On the other hand, if you’re recording a soft tune with an intimate vocal, you need to be more careful about how the compression is sounding. If the compression in this kind of setting is noticeable, it can ruin a performance; it sounds unnatural and detracts from the performance’s intimacy.
Unless you’re using heavy compression as an effect, try to make the compression as transparent as possible—it should sound like it’s not even there.
If you can hear the compression kick in during playback, try using a lower ratio as well as a slightly lower threshold. The lower ratio will make the compression sound less heavy-handed, while the lower threshold will bring down the hottest peaks’ levels to about where they were with the previous settings.
If you end up being the only person in the world who knows that a track was compressed, then you know you’ve done a good job applying compression.
Why Compression Is So Important
It might be a good idea to compress just about every track you record, before you record it.
An exception is commercially sampled sounds (including drums):
—First, because sampled sounds are often already compressed somewhat;
—Second, because you may be bringing these into the mix by way of MIDI and therefore aren’t actually recording them during the tracking phase;
—Third, because sampled sounds have more even, predictable dynamics than live acoustic sounds.
For everything else, adding compression while tracking just makes everything much easier.
Here’s why: Acoustic sounds, as well as many electronic sounds, tend to vary quite a bit in level.
If you’re recording a shaker percussion part without any compression, the loudest shakes may be a good 9 dB louder than the softer shakes.
Meanwhile, on your uncompressed rhythm-guitar track, some chords may be right up near the top of your system’s dynamic range (0 dB), while others are at –4 dB.
Figure 4: Uncompressed tracks are hard to mix because their levels are constantly changing (click to enlarge)
This kind of variation could be occurring on most of your song’s tracks, including the all-important lead vocal, throughout the whole song (Figure 4). It’s extremely difficult to create a consistently good mix of a song under these circumstances.
Even if you could get the mix perfect for one moment in time, a half-second later the mix could be totally different—the rhythm guitar could suddenly get much louder and suffocate the lead vocal for a moment, which may have gotten suddenly quiet at that moment anyway, making matters worse.
You have no real control over the blend, and the end result could be a random pastiche of obnoxiously loud moments and inaudibly buried moments.
Compression solves this by getting levels under a certain amount of control. If the levels of each instrument vary by only 2 dB or 3 dB, you can achieve a much smoother mix than if each instrument is varying by 9 dB.
From the song’s beginning to its end, the mix will be much more consistent than a similar mix of uncompressed tracks (Figure 5).
Figure 5: The same tracks with compression can be mixed in a much more smooth, consistent way (click to enlarge)
Simply stated, it makes things easier in the studio.
And since we Guerrilla recordists mix as we go, we need to have the tracks compressed as they go down. Million-dollar facilities may be able to compress each track separately during the mix; we can’t.
The process known as “de-essing” is a type of compression. If you record a vocal track very bright, adding some top-end EQ, you’ll get a very present, airy sound—two good qualities to have on a vocal track.
However, moments where words have “s” sounds can cause an explosion of noisy sibilance. Sibilance refers to excessive high end associated with the sound of the consonants S, Z, etc.
Sibilance can mess up an otherwise well-recorded track, and if you have an effective way to tame a track’s sibilance, you should use it.
The most common way to do this is by de-essing the track (Figure 6).
Figure 6: In de-essing, the signal is split, and compression is applied based on its high frequencies only. This is done by sending a high-pass-filtered signal into the compressor’s side-chain input (click to enlarge)
De-essing refers to compression based only on the signal’s high-end frequency content.
Here’s how it traditionally works using hardware compression: An audio track is split into two parts, and one of the parts goes through a high-pass filter, which removes everything but the highs—say, everything below 8 kHz.