FabFilter User Forum
What is the MB Compressor for exactly?
I'm been trying to understand this multi band compressor.
I assume you are doing the equivalent of taking the frequency range of the band, removing that part of the signal and taking it to the time domain, compressing or expanding it, then adding it back in.
Is this right?
But I'm having trouble understanding how this is different than EQ. What is the difference? And how does it sound different?
Also, when would I be tempted to use this? What musical problem would I be needing to address that requires the multi-band compressor and for which simple EQ wouldn't do the trick.
It's an interesting looking tool (and I listened to the demo twice), but I'm still not getting it.
Appreciate the info.
Mark — Mar 19, 2014
i understand how it might be a little confusing at first. i would try to think of it less in terms of a variation of equalization and more of an extremely flexible compressor/expander. you're not simply attenuating frequencies (equalization). if you're familiar with the principles of compression, it should be pretty clear where one might want to compress only a portion of a signal. compression/expansion and attenuation are not the same thing.
With an EQ, the amount of gain that you add or subtract to certain frequency areas is always static: it doesn't change over time (unless you automate the EQ controls).
A multiband compressor allows you to change the gain of a frequency band over time, depending on the input. This allows you to do really specific tweaks. I found this background article with more information:
www.soundonsound.com/sos/aug02/articles/multiband.asp
Cheers,
Frederik: hugely helpful article. At least now I have some use cases.
I still have a lot to learn about how to actually use multi-band compression. I had been trying it on a vocal and that's probably not the best place to start. My first application to a drum track was much more interesting.
The sense I'm getting is that it is primarily a mastering tool. If you had separate drum tracks for kick, high-hat etc... the usual gain, compassion (spelling error, but I like it) and EQ tools on each track would be best. But if you have a drum loop already mixed, then multi-band compression is your friend.
I also understand better the phase issues you talked about in the documentation. I was confused since a person can't really hear phase. It's in the recombination of the bands that phase becomes an issue.
I assume you do all this in the time domain. Convolutional bandpass to extract a band, then the usual compression, then add things back together. As I think about it, it's amazing this can be done more or less in real time.
Hi Mark,
If you create a single band, without doing anything (no compression), and then switch between the Linear Phase or Dynamic Phase modes on one hand, and the Minimum Phase mode on the other hand, you should hear phase distortion quite clearly in Minimum Phase mode.
Multiband compressors are often used in mastering because (if used correctly) they allow you to correct problems in a subtle way without changing the balance of the entire track.
Alternatively you can also use them to sonically shape a sound while producing/mixing. Hope this helps!
Cheers,
Vocals are actually not such a bad place to start since a DeEsser is kinda of like a specific case of an MB-compressor and really highlights how it differs from EQ. If a vocalist has strong sibilance around 6k, you could make a cut around that frequency with your EQ. But then obviously you wouldn't just be cutting the sibilance, you would also be taking out a lot of life and sparkle from the rest of the performance. Using a multiband compressor, you could create a band from 5k to 8k and let it compress itself - that way, it will only attenuate during moments of strong sibilance and not the rest of the time. It should also leave all the other bands untouched.
You can use a similar technique to negate some of the proximity effect without cutting out all the bottom of a vocal performance.