Someday I’d like to expound on this more…

As part of the loudness wars, they are compressing the dynamics of music more and more.

So we have music that’s at an average level of 90% with a 98% peak vs music that’s at an average level of 50% with a 75% peak.If you normalize both to 100%, the 98% one becomes slightly louder (average 91% with 100% peak), and the 50% one becomes 66% average with a 100% peak. Even though both peak the same, the average is very different, and one will “sound 91%” while the other will “sound 66%”.

I found a normalizer that works via gain instead of peak. It deals with averages. It prints a report. In the situation above it would actually reduce the first (90% avg/98% peak) to be quieter (45% avg/49% peak) in that situation, allowing the “quiet” encodes with more dynamics/volume variation to end up sounding as loud as the “loud” encodes with less dynamic/volume variation.

I don’t know what it’s called. I named it normalizer.exe and put it in my C:\UTIL\ folder where I keep my command-line utilit.es Anyone wants it, lemme know. Think it’s freeware.

Mood: full
Music: L7 – Deathwish