This Is Your Mix On Compression - Part I

by Donald Dinsmore

Compression is central to a modern producer's arsenal.  You've probably heard all about compressors and the amazing things they can do to your sounds (and maybe some of the not-so-amazing things they can do too).  Compressors are tricky devices to wield because often their effect is not obvious.  All of the moving parts found on a compressor only add to the confusion.  Is the compressor attacking or releasing and how do the threshold and ratio play off the other parameters?  All this confusion has manifested itself in often more perplexing articles, blog posts, YouTube videos and tutorials filled with truth, lies, myths and legend. This blog post will be the first we release in a series on compression.  Today's post will define compression and explain what the different controls found on most compressors do.  The remaining articles in the series will discuss how to use compression in specific ways on different elements in your tracks.

WHAT IS A COMPRESSOR?

Trying to answer this question isn't an easy task.  The reason being is that compressors can be used for so many different tasks.  In this case, the best definition is probably the most general.  A compressor is an audio effect reduces the volume of certain portions of a sound thereby altering the sound's dynamic range.  Before you cry bloody murder and say things like, "compressors make things louder!" hear me out.  Compressors initially designed to solve a problem.  Broadcasters found that the level of audio they were recording and transmitting would change in volume rapidly, going from quiet and to loud.  The quiet parts posed no significant issue, but the loud parts of the signal had the potential to overload circuits and fry electronics.  Consequently, for years someone sat at the mixing board with their fingers on the faders and would reduce the volume of the signal if it became too loud.  Once the volume of the signal dropped, the engineer would slide the fader back up.  This worked reasonably well but there was still an issue, sometime the engineer would respond too slowly to the spike in volume.  The solution was a circuit that would reduce the volume of the signal automatically if it became too loud then return the volume of the signal to its original point after the volume spike had passed. Thus the compressor was born.  At the end of the day a compressor is just an imaginary guy sitting in your DAW reducing the volume of tracks if they get too loud and then turning them back up.

Because compressors turn down the loudest parts of an audio signal they change the relationship between the peak volume (volume of the transients) and the RMS volume (root mean square, think of it like an average).  If the compressed signal is turned up so it's peak volume is identical to the uncompressed signal, the RMS volume of the compressed signal will be higher than the uncompressed signal and appear louder.  When the human ear picks up sounds and send them to the listener's brain, they can be perceived as quiet or loud.  Sounds our ears and brains consider loud start loud and stay loud.  This means sounds we perceive to be loud have high RMS volumes.  Certain sounds may have loud peaks but our ears are likely to underestimate the volume of those peaks and focus instead on the RMS volume.  In the future when you hear people say compressors make things louder you'll know that compressors, in a roundabout way, can be used to increase the RMS volume of the audio signal.  You'll also know compressors achieve this effect by reducing the peak volume of the loudest parts of the sound and subsequently increasing the gain of the entire signal, leaving a compressed signal with a smaller dynamic range (difference between loudest and quietest) than the uncompressed signal.

These settings were used to heavily compress the drum loop below.

These settings were used to heavily compress the drum loop below.

If the difference between peak volume and RMS volume is confusing to you visual aids may help.  When Ableton released Live 9.5 they changed the metering in Ableton to include peak and RMS volume (before 9.5the meters in Ableton only showed peak volume).  The bright green portion of the meter indicated RMS volume, the greyed out green portion of the meter indicates peak. 

WHAT'S IN A NAME?

Unfortunately our analogy of a tiny mix engineer sitting in your DAW only holds up for so long.  Most compressors whether hardware or software have several different knobs, button and meters on them.  The most common words you will run into are the following: gain reduction, threshold, attack, release, ratio, peak, RMS and makeup.

Gain Reduction

This is often displayed as a meter on a compressor and refers to how much the volume of the signal coming into the compressor is reduced by.  

Threshold

The volume level at which the compressor begins to reduce the gain of the signal.  You can almost think of it as how often does the compressor kick in.  If you have a high threshold, the compressor will only turn down the loudest peaks of the signal as they cross the threshold.  That means any part of the signal that is below the threshold will pass through the compressor unaffected (uncompressed).  If the threshold is low the compressor will catch any part of the signal that exceeds it.  Meaning the loud peaks and the less loud peaks will be compressed.  Therefore in most cases, the compressor will kick in more often with a low threshold than with a high threshold. 

Threshold.jpg

Attack

The length of time it takes for a compressor to reduce the gain of a signal once it crosses the threshold.  In other words, how fast does the imaginary engineer turn down the signal once he recognizes a peak. You may have heard different definitions of attack such as, "how long does it take for the gain reduction to kick in after the signal crosses the threshold," but this isn't accurate.  The moment the threshold is crossed the compressor starts reducing the gain of the signal (See below, A Note on Attack and Release).

