A Place For Us

[Pedaling Prince]

the use of compressors on the Master Buss isn't only engaged for loudness purposes [in order to compete], but to tame exuberant transients - the ear responds to average loudness, not short transient peaks like a raw snare drum hit, say. it is possible, in the right hands, to reduce them without any audible degradation in the delivery of the material, and thus enable the overall volume to be raised - this should NOT be confused with the loudness war because that's a different matter entirely. so, if you have an untamed peak which occurs only once in your song at say -2dB, applying peak normalisation will only raise the volume by 2dB . . . now, by applying correct compression, this transient could be reduced without acoustic consequences, i.e. implemented entirely transparently, to say -6dB where the next transients in the source material might occur. so, if you then applied Peak Normalisation to this newly compressed source, to 0dB, you'd gain 6dB in apparent loudness . . . rather than only 2dB in the previous example. this, in essence, explains why Peak Normalisation is utterly useless....and of course, that's why it's been abandoned, and why compression on the Master Buss is your friend.....but you gotta do it right, yeah?

I agree with your assertion that sometimes transient peaks like this have to be tamed in order to bring up the loudness of a track. Myself I run into this most frequently with outdoor production sound captured on my iPhone for my online videos. Unfortunately I don't have the money for professional gear and while the iPhone's mic captures reasonably decent mono audio (with FiLMiC Pro it even captures it in uncompressed 16-bit 48 kHz PCM ) it is rather sensitive to wind noise. I have had many occasions where the loudest point in an outdoor recorded track was a blast of wind noise, preventing me from bringing the speech I actually wanted up in volume.

The first time this happened I knew I could've used a compressor but, as you know, I prefer to avoid that approach simply because I like to keep the dynamics of a track as intact as possible. I suppose I could've taken the approach you describe, and probably would if I was in a rush, but since I wasn't in a rush I decided to take a different approach that, while more labour intensive, allowed me to solve the problem without having to use a compressor.

What I did was I went into the production sound, manually selected the offending wind noise peak and normalized it down below the level of the peaks surrounding it. Then I normalized again, found the next quietest bit of wind noise, took it down manually, and so on, until finally I got a peak that was actually part of the speech; at that point I normalized the production audio to that peak and voila, I had it exactly where I wanted it.

Thus far I have not yet had a situation where a snare hit was so much louder than its mates on a track that I had to tame it, but I could certainly see that happening sometime. If i ever does, I suspect my technique here would be easier to apply because drum hits are much easier to isolate than a blast of wind noise on a track of speech.

So, to sum up, I don't entirely agree with your assertion that peak normalization is useless; I always peak normalize as the final step in my mastering phase. I do acknowledge, however, that there are times when an errant peak needs to be contained for one reason or another, and I agree that something needs to be done about that peak when it happens. I disagree that master buss compression is the only way to do it, but I do concede that master buss compression can do it, in a pinch, if you don't have time to deal with the peaks by hand, as I prefer to do whenever possible.

That's one thing about me and my style. When I do use a compressor, I rarely use a ratio higher than 4:1, and generally I use it set to a gentler ratio, and the only kind of track I always use gentle compression on is a vocal track; for the most part I avoid using compressors on any of the other tracks (though, again, there have been exceptions). Given the choice, I'd rather automate a track's levels, keeping the natural short-term dynamics of a sound intact, than use compression. It's more work, yes, but at least it gives me the satisfaction of knowing I kept the processing on the sound to a minimum.