Mel Bay's Guitar Sessions® Webzine | Phil Gates - Compression: Friend or Foe?

Makin Trax

Compression: Friend or Foe?
by Phil Gates

If ever there were a tool that was used in a more broad-stroke general fashion, it would be the compressor. Going to record vocals? Put a compressor on it. Mixing vocals? Put a compressor on it. Having Problems with vocal levels? Put a compressor on it. Mixing down before mastering? Put a compressor on it.

WOW!

I should just compress the life out of everything and make it sound loud and smooth right?

OK, maybe we've hit a nerve here. I apologize. But I get tired of hearing what could be great tracks compressed so needlessly, or badly, that I want to scream. Compressors are a great tool to use in the recording & mixing process. I use them at different times, never just "Always" though. Compression is like cooking with salt. Use a little, everything's cool. Go one step too far, and you've ruined the meal. Luckily for us in recording, we can undo, or bypass to check to see if there's too much compression going on.

Just what is a compressor?
A compressor is a device that reduces (or compresses) the dynamic range in a sound source's softest point to its loudest point (dynamic range). This typically will result in a more smooth output. Recording vocalists new to recording can be a challenge, because they are typically not having a very controlled voice. So they may unexpectedly change volume on you while recording. This is usually due to poor mic technique, or just plain inexperience. If you have to do the session anyway, a compressor might help this vocalist sound a little better. It also will protect your gear from transient peaks during the recording of the track.

There are four basic components to a compressor:

Threshold: Threshold means, at what level of dB do you want the compressor to start compressing? If you want it to barely be used, a number closer to Zero would be used. For more compression, go into a more negative number.

Attack: Attack means, once the threshold level has been detected, how fast or slow do you want the compressor to kick-in? Do you want to compress right away? Or perhaps have the compressor come on more slowly.

Release: Release means, that once the compressor has kicked-in, and now is going back above the threshold level, how fast or slow to you want the compressor to "ease-up" and no longer compress the signal.

Ratio: Ratio means, this is how much the compressor is going to change, or reduce the dynamic range of the signal. For vocals, typically a setting would be around a 4:1 ratio. What this really is doing is that for every 4dB of increased volume at the input of the compressor, the output will only go up by 1dB. So a 2:1 setting would mean the compressor output would change more easily, and a 12:1 ratio would mean that compressor output is barely going to budge!

A compressor is really at its most basic level, a leveling device. While this may sound like a very tempting cool tool, there are side effects to misuse; it can take the life out of a track if used too harshly. Although, using a compressor like this can be a handy effect to use at times. But that's using a compressor as an effect device, not as a natural compressor.

Audio Example 1: No compression

Now, I'm going to add a typical compressor using Apple Logic Pro's Compressor plug-in: The settings looked like this:

With the threshold set at -14.5, the compressor can stay fairly subtle, yet useful. The compression ratio is at a good starting place: about 4:1. This will even out, but not clamp down too hard on a vocal track. The attack is set a little shorter, but not enough to take away the starting edges of words, and the release is set to not turn the compressor on and off after every word, but over each phrase. Now the vocal track sounds like this:

Audio Example 2: Normal compression

Notice that with no change in EQ (still none being used), it almost starts to get more nasal like as we start to squeeze the dynamics, although it is louder, and more smooth. As you can see EQ and compression will have to work hand in hand on your projects.
Now here's the over compressed version. Take a look at these settings. This can happen when novice engineers are desperately trying to smooth a track:

This sounds like this:

Audio Example 3: Over compression

Take a close listen to the phrases "Moving" and "Nowhere Else" in the last three examples. Notice how the "ing" of "Moving" starts to sound funny, and the "s" of else seems to really start to disappear. Then listen to example 3, then example 1 after it. You can hear the difference in the edge, or life of the sibilance in the vocal track verses a duller sounding track when it has too much compression on it.

So what is the answer?
For myself, on this project, I'm most likely to use my little vocal toolbox with Hydratone EQ, and other EQs from Tritone Digital, (http://www.tritonedigital.com) , a compressor from Logic Pro, that I'll adjust for a very light compression for that track, and ample use of automation. I like to use automation to make the volume changes that I need to keep the vocal track smooth, yet keep the original tone of the vocal. I use this technique on many instruments in my mixes, and, as with almost all audio software, this is a built in feature. So you have the tool there, might as well use it? Right? Other options are to rehearse the vocalist! Now there's a concept! Or to ride the record faders (input level) while recording the vocal, which means you should be rehearsing these moves while you're rehearsing the vocalist.

So let's look at the other usual suspect for compression, the snare. Let's hear an uncompressed example:

Audio Example 4:Snare no compression

A nice, simple track, not gated, or EQ'd at all. Now, let's throw a typical compressor on there:

Again, take a look at the settings, then listen to the example:

Audio Example 5: With compression

Using a compressor with the attack set back some like this, still gives the leading edge, or attack of the snare intact. Yet it still gives the volume boost and more full sound.

Listen to when I change only the attack time to a very short attack time. Like this:

Audio Example 6: 6 fast attack

Can you hear the difference of the leading edge of the snare? It has less snap to it; which tends to soften the snare sound. This can be useful in some cases when a drummer is hitting too hard, but if you want a crisper snare, increase the attack time some. I'm a big fan still of the combo of EQ, compression, and automation. With some gating if I think I want it.

Now here are someone's settings that are trying too hard to settle the snare into a mix:

The threshold is at almost -30dB, so the compressor is always on at this point: Take a listen:

Audio Example 7: Over compression

Very squished indeed! While I would never probably use this sound for my main snare sound, I might use it in parallel to the snare track for a more meaty snare without raising the volume fader of the snare.

As you can see and hear, compression can be helpful or hurtful depending on how used. In many cases compression isn't even needed. Don't assume that you HAVE TO use compression. Try it with and without. See what works best for you. Really play with the settings. Often the result is very subtle. Train your ears for it. Also, perhaps only a portion of the whole track needs compression. Use automation to turn it on and off. I suggest go play and see where in your mix compression is you're friend, or your foe.

Have fun,

Phil Gates

http://www.philgates.com
www.myspace.com/philgatesmusic

[ top ]