Effects and Signal Processors - De-esser

The word "de-essing" is literally what its name suggests: the elimination of that annoying whirring noise you hear in certain vocal recordings when certain letters containing higher frequencies such as "s" are pronounced. The whirring depends on the fact that the signal is in that moment saturated by the letter s's frequency, thus generating a distortion. Ones immediate response would be to equalize the disturbing frequency in question to resolve the problem. This solution however, it's not possible on a practical level, because by doing so you'd modify the entire frequency-content of the whole recording, and ruin it completely and irredeemably. To bring about a correct de-essing process we need to use a compressor combined with an equalizer, as shown in the following diagram:

Diagram of a De-esser module

This is how it works: the original signal is passed through an equalizer in which the s's are highlighted more than usual, whilst all the other frequencies are attenuated to their lowest.

Equalizing of the input signal

The signal which exits the equalizer has a relevant amplitude only when s's are present. This signal is sent to sidechain input of the compressor, thus activating it only in relation to s's'. Therefore every time an s is present, the signal coming out of the equalizer goes beyond the threshold set on the compressor which thus activates itself and lowers the volume of the voice, avoiding saturation. Once the s has passed, the voice's volume returns to its original volume.