Compressor - Lexicon 300L - REV Owner's Manual

The Algorithms and their Parameters

Compressor

The Compressor algorithm is a true digital compressor which will run in either General Description
Dual Mono or Cascade Setups. In Dual Mono Setups, it configures to mono in
mono out; in Cascade Setups it configures as a true stereo effect. The
compressor can be described as an upwards averaging compressor. Digital
compressors, like analog compressors, decrease audio above a given thresh-
old. Unlike analog compressors, they increase gain below the threshold. The
result from either analog or digital compression is exactly the same — less
dynamic range .
As shown in the following diagram, the audio path takes two routes. One path
goes through a predelay mechanism which delays the audio a maximum of 48
ms. The other path sends control information (dependent on the settings of
slope, threshold, attack and release) to "digital VCA's" labeled MAX COMP
GAIN and EXP GAIN. As the signal crosses the threshold point, both Compres-
sor Gain and Expansion Gain vary constantly. Compressor Gain determines the
maximum amount of gain increase below the threshold; Expansion Gain
determines the maximum amount of gain attenuation below the threshold. Note,
there are separate thresholds for compression and expansion.
Generally, digital compression requires lower threshold settings than analog
compression. The reason for this is that there is no such thing as headroom in
a digital system — dBfs (Digital Full scale [0VU]) is the maximum level audio.
Audio dynamics below full scale, however, can be manipulated and modified.
The whole purpose of a digital compressor is to maintain peaks while compress-
ing lower level audio signals upwards. If you think about it in terms of the most
significant and least significant bits, you would never want to reduce the most
significant bits — you want to increase the least sigificant bits.
Adding some pre-delay gives the control mechanism time to react before the
audio reaches the digital VCA. Of course, the more predelay you add, the more
"out of sync" the audio will become. Attack constants should be kept to short
values (7,15, or 30ms). A good starting point for release time is 91 or 114 ms.
4-35