The compressor is one of the mainstays of modern recording and mixing, being a versatile device that can be used as a dynamic and in a sense as a time domain effect on a signal, at the recording or post production stage. There is simply no way that a “produced” sound can be achieved in a mix without one of these devices, their correct use being a major part of the black art of creating commercial sounding products.
The earliest reference I have found to the use of a compressor is in 1934, when the device was employed by the Bell Telephone Laboratories to reduce the dynamic range of an electrical signal, in this case the human voice, to stop the signal exceeding the dynamic range of the medium that it was broadcast through. Compressors continued to be used for this purpose in recording sound tracks for films, and in recorded music, to squish the signal onto recording mediums with limited dynamic range. The dynamic range of an orchestra, for example is massive, exceeding a hundred decibels. The dynamic range of vinyl records is not, being about seventy five decibels. It stands to reason that some of the orchestra's dynamic range is going to have to be sacrificed.
In order to have the quiet parts of the signal recorded at a reasonable listening level and fit the loud parts of the signal into the dynamic range of the recording device without clipping or distorting, a threshold is set using the compressor, with zero decibels representing the maximum signal gain and the threshold set as a negative number. Let's say the threshold is set at minus fifteen decibels for example. Any signal louder than minus fifteen decibels being attenuated according to a user defined ratio that is set with the ratio control on the compressor. The crushed dynamics at the loud end of the range can then be offset by a control called the output compensation gain, which makes the entire signal louder, or quieter if so desired.
The use of the compressor has evolved dramatically since the days of early recordings, with compressors now being used as mastering limiters, or to attenuate signals in the manner described above, and as effects to smooth out the dynamics in recorded tracks, and even to create pumping and breathing effects on certain tracks, often bass and beats. It is these uses of the compressor that I wish to elaborate on in this short article, using easy to follow examples.
The simplest form of compression, as mentioned on the previous page, simply involved reducing the dynamic range of a signal to achieve a reduction in dynamic range. This is done by setting a short attack and long release, setting a threshold and a ratio. The magnitude of this reduction is controlled by lowering the threshold, and raising the ratio of the compression. The threshold controls at what level the compression kicks in, and the ratio controls the amount of reduction in the gain of any signal that exceeds the threshold value.
In order to pull a signal with uneven dynamics together, the idea is to use a threshold set very low, as low as -30 to -35dB, and a low ratio between 1.1:1 and 2:1. To make a signal louder, the way to go about it is to set a high threshold and a high ratio, and then crank up the output compensation gain to fill the extra headroom created.
These settings in the image above are fairly typical settings for squishing a signal to make it louder. The threshold is set at -20dB and the ratio at 4:1, shortest attack, longest release. It is important to be judicious when setting the threshold and ratio as over use of the squishing effect will cause side effects and make the track fatiguing to listen to.
Hard knee compression will kick in instantly as soon as a signal breaches the threshold, whereas soft knee compression is applied gradually as the threshold is approached. A hard knee can be useful in deliberately creating pumping sounds to make bass and beats punch. Soft knee compression is more subtle and therefore more suitable for compressing whole mixes or sound levels that hover around the set threshold point. In some compressor models and plugins, such as the one shown above, this setting is interchangeable.
Tightening Basslines and Beats
Before the 1960's, the received wisdom among sound engineers seemed to be that pumping and breathing should be avoided when using compressors. During the sixties, the roll of the compressor began to change, and this attitude was completely turned on its head. There is almost no modern mix that you'll hear without pumping used on the bass lines and beats.
To illustrate what this means, I have taken a screen shot of the compression used in one of my mixes.
The first thing that has to be done is that a threshold has to be set. This is the fraught part of the business. Too high a threshold and the compressor will not have much of an effect at all. Too low a threshold and the compressor will cause unwanted artifacts that will be audible in the mix. The level the threshold should be set at depends on the level of the signal passing through the compressor. In Behringer outboard compressors and the Waves Bundle compressor above, there are indicators to show how much attenuation or expansion is being allocated to the signal, this is on the red on dark gray scale on the left side of the picture above. These will show nothing until you have also set a ratio. For most applications it is best to make sure that only a few decibels are being shaved off the original signal. The ratio above has been set a five to one, but there was also compression added to this bass line with a Behringer Composer Pro at the recording stage. I personally feel that on bass lines there is an advantage of using analogue over digital compression in the form of a more natural feel on the compressed track. Theoretically, digital compressors can of course do things that analogue voltage based ones cannot, for example have an instantaneous attack and release or an infinitely hard knee.
