Where other de-essers just reduce volume – smart:deess goes far deeper and creates a more natural, better balanced sound. By harnessing the power of a fully trained neural network, smart:deess identifies the exact start and end points of phonemes without the need to set a threshold.

smart:deess analyzes the signal’s actual content – not just the level of one frequency band – before it decides what to do. As soon as the plugin is loaded on a channel, smart:deess builds a ‘voiceprint’ of the input voice signal and starts balancing any sibilants and plosives that are present. Users only have to review the analysis and refine the results. smart:deess allows users to spend time on getting the sound they want, rather than correcting the sound they don’t want.

smart:deess identifies specific individual phonemes – S, Z, Sh, Ch, T, P… – in real time, determining the exact start and end point of each sibilance without users having to set a threshold. The entire sibilant is then processed, rather than just the loudest part, making the final result far more natural sounding. Each phoneme type is identified and processed independently, an “Ess” is treated differently than a “Ch”, which is treated differently than a “T”. The result is as though the original recording had simply been better balanced.

Features

• The most natural-sounding de-esser thanks to its AI technology
• De-essing and plosive reduction in one tool
• Real-time phoneme detection to identify sibilants and plosives
• Harsh sounds can be shaped, rather than ‘just turned down,’ via spectral processing
• Processes the entire sibilant from start to finish, without the need to set a threshold
• Handles different phonemes – S, Z, Sh, Ch, K, T, P – each with distinct processing
• Will automatically “re-ess” to improve any ‘dull’ and ‘lisping’ phonemes
• Huge range of options for customizing the processing

Tech Specs