Release

The length of time it takes to restore the gain after the reduction of gain.  If we think about our little friend in our DAW it means how fast does he turn the signal back up after he turned it down.  Again you've probably heard definitions like, "release is the time it takes to restore the gain after the signal drops below the threshold."  This is false (See below, A Note on Attack and Release).

A Note on Attack and Release

If you've been paying attention you've probably realized that a compressor, when active, is either attacking or releasing.  If the gain reduction meter is increasing, the compressor it attacking and if the gain reduction meter is decreasing, the compressor is releasing.  If you want proof of this load up a drum loop in your DAW and slap a compressor on the track as an insert effect.  Pull down the threshold so that the entire signal exceeds the threshold and the gain reduction meter never goes down to zero.  If the attack and release only affect the sound above and below the threshold, you would expect them to have no effect on the sound if you change their settings.  But as you'll notice, altering the attack and release times will change the sound of the drum loop.

Ratio

The Ratio is the degree to which the compressor reduces the signal when the signal exceeds the dynamic range.  You will often see ratios expressed like this, X:1.  For example, say you have a ratio of 10:1, this means if the input signal exceeds the threshold by 10dB, the output of the compressor will be 1dB above the threshold.  In other words, for every 10dB your input exceeds the threshold you'll get 9dB of gain reduction.  The ratio is an important aspect in how hard you will be compressing the signal.

Ratio.jpg

Ratios to know

When you are dialling compressors it can be be useful to have certain benchmark values for different uses.  Here are a few suggestions:

Ratio Application How Transparent?
1:1 to 2:1 Mastering and Bus Compression Very Transparent
2:1 to 5:1 Vocals, Guitars, Bass, Drum Bus Somewhat Transparent
5:1 to 10:1 Drums, Hip Hop Vocals, EDM Leads Not Transparent
10:1 + Catching peaks, Limiting Not Transparent

*These ratios are recommendations only.  Remember that the other parameters of the compressor work together to achieve a certain effect.

Peak Mode and RMS Mode

Peak and RMS mode selectors are common on software compressors.  Back in the days of hardware you didn't have a choice, each compressor's mode was inherent to its circuitry.  Peak or RMS mode refers to how the compressor recognizes the input signal.  When the compressor is in Peak mode the input level is determined by the peak volume of the input signal i.e. the loudest part of the transient.  Therefore the peak of the signal triggers the compressor.  Peak mode is typically considered to be very fast and very sensitive.  RMS (root mean square) mode means the compressor responds to the RMS input level.  You can think about RMS as sort of an average.  In RMS mode the compressor is often thought of as being being less sensitive and better for smoothing things out and raising the average volume.

Makeup and Output Gain

Compressors turn things down, the makeup turn things back up.  Let's say you set up a compressor on your drum bus to catch stray peaks and you get 4dB of gain reduction.  The makeup gain, if enabled, will increase the output signal by 4dB to bring the signal back to the "same" peak level.  In Ableton (and in most VSTs) the makeup gain on Compressor is turned on by default.  I would advise disabling it and setting the output gain yourself.  Here's why, louder volumes often seem to sound better than quieter volumes.  If you leave the makeup gain on as soon you slap the compressor on your channel your track will sound louder which may trick you into thinking it sounds better.  The only way to know if the compression you applied actually sounds good is to A/B it to the uncompressed signal at the same volume.  This means setting the output gain yourself.  If you decide the track sounds better compressed now you can turn the track up with the output if you have the headroom.

Final Thoughts

One questions to ask yourself as you reach for the compressor is, "What am I trying to achieve with this compressor?" Is it to catch peaks?  Is to smooth out a vocal performance?  Is it to get aggressive pumping? Do you want to accentuate the transient of your drum sounds? Sometimes compressors are important to a sound and other time they are not necessary and even hurt the sound.  You have to decided if the compressor is helping or hurting and the best way to do that is to A/B the compressed and uncompressed signals together at the same volume.  

To recap, compressions take loud parts of an audio signal and turns them down.  The decrease in volume by a compressor is called gain reduction.  Gain reduction occurs when the input signal exceeds the threshold.  After the threshold is crossed by the input signal it take a certain amount of time to achieve the gain reduction, this is attack.  Once the gain reduction is achieved it take time for the compressor to relax, this is called release.  Ratio controls to what degree the signal, after crossing the threshold, is turned down.  Using the makeup gain allows you to turn up the compressed signal to compensate for the gain reduction.  Different compressor modes respond in different ways to the input signal.  Peak mode is often considered less transparent, works quickly and is extremely sensitive to transients.  RMS mode is more transparent, works more slowly and responds to the average volume of the input signal. 

Join us again next week for Part II of our series on compression and learn how you can use compression on your drums.

More more information on compression and how to use it check out our Mixing and Mastering class taught by Tim Allan.

If you liked this article consider sharing online and registering for our email newsletter.  We send exclusive content, tip, tricks and information on special offers.  We promise not to inundate your inbox or share your email with others.