The higher the ratio is set , the harder the signal above the threshold will be squished. Compressing bass lines at a high ratio can make the bass track sound smoother. Trust your ears and practice.
Now, as you can see from the diagram, the attack has been opened up here to five milliseconds and the release to fifty seven milliseconds. This was not by chance. The tempo of the song is one hundred and thirty beats per minute. In order to obtain a figure for the release to make it tempo based there is a simple equation.
Sixty thousand, divided by the number of beats per minute will give you the number of milliseconds per quarter beat in a recording with a four four time signature, provided the recording has been played to a metronome. Now, divide this number by two, again and again and again until it falls between fifty and one hundred. There you have the number to set with the compressor controls as your release. If the song has not been recorded to a click track it will be necessary to either take an average value, somewhere between fifty and a hundred, or get out a stop watch and make a rough calculation of how many BPM you have that way. Using certain tap delay plugins will give you an approximate idea of what kind of BPM you're working with as well, as you can tap along with the beats of the song and get a digital readout of the BPM. It should be possible to download tap to BPM tools from the NET as well, like this example from AnalogX.
Adjusting the attack from zero upwards will give you the amount of pumping that you desire. More milliseconds on the attack means more of a pumping sound, but watch out for warning lights on the trim at the top the channel strip, and be judicious. Overdoing this kind of effect can destroy a sound.
Finally, to take advantage of the extra headroom you've just gained, you can start to crank up the output compensation gain until the signal is coming in at an optimal level. If this is done properly on every track, there will be little or no need for huge fader movements at the mixing stage. If there are still odd parts of the signal that are clipping the trim and distorting, then a stone wall limiter plugin can be used to reign them in. Simply set the threshold of the mastering limiter to -0.5 decibels and this should be adequate to stop any peaks from clipping. If it is not, then there are serious problems with the gain structure of your track or your compressor settings that need to be looked at.
Practice the techniques described above, and it won't be long until pumping bass lines and beats that you hear in commercial recordings will be at your fingertips.
Stone Wall Limiters
A lot of limiters only have two settings, more or less. Some limiters like the one in Adobe Audition (formerly known as CoolEditPro) have settings for look ahead and release. Sometimes the release can be shortened to mitigate side effects where the sound starts to show obvious side effects and lose clarity and definition.
As well as dealing with entire mixes to bring a finalized stereo master up in volume to a commercial CD listening level without clipping the track, I've found limiting particularly useful on kick drum and bass guitar tracks where large output compensations are used that clip the trim. You will absolutely never hear a modern recording where limiting has not been used to increase the volume of the stereo mixdown at the mastering stage.
Companding is short for compressing expanding, and is a technique used in broadcasting and in digital recording which involves strategically reducing the dynamic range of loud parts of a signal and leaving the dynamics of the quiet parts, then broadcasting them across the airwaves as pulse code modulated signals then expanding them again at the receiving device. Paradoxically, this creates subjective impression that the dynamic range of the track is greater than in fact it is. This is apparent in NICAM television broadcasting (Near Instantaneous Companded Audio Multiplex), a form of Pulse Code Modulation where companding is used to effectively create the subjective feel of a fourteen bit dynamic range from a ten bit signal.
Copyright Graeme Young
Published with permission
Discuss this article in our Music Forum.
About Graeme Young
Graeme Young is a sound engineer based in the South of Scotland, starting his career after being presented with a Yamaha MD8 in the late 1990's by a friend who had despaired of figuring out how "the damned thing" worked. Brimming with enthusiasm as he learned his way around his new toy, he immediately set to work with gusto, creating some of the most abominable sound recordings of musical doodling ever committed to tape.
Nevertheless, the heady times of fun and friendship that were built up in the early days convinced Graeme to go back to college and expand his knowledge and skill set, meeting contacts and learning the tradesman's tricks from industry professionals. In the meantime student loans were spent on studio equipment to expand on the trusty MD8 while the O2R96 and the old StudioMaster at the college provided experience of working in both the analogue and digital domains.
Now, Graeme has gained years of industry experience working on a number of studio recordings and location recordings with professional musicians and in directing and editing video projects for professionals in various fields. Graeme has fronted two bands, the now discontinued Popping666Cherries alongside Gwen Smith, who had fifteen minutes of fame on the X-Factor before being told by Simon Cowell that she would "never do anything" and now his own Moonstruck Project, with a loose collaboration of musical friends.