I noticed that an interesting question came up which lead me to think back on my days as a student. The original question was how to use compressors in orchestral music. As I was thinking on how I could explain this as easy as possible I realized that we needed to go back to the basics first. What is a compressor and what does it do?
Why another article?
There has been written, probably, thousands of articles on this before so why would I make another one? Well, I have had so many lectures on this but only one that stuck with me. A compressor can be explained so easily, still so many dive straight into the hard core mathematics of it. This is meant to be an easy guide for beginners.
I remember we talked about this in one of the first lectures at the Notoff Institute. My tutor had a very simplifying example, and after many years I keep coming back to this example.
Imagine you are mixing a song and you have a couple of tracks. You have recorded vocals for the track and while you listen you notice that the vocalist sings very loud on some parts and rather soft on other parts. You would like to adjust the volume for the vocals only so that it is more even. You could use automation (read/write) the fader so that the volume gets averaged out. But this is time consuming and it is hard to get the levels precise and consistent. So instead of doing all this manually we use a compressor.
The fader robot
The easiest way to explain a compressor is to compare it to a little robot that sits at the mixer desk and adjusts the faders. Whenever the vocals get loud, he turns the volume down with the fader. There are many different compressors available. If we are talking about plugins, they usually consist of the same four components. Lets go back to the example of a robot controlling the faders on the mixer. Since he is a robot you will have to tell him what to do and give him instructions. These instructions will have to include when the robot should turn down the volume, how much and so on. On a conventional compressor you will have to do the same. The four basic elements (the four basic instructions) of a compressor are:
The four elements of a dynamic range compressor
When it should reduce the volume
If the vocals get loud enough the compressor will be activated and reduce the volume. The threshold defines at what point the compressor is activated. In other words, how loud the track has to get before the volume is reduced. Usually this is measured in dBFS where 0 (zero) is maximum possible level.
How much it should reduce the volume
When the track gets loud, above the threshold, the volume has to be reduced. The ratio will define how much the volume is reduced starting from the minimum level 1. The ratio often goes from 1 to 8 or from 1 to 5 depending on the plug-in used.
How quickly it should reduce the volume
When the threshold is reached and the compressor needs to know quick it should turn the volume down. Should it do this instantly? Should it do it quickly? Or should it do it slowly? The Attack defines how long it should take from the audio reached the threshold to when it is reduced back to its original state. If you think about the robot, it’s sort of like the robots reflex response time. The amount of time is measured in milliseconds.
How long before it should restore the volume back
When the audio signal goes back below the threshold, it is time to release the compression. The release function is how long it will take for the compressor to turn the level back to it’s original state. In other words how long it should take before the volume is restored back to where is was again.
Other elements in modern compressors:
In addition to these four elements many compressors have other options. These are modern additions to the compressor. In other words these elements are usually not in hardware, old compressors:
Make-Up, Auto or Auto Gain
When a compressor is used the highest peaks (the loudest parts) of the audio signal will be reduced. This will cause the overall loudness to be reduced a bit. A makeup gain function will allow you to add some overall gain to the signal so it will not become too soft. It is important to understand that this function is a post-process function. In other words, the gain is applied only after compression, so it is not going to affect any of the compression settings.
The hold function will delay the time before the release is activated. This allows the compressor to “compress for a while longer” even if the original audio signal is below the threshold again. The hold is also often measured in milliseconds.
Soft-knee or envelope
You would not actually want to have a robot sitting at the mixer desk, would you? If a robot were to turn down the fader every time the audio signal got too loud, he would probably have very mechanic movements. Just like robotic dance moves. When mixing, it is quite important that adjustments done on the faders are gradual and natural sounding. Compared to a robot a human would maybe turn down the fader gradually at the beginning, then faster and then a little slow again depending on the feeling. A soft-knee is a great way to emulate this “human behavior” or “humanize” the compressor a bit. It will cause the compressor to reduce the volume at an exponentially. A normal compressor ratio curve is often refereed to as “hard knee”:
Cubase has an additional option called Analysis. This is probably a bit too complicated for this article but, I will explain it briefly. When the sound signal is coming in as continuous sound the compressor will measure the peaks (the loudest parts). Here you have the option to measure the average instead (or the RMS, root mean square). The compressor will “look ahead” for audio signal and calculate the average level of the signal. Measuring RMS is great for vocals with few transients but peak might be more suitable for percussive instruments such as drums.