Since we're already this far off topic...Until I bought my Alesis MasterLink, which performs both compression and peak-limiting, I probably would've described peak-limiting the same way you have. Apparently, I hadn't thought about it a lot in relation to compression, technically speaking. The compressor has all these adjustable parameters:
Threshold (level relative to full-scale 0dBfs the compressor begins to affect the audio)
Ratio (of input level to output level of the compressed audio)
Make-up gain (applied after the compressor to compensate for the level lost in the compression process)
Attack (time the compressor takes to begin affecting the audio after it's risen above the threshold)
Release (time the compressor takes to stop compressing the audio after it's fallen below the threshold)
Knee (controls the way the compressor behaves around the threshold by varying how quickly it ramps up to the full selected compression ratio -- a "hard" knee applies the full selected ratio at the selected threshold level, while "softer" knees progressively take effect beginning a selected number of dB's below the selected threshold level, increasing the applied ratio until at some level *above* the threshold it is fully equal to the selected ratio)
There are also controls determining whether the compressor is keyed to peak or RMS input levels, of the left or right or both channels, plus for what the meters measure. Fancier compressors than this one can have multiple frequency bands of compression per channel, each with independently variable parameters.
The peak-limiter in the MasterLink doesn't have all these adjustable parameters (just Threshold and Release, plus an adjustable Output level control, for if you want to preset the maximum output at something less than 0dBfs; make-up gain is applied automatically, equally and in opposition to the selected threshold level), so it's simpler to use. And because it's all-digital, it's able to "look ahead" and begin progressively reducing gain in advance of a peak exceeding the selected threshold (similar to the Knee control, but fixed), for a smoother dynamic compared to traditional analog peak-limiters. But conceptually speaking the two functions involve all the same processing.
According to Alesis, a traditional peak-limiter "is typically thought of as a compressor with a high ratio setting". Due to the look-ahead feature of theirs, they say it functions with essentially an infinity:1 gain-reduction ratio. (I presume the attack time must also be fixed at zero and the compressor keyed to peak level for both channels.) In other words, any signal exceeding the selected threshold level at the input is held to that threshold level at the output -- and that threshold level could be either high, low, or anywhere in between. So peak-limiting is compression with less potentially varied, more narrowly defined behavioral parameters.
Csmithbarc: Sorry man, I'm done! :-)
