Multi band compression

R

rj

New member
A term I have heard in the topic of Mastering is 'multi band compression', and by my meager understanding, this means to add compression only to certain bands or "frequencies".
Now, I know every mix is different, but, are there certain "frequencies" that would usually be compressed in mastering.
The vocals? The snare? 3k 5k 7k 10k?
Kind of like the way there are certain guidelines for us ameteurs to start boosting/cutting when mixing a bass..100 800 3k etc...
Does somebody have ANY input here?
Lemme save the guy thats gonna tell me to trust my ears.. I WAS BORN WITHOUT THEM AND i AM ASKING YOU!
Thanks, Rj
 
You might start your search with DBX. They are offering a product that does multi band compression as part of the mastering process. Perhaps the info on the product would have some clues.
 
Low end information has much more acoustical energy tha hi end. With full range compression, you can easily get the boomy parts of your signal (bass, kick drum, etc.)triggering compression that attenuates your hi end overly much. (I. e., the comp. is pulling down your high end while responding to the low end.) That is why compressed signals lack presence compared to the original signal. After-the-fact corrective eq doesn't quite compensate for this effect.

Multiband compression employs an internal crossover to separate lows and highs and assign separate compression parameters to each range. This allows you to dial in low end punch while leaving the highs unaffected.

I used 3-band compression mastering my last disc, with a TC Electronic M3000. The results were fantastic. Behringer has a new split band compressor for $200. Get down with it!
 
The poor man's workaround for the above problem is to trigger your compressor with a sidechain signal consisting of the original signal that has had the highs boosted with eq so that their energy levels match or compete with the lo end. Experimenting with sidechain eq is an interesting experience, and you can try this with your existing full range compressor, if it has a sidechain trigger input. Squash on!
 
Some important things have to be remembered regarding multiband processing.

First, one has to be careful when choosing the crossover frequencies and compression parameters, i.e. compression ratio and attack and release time.

The correct crossover frequencies will influence directly in the compression performance. The best try for a begginer is a 2 band compression, so he (or she) can get the feel of the stuff, without messing up things.

Another point that one must have in mind is that the high frequencies (above 3.2 kHz) should be limited, and not compressed. The pumping effects are more noticeable in the highs (pumping is a sudden increase of gain after a loud passage followed by a light one).

When compressing the mid and low-frequency content, be aware that the compressor should work with both channels linked. This small trick provides a natural sound. On the other hand, the highs should be limited separately (not linked) so the stereo balance will be conserved (the human ear is more spatially sensitive to the mid-highs and highs).

The key for a good multiband processing is trying to keep the overall sound dense, but not fatiguing. Women are extremely more sensitive than man to ear fatigue. If you are processing dance or high-energy music, let the GR meters lit, watch the VU and go for it. But if you are processing soft music or anything light, take it easy.

Another point: sound is subjective. What sound great and loud to you, can sound terrible to another person. This happened at the radio station I work, the programation manager messed up the multiband processor, and told me that the station sound was awesome. In fact, it was a real crap. A heavily processed material is not always the best choice.

Try processing different songs, with different adjustments, and soon you'll get the feel for it.
 
You must log in or register to reply here.
Back
Top