XILS DeeS: De-Esser Plugin for Vocals by XILS Lab Free Download

Once the signal enters DeeS, it is split into two independent channels: the voice with sibilants removed, and the sibilants only. The detection module lets you choose the cutoff frequency between voice and sibilants, the amount of reduction, and the quality of the processing, from smooth and slow to rough and fast. A plosive detection option helps the plugin distinguish sibilants from plosive bursts.

Each channel has its own EQ. The sibilant path offers Low Pass 6 dB, Low Pass 12 dB, Bell and High Shelf filter types, with adjustable cutoff, resonance and gain, plus a level knob to reduce or increase the sibilant path. The voice path uses a similar EQ without a level control. The developer describes the EQs as analog-style zero-delay filters designed to avoid latency artifacts.

The interface uses a shared color code across knobs and graphs, so the voice and sibilant channels are easy to tell apart. The real-time frequency graph displays both spectra and the applied EQ curves at the same time, which helps you see what the processing is doing to the track.

XILS Lab says the side-chain detection is independent of input level, so the plugin should respond consistently when the incoming gain changes, rather than over- or under-compressing as many de-essers can. The developer also notes that the detection works with or without latency, depending on the side-chain setting.