mjhjmtu

If limiters on master buses can control peaks, then why is it necessary to add a compressor on each track to control peaks?

(I am not talking about changing the dynamic of tracks, just pick control gain reduction for preventing distortion.)

If you run a compressor on the master, a peak being reduced will affect every channel - so for example, you could hear all your instruments dropping volume.

It does change the dynamic of the whole track!

If you run compression per channel, you'll just reduce an individual peak, while leaving all other channels/instruments - this is far better for the overall sound.

It's necessary to add compressors on each track to change the dynamics of the tracks. Generally you should record and mix at appropriate levels so that you don't need to do any peak reduction to prevent distortion.

Compressors give us control over the dynamics of a track. With that control, we can choose more precisely which is the loudest instrument at any time. Combined with automation, the level of each instrument can be completely controlled at every moment for the whole song.

Compressors also change the way an instrument sounds. Drums are a good example. As you put more and more compression on a drum track, it changes the balance of high and low frequencies, and it changes the decay time of the drum hit, and also the transients. With a compressor you can make a drum sound like something different from anything you can get from a real drum without a compressor. So if you want certain sounds, you have to compress, no other process will make that sound.

The truth is, the best mixers are likely to not use compressors on every track. For pop and rock music, vocals and drums are 99.99% of the time compressed. Distorted electric guitars and synthesizers might not be, though. Also, sometimes a track was compressed before it was recorded, so the mix engineer may not want to add any compression to the tracks. A lot of vocals are compressed when they are recorded and then compressed two more times when mixed and then of course compressed again with the whole mix during mastering. But not always. And in genres like jazz and classical and some other cases, hardly any compression may be used at all in any stage of production.

It's not a matter of "controlling peaks" but of having well-manageable tracks to mix: it allows you to position your faders according to the resulting balance you want to be hearing rather than what the musician happens to be doing with the microphone at a given moment.

Basically that's a valid technique for working with close-miked studio recordings of combo music (rock/pop). For orchestral recordings where the dynamics and the balance are more a matter of the conductor than the sound engineer and individual musician's technique, there is quite less use for per-track compression.

I've also found that for more classically inclined accordion music, compression does not really help significantly since the player is already required to put several different voices and parts of the instrument into an overall balance. Using per-channel and/or global compression in significant manner makes the results less transparent as it interferes with the player's effort.

In a similar vein, I'd not expect compression to be greatly useful for complex organ music (though the larger overall dynamics might at least warrant some manual adjustments as registrations change). And typical arranger output will tend to be reasonably useful as is, too.

It's really mostly combo music where per-track compression can make for a good starting point for mastering.