Understanding audio compression…
Take the mystery out of audio compression by learning how to set a compressor so you improve your audio tracks instead of demolishing them.
Audio compression mystified me. It made me crazy when I was trying to figure it out. I’d turn knobs and press buttons and it either made no difference that I could hear, or the track suddenly sounded like I’d destroyed it. Compression felt totally random until I finally understood how all of the settings worked on a compressor and learned how to set a compressor properly.
In this article, you’ll skip the “it’s making me crazy” phase so you can not only
- set a compressor to improve the sound of your tracks, but you’ll
- learn the basics of audio compression
- so you understand what each setting can do and
- how to get the sound you want out of any audio compressor plugin.
Below the video lesson is a summary of the strategies, with bonus ideas and links to other songwriting resources to improve your songwriting!
Audio Compression is…
At its simplest:
Audio compression makes the quieter parts of a recorded vocal or instrumental notes louder so they sound better.
This article and video are practical so some of the deep acoustic theory is simplified or skipped. You’ll be able to set your compressors so your mix sounds better and understand how to use audio compression without making your head hurt or your eyes glaze over!
Audio Compression is Confusing Because…
Audio compression is a two step process.
First, a compressor lets the start of the sound (attack) through, then activates and turns down the volume (compresses) of the rest of the sound.
Then the compressor turns the sound it just processed (compressed) back up, making the quieter parts of the sound louder.
When it’s set correctly, a compressor makes the softer parts of the sound seem louder, without making everything too loud.
The trick is understanding how to set a compressor properly!
There are too many parameters to figure it out using the “try random things until it works” method…
Audio Compression Parameters
The sound in the track you’ve recorded and are processing in your DAW. Measured in -dB, 0dB is as loud as possible. The higher number in -dB the softer the sound.
The threshold is the signal level that activates the compressor, measured in -dB. The higher the number, the lower (softer) the signal needs to be to turn on the compressor.
For example, if the threshold is set to -12dB… when the signal goes above -12dB the compressor turns on. Set the compressor below the loudest level of the signal, otherwise it won’t turn on, and above the softest level, otherwise it will be on all the time and you’ll over compress the signal.
The ratio is the amount that the compressor reduces the signal. It’s measured as a ratio X:1, such as 2:1 or 4:1. The higher the ratio the more the compressor reduces the signal when it’s louder than the threshold.
At 2:1 every for 2dB above the threshold compressor reduces the signal to 1dB louder. So each 2dB louder than the threshold is reduced to 1dB louder, 20dB louder becomes only 10dB louder.
At 4:1 every for 4dB above the threshold the compressor reduces the signal to 1 dB louder. So 2dB louder than the threshold is reduced to .5dB louder, 20dB louder becomes only 5dB louder.
Most of my compression is between 2:1 and 4:1 for vocals and instrumental tracks.
The attack is how long the compressor waits before turning on after the signal crosses the threshold, measured in milliseconds (ms). The lower the attack, the faster the compressor will turn on and reduce the signal.
A fast attack reduces the volume of the start of the sound. Use a fast attack when you want to hear more of the end of the sound than the attack. Generally, vocals use a faster attack than instrumental tracks… use your ears. If you have harsh consonants that you’re having trouble de-essing (see below) you can sometimes use a very fast attack to help reduce them.
A slower attack lets more of the start of the sound through, use a slow attack when you want to hear some of the articulation of the vocal or instrument. The more percussive the sound, the slower (up to a point) you want to set the attack.
Compressor plugins can have an attack of 0ms (this isn’t possible with hardware because it takes at least a millisecond for the signal to trigger a hardware compressor) but plugins can “look ahead” and be ready to instantly trigger the compressor. A limiter is a specialized compressor with a really fast attack time.
The release sets how long the compressor stays on after the signal falls back below the threshold, measured in ms. Be sure to set the release time so the compressor turns off before the next note begins, otherwise the compressor just stays on all of the time.
A release time that’s too short won’t compress the signal long enough. When it’s noticeable enough to be irritating it’s called pumping.
The knee is the dB range above and below the threshold where the compression transitions from 1:1 (no compression) to the ratio you’ve chosen. Many compressors have a fixed knee that can’t be changed.
