How to Record Using Compression

Compression is used regularly in the sound editing process. A piece of music can go through multiple compressions before it is completed, because compression can be used at multiple points during the creation and revision process, from recording to mastering. Essentially, the tool is a variable gain device. The input determines the level of gain.

Compression works by adjusting the sound volume; it quiets loud sections and makes quiet ones seem louder. Quiet sections do not actually increase in volume, but appear to do so as the result of a secondary gain stage that occurs during compression. Essentially, it diminishes the range of a sound. In doing so, compression can perform several functions including accentuating specific elements of sounds, reducing distortion, and making the sound seem louder. How much compression to use depends on the exact sound desired; it is not an essential part of the editing process, and may not even be used for every single piece being done.

A threshold determines the required volume of a signal before volume decreases. A high threshold means less compression can occur, because there will not be as much signal exceeding the threshold.  Low thresholds work in the opposite way. A compressor detects sound levels by averaging input time. A compressor responds to a change in input levels. When these levels exceed the sound threshold is called attack time. 

Attack time tends to be short in duration, lasting well under a second. Release works the opposite way of attack; it occurs when the signal falls below the sound threshold. Attack and release work together; they also work with threshold to determine the beginning and ending of compression. Attack setting occurs at the beginning of the track, and release at the end. Because attack and release are so closely linked, if an input signal is less than the time of attack, the amount of compression will be reduced. Setting a long attack means that there is less risk of over-compressing a track. However, if an attack is set too long, compression will not occur.  Compression length is determined by release.

Another form of control, ratio, controls the gain of a signal if it goes above the threshold. If a signal ratio is set 4:1 and there is a signal increase of 4 dB, the compression will result in an output increase of 1 dB. During compression, gain can decrease if the signal is high, so it is best to make sure to set gain levels to compensate for this.   

To decrease dynamic range, the volume of loud sections is decreased and then the volume of the entire sound is increased. This keeps the volume ratios between loud and quiet parts remain the same, although the quiet parts of the song have been made louder. The overall volume stays the same, but the average volume has been increased.  When beginning the compression on an individual track, it is helpful to begin with a low ratio, high threshold, and mid-range attack and release. Attack and release should also relate to the speed of the instrument, voice, or other sound. On the other hand, if compressing a mix, these adjustments should be varied based on personal artistic preference.

In addition to reducing dynamics, compression can also be used for other aspects of sound editing. It can help fix poor sound quality caused by a poor recording. Compression can be useful not just for adjusting sound, but for more artistic purposes as well to create interesting special effects. Finally, compression does not have to be used just to decrease dynamics; it can increase them as well. This can be accomplished by compressing with a long attack. This can help a specific sound stand out. These effects can be created by using a feature called the side-chain, a tool used specifically to control a compressor. 

There is an extreme form of compression known as limiting. During limiting, the signal ratio is flattened, and the ratio is altered; an example of a limited signal ratio is 10:1. While both limiting and regular compressing have their benefits, there is some loss of overall sound because both processes can create unusual sounds or decrease musical accents. Both are useful, but one should be careful not to over-compress a piece of music.