Basic Audio 101: Audio Compression and Audio Compressors
Compression is a very common audio process and yet many people do not fully comprehend how or when to implement it. Audio compression is used to reduce the overall dynamic range of an audio signal. Compression may be used in the recording process, live sound reinforcement, and even when playing back or broadcasting an audio source. Simply put, an audio compressor is some type of device (physical or virtual) that applies the “compression”.
In order to understand audio compression a basic knowledge of “compressor terminology” must be understood:
The four most basic terms used in conjunction with compressors are Threshold, Ratio, Attack, and Release:
Threshold: With regard to audio compression, the threshold is the user-defined (or set) level (usually in dB) at which the audio signal will begin to be compressed. A lower threshold hold will result in more of the overall signal being compressed while a higher threshold will result in less of the overall signal being compressed. Audio below the threshold will not be compressed; audio above the threshold will be compressed at the set ratio.
Ratio: The ratio determines at what rate the audio ABOVE the threshold will be reduced. Ratio options commonly include 2:1, 3:1, 4:1, 8:1, 16:1 and greater and are spoken as two to one, three to one, etc. The ratios work in this matter: a ratio of 4:1 ratio means that a signal that is 4dB above the set threshold will leave the compressor 1dB above the threshold, at the 8:1 setting this ratio would set the same 4dB above the threshold at .5 dB, etc. The higher the ratio the more compression that is applied to the audio above the threshold.
Attack: This setting determines the “reaction time” the compressor will utilize to apply the compression from the point the Threshold is crossed. The attack time is usually given in milliseconds, typical ranges go from 1 millisecond to 100 milliseconds or more.
Release: This setting determines the time in which the “compressed” sound will return back to normal gain, or volume settings after the threshold is crossed in the reverse direction. The release time is usually given in milliseconds to seconds and typical ranges are from 25 milliseconds to 4 or more seconds.
Another common feature of compressors is a “soft-knee” function. Engaging the soft-knee feature of a compressor typically results in a smoother change from the uncompressed to compressed audio, especially when the compression settings are set high.
Common Examples of Using Compression for Live Sound:
Bass Guitar: Used in the typical fashion, the bass guitar has a very wide dynamic range and many times it may “walk” over top of other instruments. Because of this the Bass Guitar is typically compressed and many bass amplifiers have a built-in compressor. Compression settings vary depending on the type of player and genre of music but in general, a “light” compression is a good staring place.
A typical compression setup might be to set the thresholds at -5 to -10 dB, an attack of 20-60 ms, release time of 200-500 ms with a ratio of 3:1 or 4:1. If the compressor has a soft-knee function, using it may provide a more natural sound, especially if the bass is acoustic.
Vocals: Because vocals are very transient nature, a good knowledge of overall audio must be understood to make the best of vocal compression.
While I typically will use compression after I’ve applied equalization to control for any mild volume changes that it may/might have caused, a good argument can be made that pre-compression equalization is less responsive than post-compression equalization. The truth is that both applications are used.
One of the typical issues with a vocalist is a wide and varying range of dynamics often due to the varying nature of the microphone from the singers mouth. Depending on the mic technique and experience of the singer you may need to more or less aggressive in the settings. A typical setup for an inexperienced vocalist might be to set the threshold between 4:1 and 8:1, set all attacks as fast or almost as fast as possible, a release time of 300-600 milliseconds and a threshold of -4 to -8 db.
It is important to note that vocal compression works very well in a studio setting but presents additional feedback issues in “live” settings and the audio engineer must be particularly aware of these issues when working with live sound. The reason for this phenonmea is that compression works on reducing peak levels and by applying compression peeks are reduced but not the “lower” volume signals.
Typically, additional gain (or make-up gain) is added to “make-up” for the overall loss of peek volume and by doing this -it will reduce your gain-before-feedback margins. For example, if you have only 9dB of gain (or headroom) before feedback in your system prior to applying compression and you then apply 9dB of compression, you would typically add additional gain (maybe 6-9dB of make-up gain) to bring the overall peak volumes back to where they were. The issue with this scenario is that the non-peaked levels are now brought up another 6-9dB as well and are now surfing the boarder of feedback because the gain-before-feedback margin has been reduced by the 6-9 dB of the additional make-up gain.
Typically in a live setting you would not want to add any more than 3dB-6dB of compression for these reasons and minimize any chance of feedback that might be caused because of it. Make sure that you have the overall headroom to add the compression if you are going to use it in a live setting, otherwise the results may be disastrous and feedback may be rampant in your audio system.
Common Example of Adding Compression to the Overall Audio mix:
Many people commonly apply a compressor to “reduce the peaks” of an overall mix after a song has been recorded to two-track (stereo mix) or during the mastering phase. It is important to note that any individual track or input compression and all audio leveling should be accomplished before you attempt to compress the overall 2-track or stereo mix as any compression applied to the stereo mix will have very little effect on individual audio components or produce effects that are unwanted. This type of compression is typically used to reduce the overall dynamics of the song or track.
By reducing the extreme “peaks” of a stereo mix you create more headroom and if you choose to normalize the audio for overall maximum playback volume (useful if you are broadcasting the audio and/or trying to get consistent levels across songs/tracks) will create a more even track volume (albeit sacrificing some of the lower dynamic levels).
Whether you are recording, editing, or dealing with live sound reinforcement, proper use of compression can and will make your sound and mixes more coherent and provide you with a more consistent level of sound and blends. There are no “cookie cutter” settings for compression and the best results are obtained by sound engineers and musicians that have spent much time trying various compression settings in various situations.
In the end, if it sounds good -implement it, if it does not -try another combination. How do you learn to determine what sounds good? Spend time actually LISTENING with an active “ear” to established artists and good recordings. Experiment with your gear and equipment and realize that good recordings and performances start with good musicians, good audio gear (starting with the mic and mic-preamp) and proper placement and/or recording techniques. If anything in the beginning of that chain is sub-standard, YOU WILL NOT BE ABLE TO FIX THE AUDIO IN THE MIX LATER!