Audio compression is used to reduce higher signal parts above a certain threshold while the signal below the threshold remains un-effected. with a good compression music sound more compact and louder. With compression used for speech distortion can be avoided easily by diminishing loud sections or loud voices in total.

control circuit level compression above threshold
(center and right)
compression ratios

To avoid a too sharp and audible changing the knee of the ratio curve (see above) can be set as a 'rounded' knee. This is common in more sophisticated compressor designs.

The threshold is the level from which a level reduction will be executed. Threshold is set in dB, counted from full level downwards. With a threshold of -20dB level reduction starts at -20dB input signal, with a threshold of -40dB level reduction starts at -40dB input signal.

The ratio defines the ratio between output and input signal (above threshold). A ratio of 1:10 means and output of 1dB for an input of 10dB (always only above threshold).

The attack time is the time until the level eduction will be activated. A higher attack time let music sound more lively, there as a too short attack time lets music sound dull and dead. A good start is 50 msec.

The release time is the time until the control goes back to an uncompressed state. Also the release time should not be set too short.

uncompressed signal and compressed signal 2:1, 4:1, 10:1 (signal below theshold is un-effected!)

Rolling Stones
Time is on my Side, 1964
Norah Jones
Sunrise, 2004
highly compressed
Metallica, My Apocalypse, 2008