term and substitute an algorithm that effectively compresses the music. Compression is a type of lossiness where the dynamic range is lost forever, to be avoided. However, the usual kind of normalization is "by peak", where the software searches for the largest transient in the selected audio and calculates the factor that would result in that peak being very close to "1" (usually you have the option to set it to a little less, like 0.98). Then that same factor is used to simply scale the rest of the audio data. Since this is being done in 32-bit math there is no chance that it leaves any audible artifacts or accidental clipping. It would only be proper to normalize in this manner if you process the entire show at once, as one of the last operations before tracking. Also, there may be transients such as loud claps between songs that you may want to remove first to make the process more effective. As for tracking, good old CD Wave still works just about as well as anything else:
http://etree.org/cdwave.html
No members have liked this post.