Citaat:


In fact he's got a nice example for it too. If you throw all the song of an album in whatever sequencer you're using, and normalize all of them thinking you're getting them to a nice loud level, it's very much possible that a Ballad will sound louder than a rock song in the same album...other surprises could happen too, that's why he says Normalization should never be used to regulate song levels in an album.




There is a big difference in normalizing tracks of a song still to be mixed (as I assume we were discussing here) and the final tracks for an album. What you talk about is mastering. And well in that case it most of the times makes no sense to use normalization. At least not peak normalization, because that can indeed give bad level balances amongst the tracks. If used already in that case (which I wouldn't) you would use RMS normalization, since this is more like an average energy level to accomplish. Still however I would not recommend it, since mastering is not like you throw all tracks through a boosting and limiting piece of software in order to get the same average level, but on the hearing and feeling. In that way every track has to get it's own special treatment while still in good balance with all the other tracks on the album. But again, this is about mastering and the type of normalisation discussed here is merely for peakboosting and lifting the whole track into workable levels, before mixing down...

I'll be back...