ICEinLAVA's Domain

Have we forgotten how to cook?

Has the microwave generation taken hold in the sound industry? What would we do without a preset for this and automatic function for that? Dynamics have often been a mystery, much to do with the abstract ways product designers have given the sound- mixing community processors, controls and metering functions. In addition, many devices perform multiple dynamic processes simultaneously; others perform one at a time and are dependent upon set-up in order to determine their function. In the following series of columns, we’ll provide some basic guidelines on the theory and operation of the most popular dynamic functions, basic control-setting used to achieve a processing type and how to apply them to our sound-mixing tasks.

In audio, we deal in two basic domains: frequency and level. At our disposal are many tools to help tame or contour these realms. Equalization helps us with the frequency area, by statically adjusting a range of audio bandwidth. Dynamics allow us to deal with the changing of levels. There is a point where these two processes converge: dynamic EQ; and we will cover that in the “Special Devices” section at the end of this series. This month, the focus will be on the two primary dynamic functions: compression and limiting. It’s not uncommon to use these terms interchangeably, but by article’s end, you will know the precise difference and how and where you would choose to apply the process.

With all dynamic processors, there are some common terms and not-so-common terms used to describe the operation. Of the common ones, THRESHOLD, RATIO, ATTACK and RELEASE describe the four principal operational parameters of many dynamic processors. Some of these are internally fixed, but still there. Threshold describes at what level the process occurs, either above or below; ratio describes how much processing takes place once the threshold has been reached; and attack and release determine the speed of the envelope. Additionally, you might encounter HOLD (another envelope function, the transition point between attack and release), CREST FACTOR (one way of describing the action of ratio as the dynamic transitions from below threshold to above) and KNEE (in the soft or hard variety; these are other ways of describing how the transition from under to over threshold occurs). Other common controls found are: GAIN (the make-up after the dynamic process); PROCESS (generally a threshold control, sometime a combination threshold, ratio; and COMPRESSION (again a combination of threshold and ratio).

Compression

As the term implies, compression makes a reduction in the total dynamic range. Often, this is required in order to get a given audio signal to “fit” within a smaller performance range. Case in point: take the average vocalist. A pop/rock singer could have a whisper of 15dB and a scream of 90dB. Some operatic vocalists are measured to produce 120 dB at one meter! This dynamic range couldn’t fit in a performance environment where the average noise level is 60dB. The bottom 45dB would be inaudible.

A compressor takes the soft parts of the performance, and brings them up. Now, I can hear you saying, no way, it brings the loud stuff down, and you’d be partially right, if it were not for something called make-up gain. A compressor works by taking away gain when the signal is at its loudest (referred to as gain reduction), and gives it back when the signal falls below the set threshold. It is this basic action of giving back the gain below threshold that creates the action of bringing up the soft parts of the performance, but there are other factors too.

The setting for the dynamic processor needs to tell the device it is to act as a compressor. For example, the threshold is best adjusted to a low setting, say -10 to -15 dB. Low ratio settings of 2:1 to 6:1 are to be used. Lastly, moderate attack and release times are employed. The combination of all of these settings ensure that the signal is processed gently by the ratio and attack and release settings, so as not to distort the overall envelope of the program.

At this point, the operator should meter the equivalent amount of gain reduction for the loudest parts that he wants to have “brought up” during the quiet ones. This probably means that for most of the performance, the mixer will be metering some amount of gain reduction; this is normal.

It is important to remember that the ratio and attack/release are set to ensure the integrity of the program envelope. If you hear artifacts at this point, lower the ratio or slow down the attack and release.

This, by the way, is where many go wrong. I often hear engineers claiming, “I can’t stand the sound of the comp,” but invariably, the setting for attack, and particularly the release time, are set far too fast for the job at hand. Different instruments and performers have different requirements. Vocals, with their medium wavelengths and fast dynamics, can tolerate faster attack and release times compared to bass guitar and low-synth sounds. Case in point, compressing a bass guitar is a frequently performed task. Consider that the lowest frequency on a normal bass (the low E string) is a 45 Hz wave, 40 Hz for a dropped D, and about 30 Hz for a Low B on a 5 or 6 stringer. If you wish to select an attack and release time for these wave forms which won’t distort the waveform, you will need to choose a time constant that is at least 1/4 wavelength of the lowest note for the attack and perhaps .75 to 2x the wave on the release, again source dependent. It is very important to try different settings, (if your compressor allows it) to adapt it to the instrument at hand. Keep in mind, faster is almost always more destructive.

Other operational tips: You will notice an interaction between ratio and threshold. For example, lowering the threshold will result in higher amounts of gain reduction much as increasing the ratio will. Think of the amount you want to compress the signal, say 10 dB, set your attack and release controls to the middle settings and then adjust the threshold and ratio to achieve 10dB of gain reduction; always judged by the loudest part of the performance.

There are primarily two types of compressors, those with variable thresholds such as a dbx 160 or Aphex 662 and should be adjusted using the methods described above. Those with fixed such as UREI LA-2 or a Summit TLA-50 are generally easier to set as the threshold and ratio controls are set internally. As you adjust the “process” or “peak reduction” knob, you are adjusting the input gain into a fixed threshold, then, you adjust the output gain for unity. In these vintage units, attack and release are determined by an optical circuit and in some cases you can modify these slightly with a switch selection. Since the nature of the optical device is slowish, the effect is often musically pleasing on some instruments and vocals, largely due to the lack of intrusive action. These “boutique” devices are also chosen for the unique “signature” they impart upon the sound.

Limiting

As compression is used to define the lower part or the “floor” of the audible dynamic, limiting is used to define the ceiling. Therefore, the requirement for the settings is different. The threshold is set high; perhaps 0 to +10 dB and the ratio is set high as well – 10:1 to 50:1 or higher. Attack and release times are set as fast as necessary to catch the dynamic; always balancing peak overshoot and sonic artifacts that come from the fast action. Limiters come in various forms as well: peak, clipper, program, etc. These differences describe the application and performance to a certain degree.

Peak limiters should be used on fast transients like percussion, piano and perhaps vocals while clip limiters are used as a final “brick-wall” to prevent digital overload. Program limiters are used when fully mixed audio is being processed for recording and television or radio broadcasts.

Next month we’ll get down and dirty with gating, expanding and other forms of advanced processing like leveling, De-essing Dynamic EQ and more!