Search Gear

Vocal Cords: Compressing Vocals

February 1, 2009

My first ex-wife hated it whenever I bought any musical equipment, because that meant she wouldn’t be able to spend the money on herself. Never mind that I was going to be a famous recording artist some day, and share all the loot with her. (Ha! I guess I showed her!) One time I bought a limiter/compressor (L/C), recorded my vocal through it, and played her the result. When she said it was the best piece of equipment I ever bought, I knew I was really onto something. Honestly, that’s how much difference it made to my recordings—and how much it can make to yours. Fortunately L/C plug-ins come with most recording software, and they’re really not that hard to use.


A compressor controls a signal’s volume automatically by lowering the level on peaks, and raising the level during quiet parts. This produces a smaller level change at the output for a given level change at the input. For example, if increasing the input level by 4dB yields an output level change of 2dB, the compressor is said to have a 2:1 compression ratio. With vocals, compression not only brings level variations under control but brings up softer elements of the voice that add “character.” (Compression is also used with instruments like guitar and bass to add sustain, or drums to tame peaks.)

Another advantage is that if the singer can maintain the same distance from the mic, an L/C can help keep the same intimacy during the entire performance because the tone doesn’t change due to the proximity effect. But a compressor can also be very helpful in the opposite situation: a singer who can’t perform without leaping all over the place. You can level out the performance, even though the tone may change.

Limiting is a variation on compression that’s designed more to clamp peaks to a certain level, but leave signals under that level alone. Limiters typically use a fast attack time (under 10ms, but often less), a high compression ratio (around 20:1 or even 100:1), and a high threshold (the level where limiting kicks in).

Now that I’ve started tossing around terminology, let’s look at the controls and what they do. Most of these devices come with a batch of presets to suit most common needs; Figure 1 shows the list of the Dyn3 presets that come with Digidesign’s Pro Tools.


The human voice has a wide dynamic range, from a whisper to a shout. Because of that variation, it can be difficult to get a vocal to sit nicely in the mix where you can hear all the lyrics. A compressor is often a big help.

Generally you want to set up dynamics control so you don’t notice it’s there, but plenty of engineers like to use compression as a special effect where it actually is noticeable. For example, the only way you can get that in-your-face effect used for the vocals in a lot of headbanger music is with a compressor (think AC/DC).

The graph in Figure 2 on page 40 really helps to see how the controls affect the signal. The line begins moving upward at 90 degrees from left to right in a one-to-one relationship: When the input signal increases by one decibel, the output also increases by one decibel. That’s what we hear in real life. After the signal exceeds the threshold or knee (indicated by the vertical red line), the compressor attenuates (reduces) the signal based on the ratio you select. When using the Vocal Levelor preset with a 3:1 setting, for every 3dB the input signal increases above the threshold, the output increases by only 1dB.


By setting a low threshold (–30dB), compression kicks in at a correspondingly low volume, reducing high levels and raising lows. But now, we’re no longer taking advantage of the available headroom because of the reduced levels for the peaks. So, we boost the plugin’s output gain so that the peaks once again hit 0 at the output; the overall average level is now much higher. To give you some idea of how compression affects the audio, take a look at the waveforms in Figure 3. This is before raising the compressed signal’s gain so that its peak levels match the uncompressed version.

On the downside, a low threshold setting also makes background noises more audible because it brings up lower levels. You need to be particularly cautious about sounds like heating, air conditioning, computer fans, and buzzing amplifiers.


Now let’s looks at the other controls. The attack is measured in milliseconds, and determines how quickly the unit responds after the input signal crosses the threshold setting. Increasing the attack will let some transients through, making the material sound more natural and a little less squashed.

The release setting determines how quickly the compressor returns to the uncompressed state after the signal drops below the threshold. Adding some release time (80–250ms) can help the material sound more natural. It’s also common for compressors to have an "auto” option that changes the release time dynamically to best accommodate changes in the input signal.


The side-chain section (also known as “detector”) in the upper right in Figure 2 allows some other signal to control the compression. For example, a side-chain can provide automatic gain control or “ducking” of background music when a narrator or DJ is talking—just send a bus out from the announcer’s voice track to the side-chain input on the music track. The sidechain input also has musical uses, like making a sustained instrument “pulse” when another instrument (like a kick drum) controls the compression.


While you can certainly over-compress, I think it’s also possible to under-compress. I have a Diana Krall CD (Only Trust Your Heart, GRP Records) that was recorded live to two-track with no compression. When I listen in the car, I have to keep my hand on the volume knob because her voice keeps dropping below the road noise. The same is true of classical music.

I’ve also found that having a compressor in my stage rack is invaluable when singing live, as it can increase intelligibility. The unfortunate side effect is that it also increases the possibility of feedback; to control that, I use a pink noise generator and calibrated mic to tune my EQ unit to the room, along with a feedback exterminator to catch any residual feedback issues.


Except for special effects like those we’ve already discussed, be careful not to squeeze the life out of the performance. Use as little as you can get away with that still gives the performance all the attention it deserves. You can start with the presets and tweak as you learn; just don’t get carried away with your new toy and create something you’ll be embarrassed by a year from now. Do it right, and your spouse might even say it’s the best piece of equipment you ever bought.

Keep up-to-date on the latest news
Get our Free Newsletter Here!
Show Comments

These are my comments.

Reader Poll

Are you a gear DIY-er?

See results without voting »