posted 09 October 2001 16:21            

compresstest.kym

================================

The compressor in Kyma behaves like other compressors -- with one addition.

As you know, a compressor measures the input amplitude in dB, computes how many dB above the threshold the input is, divides by the ratio parameter, and attenuates by the difference between this number and the amount the input is above the threshold. When the input amplitude is less than the threshold, the compressor does nothing.

For example,

input = -20 dB
thresh = -40 dB
ratio = 20

(input - thresh) - ( (input - thresh) / ratio )

(-20 - -40) - ( (-20 - -40 ) / 20 )
20 - ( 20 / 20 )
20 - 1
19

so, the -20 dB input is attenuated by 19 dB to give -39 dB. Another way to look at it is that the 20 dB above the threshold has been compressed by a factor of 20 to be 1 dB.

Using this method, when the input level is the same as the threshold, the output is unchanged (just the input copied), regardless of the ratio.

If you put in 0 dB (max input), you will get out thresh * ( 1 - (1 / ratio) ). (In this example, that would be -38 dB). This means that the larger the ratio, the closer to the threshold the output will be.

If you listen to the output of a compressor, the lower the threshold, the softer the output, because the dynamic range has been compressed to a smaller range. In other words, there is less difference between the highest and the lowest amplitudes. This narrower dynamic range allows you to add gain to the compressed signal, making the overall effect "louder".

Since you almost always want to add gain after compression, in the Kyma compressor we have built in a gain. This gain is set so that the fixed value is not the threshold, but the maximum possible input (0 dB in causes 0 dB out) with an additional attenuation of 10 dB (to allow for errors between the amplitude measurement and the actual program material). In addition, the Gain parameter can be used to provide additional gain or attenuation.