by Dave Foxx
So how do you use your compression? Are you a virtual voiceover head-banger, pushing the compression threshold to the max every time? Or are you more of an audio purist who strives to keep the voice as pristine as possible, depending on an excellent microphone and super clean mic-pre? Chances are excellent that you fall someplace between, pushing for a bit more loudness without thrashing the plosive diphthongs.
This month I’m going to let you in on a little secret of the music recording industry that somehow never found its way into everyday broadcasting use. I know a few VO artists who use it to great effect and are probably willing me to stop writing immediately so they can keep that competitive edge over the other VO guys and gals. I can practically see a macho VO artist, sitting at the breakfast table in his robe, drinking a bucket of water to wash down the yogurt and fruit, casually perusing this article, suddenly saying, “No. NO! Don’t you DARE say it!”
Parallel compression. There, I said it.
Now, some of you have been using side-chain processing to duck your music/effects tracks automatically to the voiceover for years now. You add processing directly to your music/effects track, then side-chain it to your VO track. Every time the VO is going on, the music is ducked automatically to the parameters you’ve set in the compressor. With a bit of finesse, you can make the music track respond directly to the VO perfectly, every time and never have to duck the music manually again. This works exceedingly well in Pro Tools AND Adobe Audition. It’s easy to set to set up and adjust on both workstations, and I’m sure it’s the same with others.
Music producers have been using a similar technique for many years to create a truly amazing dynamic with the kick drum, particularly in dance music. Imagine having the bass get louder when the kick drum hits, punching up the rhythm by a factor of four or five. It gives the music an absolute solid bass line without adding an ounce of distortion. Bring EQ into the side-chain and the rhythm fairly pops out of the speakers. Judicious use of side-chaining can open up the dynamics of a music track tremendously, regardless of the format.
One of the most famous examples of this process is the group Daft Punk. Given the fact that they play keyboard and bass, with an electronic drum kit, they should have an extremely basic sound that is relatively clean. However, they use the Alesis 3630, a notoriously dirty-sounding analogue compressor to great effect as a side-chain processor, giving them a marvelous grungy, growling, howling sound that is uniquely Daft Punk.
Parallel compression is side-chain processing… modified to have a track react to itself. The net effect in voiceover is a louder VO without losing all of the ‘natural’ parts of the human voice, the parts that make it sound real and down to earth. To me, this is the promise of compression, delivered. The best part is, it gives you absolutely precise control of the VO track, making it strong AND vibrant, dynamic AND powerful, all at once.
Parallel compression is set up pretty much the way reverb is configured and really only works well with a handful of compressors. The original signal is split, using the SEND bus, rather than a regular INSERT on a channel. The original channel gets some very light compression to make sure the beginnings and ends of words come through. The NEW channel gets the heavy duty compression, which is where the loudness really happens. If your system allows for it, make sure the SEND gets the raw audio, and not the lightly compressed track. This will keep the process streamlined and easy to troubleshoot. If you can’t do this on your DAW, you’re probably better off with NO compression on the original track.
You now have instant access to a WETNESS factor or… how much of the signal is über-compressed and how much is allowed to come through in its raw state. You can drag the bus RETURN fader up and down any time while leaving the original track at parity (±0.0db). The higher the gain is on the return, the higher the wetness of the effect. A good starting point is always 50%. I tend to go down from that, but not always, and that will change depending on the VO artist and the kind of timbre he or she naturally projects.
On the RETURN track, the compressor threshold is typically set pretty low so that practically everything is compressed. The ratio is also high, say 20:1, the knee should be soft and the attack is set to slow and release set to a bare minimum. This sets up a kind of ‘umbrella’ you can put over the original to boost the main parts of the original track, but still allow the original track to handle the beginning and end of individual sounds, giving an overall feeling of a non-compressed signal. Once set up, you might find it best to play with the knee first, if you’re not getting the impact you like.
It is absolutely essential that you use a high-end compressor, capable of handling just about anything without breaking down into distortion. I have found that the ‘built-in’ plug-ins seldom have the juice to deliver the kind of clarity you want. The single instance I’ve found of a “comes-with” plug-in that is up to the task is the Avid Dynamics III compressor. Otherwise, you will probably have to get a third party compressor by WAVES or PSP to make this really magical. The RETURN track must be pristine or the whole effect is lost and you are really just adding another layer of noise to your track that will pretty much undo everything you’ve tried to accomplish. ALSO, this is one instance where you cannot use a limiter as a compressor. You really need to have the knee, attack and release controls to be effective with this and very few limiters give you all of that.
In addition to side-chain processing, music producers will often use parallel compression (having a track react to itself) on individual instrument tracks that tend to be much more quiet than others. You might see it on a bell tree track, but not on the tympani. You might see it on a French horn, but almost never on a trumpet. Of course, it can be used for effect too, so nothing is ever absolute.
Using it on VO has its own set of issues, mainly that when you amplify a whisper, you not only bring up the VO, you also tend to bring up all the mouth noise, clicks and glottal junk, room noise, echo and hum that are easier to ignore or otherwise take care of with good noise reduction and EQ. Remember that compression of any kind means less dynamic range, which is really easy to overdo. Like a really good pepper sauce, too much can burn out the taste buds and leave you craving relief. Just the right amount can add a delightful sting that will leave your audience wanting more.
My big hope is that this will lead you to experiment a bit with your track setup. Experimentation leads to all kinds of exciting discoveries and usually makes a producer’s work stand out from the pack. THAT, dear reader, is the ultimate goal of this column.
For my entry on the SOUNDSTAGE this month, a seasonal throwback from 2011 for the Z100 Z-Free Money Sticker! The “old bumper sticker” promotion is still one of the simplest, most direct and exciting promotions any radio station can do. The winners get extremely excited (which ALWAYS makes for better promos) and there are always so many to choose from. As it happens, I produced this at just about the time I was first experimenting with parallel compression. I was also using side-chain processing to auto-duck the music and effects. My goal at the time was to have the entire promo mix down without any of my faders moving at all. Yes, mission accomplished.
Dave welcomes your correspondence at