Photo: Doug Eisengrein
Limiting, also referred to as compression, is one of the most oft-used (and misused) ingredients in a recording engineer's cookbook. That's interesting because as compared with many other recording and mixing tools, limiters (whether software or hardware) often host only a few controls. However, because the relationships between the settings on these seemingly simple beasts heavily weigh in on the resulting sound, and because that result can be very subtle compared to other processing tools such as EQ, knowledge of what a limiter is designed to do and how its controls actually work is imperative to successful use.
Broadly speaking, a limiter (or its longhand term, a limiting amplifier) is designed to limit an incoming signal's dynamic range, generally while raising the resulting signal's output level (or perceived level). The most common uses of limiters are: taming the loudest peaks to avoid distortion in the following audio stages and therefore the resulting signal (known as “peak limiting”); performing relatively smooth, uniform level/gain control (leading to another name for limiters, “leveling amplifiers”); and in radio, Internet and other broadcast audio, for radically raising the overall perceived level to compete with other stations that do the same over less-than-stellar sound systems. More specialized, less common uses for limiting that use a qualified limiter's sidechain circuitry are de-essing and ducking. De-essing is a technique for reducing excessive ess, ch and other similar consonant sounds from vocal tracks, while the ducking involves triggering the limiter to act on one audio source via an entirely different source (such as limiting a vocal track with a guitar track), generally to save both sources from competing for “sonic space.”
The answer to a common question — “Is there a difference between a compressor and a limiter?” — is yes, albeit a small difference. While the two terms are often used interchangeably, limiters generally have higher ratio settings than compressors, and they often sport fast attack time settings (without a variable control). Small differences aside, here I focus on describing the various controls of compressor/limiters, without delving much into the different scenarios in which they are applied. There are five most common controls on limiters: threshold, ratio, attack, release and output gain, sometimes makeup gain. Some limiters, notably higher end units and classics such as Manley Laboratories Variable-Mu Limiter Compressor (www.manleylabs.com) or the Teletronix LA-2A (hardware and plug-in versions at www.uaudio.com) are absent of one or more of those controls; instead, they favor predefined circuits or algorithms. Their sparse front panels should never be mistaken for marks of lower quality products. A limiter's price tag will usually reveal volumes about its ability and sound quality.
The threshold control designates at what volume level an incoming signal will start to be limited. If you're dealing with a lower level source signal, you'll generally use a lower threshold. Conversely, if your source is already at optimum gain and you still use a lower threshold, more of the signal will be compressed. For peak limiting, you will generally use a high threshold setting, whereas for leveling you will often use a lower threshold.
Bypassing some of the mathematical jargon, the ratio designates the degree (percentage) of compression taking place. A 2:1 setting is lower than an 8:1 setting, which is lower than a 10:1 setting, and so on. The higher the setting, the more heavily compressed (or reduced in dynamic range) the signal that reaches above the threshold will become. Herein lies the reason for output gain: If a certain portion of a signal's dynamics is reduced, raising the output gain brings the final, limited signal back up to optimum gain. In the simplest terms, that means the resulting overall volume is greater, since the higher level portions have been leveled (“limited”), just as the lower level portions have been amplified.
Attack and release controls are arguably the least understood of the bunch. That may be due to their relatively subtle effects (compared to threshold and ratio, at least), and also because they are time based (in milliseconds and seconds). For example, if a limiter has a variable Attack range of “1 to 100,” it's easy to think of “100” as “fast” because it's the highest number and at the farthest clockwise setting. But the opposite is true: 100 ms is a very slow attack time, meaning that limiting will kick in 100 ms later; for some purposes, such as peak limiting, that's way too slow.
Slower attack times, however, can present a smoother, less compressed-sounding result. Variable-release settings typically range from about 20 ms to 2 seconds, 5 seconds or more. Finding suitable release settings may be the trickiest of all because if it is set either too fast or too slow for the program material, audible distortion or obvious “pumping” can occur — that is, an audible jump from the compressed to the noncompressed portions of the signal. Generally, you'll want to find a balance where you limit exactly the portions you need to, without making the overall track sound unnatural. Only your ears can tell you where those magic settings will be; don't rely strictly on any “rules” you've heard. Learning not only about the proper use of limiters, but also about the sonic character of the particular limiters you have on hand, are lessons learned best on the job. For a detailed look at the more specialized de-essing and ducking uses of limiters, refer to the Phantom Power article “Do the Dip” in the August 2007 issue of Remix (Remixmag.com/issue_20070801).