Question from an Audio Masterclass visitor…
“How do you get the maximum volume out of your mix before mastering while ensuring the master channel doesn't clip? Working in DAWs the crucial thing is to prevent the master channel from going in to the red, but inevitably the easy answer is to just turn the master down. But this effects the overall volume at mixdown and makes the mastering session, particularly the all important 'loudness factor' more difficult. I suppose this question is a fairly comprehensive issue of an overall approach to a mixing session. I've got fairly good info about how to approach a mix, i.e. matching the volumes of bass and bass drum, mixing everything in reference to the vocal etc, but as I go along thinking I'm getting the right volumes, a peak in the song, or a synth or any number of irritating factors clips the master forcing me to drop the master volume. I'm sure this is not the ideal situation… please advise.”
The first and very most important advice I have to give is that you have to be TOTALLY capable of producing a mix that does not clip. No red lights, all the way from beginning to end.
If that sounds too basic, then I'd say that you haven't heard much of the stuff that gets sent to me. It is just so important to keep that red light off.
So the question becomes, how do you get maximum loudness with the red lights off?
One common answer is to use the 'normalize' function that most digital audio workstations have.
Don't. It's a waste of time. It achieves nothing at the expense of yet another cycle of processing. The only time I would use normalization is if I did not intend to optimize a track for loudness, which actually is surprisingly often.
The first thing you need to do is look at where the peak levels are in the track. You may find there are three or four places in the track where the level is higher than everywhere else.
Usually, this points to a fault in the mixing. Ask yourself whether it really sounds right. It probably doesn't.
But there occasions where this can happen, even when the mix is perfect. In recordings of acoustic instruments for example… it isn't at all uncommon to have a small number of peaks that occur infrequently.
What these peaks are doing is eating up level. If you lower the peaks, using automation possibly, then you can bring up the level of everything else.
Put it this way… if you had one peak that was 6 dB louder than everything else, then if you can bring it down to normal level, then your whole track can be 6 dB louder.
Controlling occasional peaks is best done manually. Often though the peaks are much more frequent.
There could for instance be a loud, impulsive snare drum. Once again, I would go back to the mix in the first instance to see if it can be fixed there – possibly by compressing just that one channel.
Otherwise, you can use a limiter to bring down short-duration peaks, often without noticeable effect. Then you can bring the level of everything else up.
If the peaks are of longer-term duration, then you can use a compressor. For this purpose, and to avoid obvious compression (unless that's what you want), set a low ratio and a low threshold, and a fairly long release time.
Once again, you will find that you can bring up the level of the whole track so everything is closer to the red light level, without actually getting there.
I'm going to stop at this point, and so should you. These are the fundamental skills of loudness optimization. They need to be mastered, and well mastered, before trying other, potentially more damaging, techniques.
Let us know how you get on.