Sonible smart:deess

What smart:deess is for

smart:deess is a vocal processing plugin designed to instantly enhance the intelligibility and smoothness of vocal takes, podcasts, voice-overs, or singing. It targets areas where consonants become intrusive (S, Z, Sh, Ch) and where certain attacks can "pop" on the microphone (T, P, K), to achieve a more stable, radio-ready, and pleasant voice. Its content-oriented approach preserves the naturalness of the performance and avoids the "muffled voice" effect often associated with classic de-essers.

Main Features

Intelligent analysis without threshold

Unlike traditional de-essers that rely on a threshold and detection based on the energy of a frequency band, smart:deess uses a neural network trained to recognize phoneme structures. Upon insertion on a track, the plugin builds a "vocal fingerprint" of the incoming signal and begins balancing sibilance and plosives. You save time on technical settings and can focus on the color and placement of the voice in the mix.

Phoneme detection and processing throughout the entire duration

smart:deess identifies specific phonemes in real time (S, Z, Sh, Ch, K, T, P) and determines their precise start and end points. The processing is then applied over the entire sibilance, not just its strongest peak, which helps preserve the timbre and achieve a more transparent result. Each phoneme type is analyzed and processed independently, allowing for finer correction than a simple "high-frequency cut."

Spectral shaping and automatic re-essing

Thanks to spectral processing, harshness can be musically sculpted rather than just attenuated. The plugin can also automatically "re-ess" to enhance phonemes that have become too dull or exhibit a lisping effect after correction, useful for maintaining natural diction while controlling aggressiveness.

Advanced customization

smart:deess offers a wide range of options to adjust the processing behavior according to the source: soft or percussive voice, microphone proximity, singing style, fast spoken voice, etc. The idea is to start with a usable result right from loading, then refine according to your artistic direction, instead of spending time "fixing" the take.

Technical specifications