With a hard knee (0dB) the compression ratio is constant and starts immediately. A softer knee adds compression less suddenly, starting with less than the ratio the compressor gradually increases to the ratio you’ve chosen.
The amount of signal that’s lowered by the compressor (in dB). I prefer 2dB to 6 dB gain reduction for most situations. I find more than 6dB gain reduction for acoustic instruments or vocals is usually too much. It starts to sound like the dynamics of a sound, a note or a performance are lost and the track becomes mechanical and boring instead of lively or lifelike.
Over compression is often called “squashed” or “flattened” because the natural dynamics of a sound or performance are eliminated by the compression.
The amount of compression you want to hear depends on the style of music. For example, Hip Hop and EDM (electronic dance music) use more compression than an acoustic folk performance. An EDM track without compression would sound odd, just as an acoustic performance with excessive compression would sound strange. Use your ears and compare your mix to other songs in the same genre.
Gain (Make Up Gain)
Once a compressor reduces the signal, make up gain brings it back up to a level you’ve specified. The idea is to match the gain reduction and the make up gain. This leaves the attack (start) of each sound/note the same (because the compressor hasn’t started compressing yet) but makes the rest of the note (that’s been compressed) louder.
Balancing the gain reduction with make up gain makes the softer part of a sound seem louder (without making the attack louder).
Bypass temporarily turns off the compressor. Use the bypass button to compare the sound with and without compression to be sure you’ve improved the sound with the compression settings you’ve chosen.
The help button is your friend… read the help file, especially with a compressor you’re not familiar with. It’s surprising how many problems you can avoid with a little reading.
Other Compression Terms
Many compressors have a mix knob to blend the (wet) sound of the compressor with the (dry) unaltered signal. This lets you compress the signal more heavily and then mix the two together to your taste.
This control isn’t on the Sonnitus compressor in the video demonstration.
a compressor with a ratio of infinite:1, all signal above the threshold is set at the threshold, for example if a limiter is set to -5dB if the signal is louder than -5dB it stays at -5dB and can’t go higher. Some compressors can be set to act as a limiter.
Literally removes “s” sounds… a specialized compressor that reduces sibilance (harsh consonant sounds) between 2kHz and 10kHz.
See how I set a de-esser to reduce the consonant sounds in this song when I set the EQ for vocals in Mixing with EQ.
Audio Compression Best Practices
Compression feels confusing because a compressors main function is to make sounds softer by compressing them. The missing step is adding makeup gain, so the attack of a note sounds the same but the quieter part of the rest of the sound seems louder. This makes it seem like the quieter part of the sound is louder.
The attack time lets the start of the sound through normally. When the compressor cuts in, the ratio sets how much the sound is turned down by compressing it but then compensates with make up gain. This makes the softer parts of the sound seem louder compared to the original sound.
Some compressors only have a few settings you can change, some parameters are constant because these plugins simulate classic hardware compressors. They still sound good if you experiment a little and learn how to set them.
Use low ratios between 2:1 and 4:1 for most situations, don’t over compress your tracks.
Adjust the threshold to compress a few decibels of the signal (I usually keep it between 2 and 5 dB for most situations)
Keep the release time shorter than the time between new notes, otherwise the compressor just stays on all the time and doesn’t improve the sound.
The key to using compression is properly setting the make up gain. Check the compression meter so you know how hard it’s compressing and match it with make up gain. You can also set a compressor by ear. Switch it on and off and adjust the makeup gain so the track seems to be the same volume whether the compressor is on or off.
When you’re done, just like EQ, listen to your track and your mix while you turn compressors on and off so you’re sure you’ve improved the sound of your tracks and your mix.
Audio Compression Summary
The start of the sound (attack) is the same, but everything afterwards seems to sound louder. It’s confusing because compression is a counter intuitive, two step process, compressing a sound then turning up again to make it sound better. Less compression is usually better than more compression.
- Set a small ratio between 2:1 and 4:1
- Setting the threshold to trigger the compressor
- Set the attack to let the beginning of the sound through before the compressor activates
- Check that the release time is less than the time between notes
- Click the gain reduction meter to reset it and check the reading
- Set the make up gain to match the gain reduction
- Bypass the compressor to compare the sound with and without the compressor.