How to De-Ess Vocals: Sibilance De-Esser Plugin Tutorial
January 21, 201932,930 Views
2x Grammy-winning mix engineer Dave Darlington shows how he likes to de-ess vocals using the Waves Sibilance plugin, explains the plugin’s features – Threshold, Range, Detection, Lookahead and Wide vs. Split mode – and shows where he recommends placing Sibilance in a vocal chain.
Why Use a De-Esser
De-essers take away the harsh S, Sh, and T sounds in a voice that are not musical.
We need some of those sounds to be able to understand the lyrics and the content, but we don't want an overabundance of them, because especially at low volumes, those non-musical sounds can poke through a mix and be very distracting.
How the Sibilance Plugin De-Esses Vocals
The Sibilance plugin uses Waves' Organic ReSynthesis technology to separate the vowels and formants of a voice from the sibilant artifacts. In other words, it separates the sibilance sounds from the rest of the vocal signal. That makes Sibilance especially surgical when it comes to de-essing vocals.
If you watch the waveform display in the Sibilance plugin interface, you can actually see how that separation is happening.
Every time an S or a T sound or something high up in the frequency range is hit, you can see the detection circuit react, and then you can see how much it's doing by the display at the top (the green line).
How to De-Ess with Sibilance: The Controls
Sibilance only has a few controls, it's very simple.
The Threshold control determines at which level the plugin is actually going to work. As I turn the threshold lower, you can see it closing in on the waveform, so when you watch the wave go by you can set it so that it's hitting the parts you want to hit, and not others.
The Range control determines how much of the sibilant sound you're taking out. As you move it down to take out more, you'll see the waves dip down deeper.
As always, the trick with any plugin is to make it not really audible to the listener. You want to do the job that you're trying to do, but you don't want to draw attention to the fact that something's actually changing. Just rock the range amount so it's getting rid of the bad stuff, but not so much that it makes it sound fuzzy or unnatural.
The Detection control is sort of the Q or the width of the signal you're manipulating. Lower values would zero in on just the ‘ess-y’ part of things, and higher values would be a little wider, so you get detection on more ‘wooshy’ kinds of sounds.
The Mode control determines the frequency roll-off, ranging from Wide to Split. If you turn this knob all the way to the Wide end of things, you'll be getting a full-range, full-band roll-off. If you turn it all the way to Split, you're only looking at frequencies above 4 kHz. For de-essing, try it right in the middle between Wide and Split, and see if you want to adjust it to taste. I start with this knob fairly centered, and that seems to work pretty well.
The Lookahead control allows the program to look ahead at the sound wave as it's coming, which makes the plugin's detection of sibilant sounds even more subtle and nuanced. The Lookahead option does introduce some latency, so you'll need to see whether it serves your needs.
Finally, you can use the Monitor control (with the little speaker icon) to solo the part that you're getting rid of. This way, you can adjust the controls while listening only to the S, Sh and T sounds that you're trying to remove. That's a very valuable tool: it lets you zero in on what you're trying to clean up.
Where to Place the Sibilance De-esser in Your Vocal Chain
I put the de-esser late in the chain, because if I take the S’s out early and then boost the highs, sometimes they tend to creep back in and sound a little less natural. But that's a personal decision. Different engineers like to do it different ways. Some people like to do the cleaning part early in the chain, and then EQ; and you can also just drag it around to different positions in the chain and see what sounds best.