Esspresso is a new plugin/app from Swedish developer Klevgrand that promises to cut down on sibilance, the intrusive "sss" sounds and hissing noises that can make for piercing audio.
With Klevgrand's usual mix of no-nonsense interface and under-the-hood controls, Esspresso doesn't make too much of a fuss about advanced parameter tweaking, and is aimed at getting the job done fast, whether you're on the desktop or rocking an iPad.
Esspresso's interface is split into two halves, the upper dealing with its detector circuit and the lower dealing with the surpressor (see below for more on how it works). Seven controls enable you to set the Dectector's Sensitivity (input gain), bandwidth and Threshold, get control over the Surpressor's filter type, bandwidth and Ratio, and to Solo the detector circuit to home in on the excruciating ssstuff you want to ssurpresss.
The interesting thing about Esspresso is that its Detector and Surpressor circuits are completely separated, meaning that you can use a band at any frequency to trigger reduction in a band at any other frequency - interesting on drum busses, for example.
You can grab an Esspresso as a VST/AU plugin at the Klevgrand website for $49.99, or on the App Store for $7.99 (£4.99)
When to de-ess
De-essing tackles a sort-of minor problem in audio production - hiss. Listen to an old '60s or '70s vocal (throw the volume up a bit), pick out the "sss" sounds, and you'll start to hear they're grating, and they're everywhere - there've been 71 of them in this article so far.
The problem can come from the closeness of the singer to the microphone, problems with the microphone itself, or a singer with large, gappy teeth. Be careful, though, since once your mind's wired up to listen to overactive "sss" sounds, watching TV can become an infuriating trip through hell (if it wasn't already).
De-essing helps to crack down on these sounds, allowing for smoother audio on the channel and therefore a smoother mix. Careful, though, as reducing the sibilance too much can result in a lisping sound. Other high frequencies can be subject to de-essing, too - whatever's problematic can be fair game, but cymbals are the most likely culprit.
How de-essers work
There are several ways a developer can go about creating a de-esser, and you can try to take some of the same principles into account to make your own!
Perhaps the most popular way of achieving the effect is to have the processor duplicate the audio, then filter the duplicated signal, leaving just the offending sounds, then using it as a sidechain signal to trigger a compressor over the original signal. This is effective, but a blunt approach that will compress the full frequency range - fine for isolated vocal recordings, not ideal for buss processing or engineering live recordings…
Multiband compression, or a version thereof, can also be used to crack down on just the sibilant frequencies, and only when they happen. Similarly, a dynamic EQ can react to the input signal, affecting the gain of a band only when audio material is present within it.