Fig. 1. This screenshot shows parallel compression applied to drum tracks in Pro Tools. Sends are used to route the drum tracks to an Aux track labeled “CmprDrms” (Compressed Drums). This track has a compressor (Kramer PIE) plug-in inserted, creating a new compressed version of the drums which is mixed in with the original drum sounds for a unique effect.

Watch Yer’ Bottom End Many compressors feature a “sidechain,” which acts sort of like a remote-control path. The compressor “listens” to the sidechain, compressing audio based on the sidechain input. Usually, the signal in the sidechain is the same signal being compressed, but what if we could trigger compression by sending a different signal to the sidechain?

Here’s an example: You patch a compressor on a guitar channel and use an aux send to route the lead vocal to that compressor’s sidechain. Now the vocal controls compression on the guitar track. Every time the vocal enters, the guitar level is automatically ducked down a bit (thus the term “ducker”). Attack and release times are usually set fast, so that as soon as the vocal starts, the guitar level drops and vice-versa. This technique is used in radio commercials to create a “donut”—an ad spot where music automatically drops to background levels when the voiceover enters and becomes louder when the voiceover ends. You can use that technique in karaoke situations where you want the MC’s voice to duck the music.

Sidechain filtering is useful when compressing signals with strong content in a particular frequency range. If you compress the L/R mix of a dance track, you may find that every time the kick drum hits, it sucks down the level of the lead vocal. That’s because the compressor is sensitive to the low-frequency energy from the kick. The cure is to apply a filter to the sidechain and remove some of the low frequencies (say everything below 200Hz). Filtering the bottom end stops the compressor from reacting every time a kick drum hits.

A de-esser is actually a compressor with an EQ’d sidechain. The sidechain has an EQ boost in the upper mids and highs (say anywhere from 3.5kHz up to 8 or 9kHz), making the compressor very sensitive to sibilance. Every time there’s an “s” sound, the compressor attacks it and brings it down, but other sounds do not trigger compression. It’s worth repeating that the sidechain is not the audio path. When you see a button labeled “sidechain listen” on a compressor, it lets you temporarily hear the signal that is triggering compression.

Parallel Worlds Another interesting technique is known as parallel compression, whereby multiple versions of the same signal are processed differently. A popular application is to have two versions of the drums in a mix: the ‘normal’ version and then a super-compressed version. Adding the squashed signal into the mix helps stabilize the drum kit’s level and adds body without sucking the life out of them. If you want to get nuts, you can compress for distortion as discussed in the July issue. With a hardware console, you’ll need to route the drums to the L/R master as well as to a bus or aux out. (In live applications, I find it easier to use a bus output—two for stereo.) In the case of the bus out, patch the compressor on the bus insert and set the compressor for a low threshold, fast attack and release, and high ratio. Assign the bus faders to the L/R master and mix to taste. You can also use an aux send to feed the drums to a compressor and return the compressor back into an unused channel where you can add EQ; if you crank the lows at 100Hz and the highs at 10kHz on the compressed signal, you have what is known as the New York Compression Trick.

To do this in a DAW, create an aux send on each channel and route it to a new aux input channel (see Figure 1 on page 98). Insert a compressor on the aux channel and feed the drums to the aux sends. The aux sends are set pre-fader so that moving the drum channel faders does not change the compression. Set the compressor to squash, and you’re in business. In a DAW or digital console, you’ll have to beware of latency, which can at times cause phase problems between the original tracks and the compressed tracks. This is due to the fact that the compressed versions are processed an extra step, which may result in a slight delay. If the DAW has automatic delay compensation, turn it on. If that doesn’t solve the problem, record the processed signal, zoom way in on the original and processed tracks, and manually move the compressed track until it lines up with the original.

Next time we’ll check out advanced gating techniques.

Steve La Cerra is an independent audio engineer based in NY. In addition to being an Electronic Musician contributor, he mixes front-of-house for Blue Oyster Cult and teaches audio at Mercy College White Plains campus.