Following up on our article about de-essers, today we look at multiband compressors. Both are frequency-specific (a la EQ), compressors, but whereas de-essers only reduce sibilance, multiband compressors can be used on any part of the spectrum. For podcasters there is one great and simple way to implement multiband compression on our vocal tracks.

This is to bring down harshness in the high end without totally flattening the sound. As we’ve mentioned before, a lot of inexpensive microphones can have uneven response curves in the high end, making voices sound overly sibilant. This, coupled with low-bitrate mp3s propensity to distort high end information, can make for a nasty listening experience. To mitigate this, we use a multiband compressor on the 8-20 kHz range, much in the same way as we de-ess the 4-10 kHz range.



Unlike the sibilant range, however, the 8-20 kHz range still has information that can be nice sounding in moderation. So setting a threshold level of -20 dB, and a fairly high ratio (10:1 or so), is a good place to start. From there, adjust the levels until you start to hear the high end smooth out. It shouldn’t sound lopped off, more contained.  

Here we’re compressing the high end to make it sound smoother to listeners. Notice the relatively low threshold (-28 dB) and high ratio (10:1), still allows information from the high end to be audible, but we’re setting an upper limit to how loud it can get. This gives our audio a “smoother” sound, reminiscent to how analog recordings have a tendency to “roll off” the high end, or have a decreasing frequency response over 10 kHz. The Beatles would often record tambourine overdubs to help give their cymbals more definition in the mix because of this rolloff.

Use Sparingly. Apart from taming the high end, you won’t need a multiband compressor very often. In fact, most music production sites caution against using multiband compression unless you’re an expert, because of the skill needed to blend crossover points well. Our example isn’t at as much risk for this, as our crossover point brushes up against , which is already heavily compressed. 

Trust your ears, again! This is another one that goes without saying. Your ears will be the best indicator of when you’ve taken enough off the top, as it were. If you’ve spent too long trying to get the level right, take a 15 minute break, then pop on your favourite podcast and listen to their high end for a bit. When you go back to editing, you’ll immediately know how close to the mark you are.

Compressor, De-esser then MBC. At least this is the order that has worked best for me. Again, we don’t want things that we are trying to bring down with the multiband compressor brought up again by our level-matching compressors.

This is a rather short article because so much of what we needed to cover was already done in the EQ, compression, and de-esser articles! And it’s the last in our series on per-track processing. Next article will be about the (light) processing to do on the master bus, and after that will be my next guide on audio repair for free!