Sibilance is what we call the hissing noise made by “ess” and “shh” sounds in the voice. We typically don’t notice it in conversation, but it can get really harsh on a recorded vocal! Naturally, the digital audio world came up with the de-esser to take care of excess sibilance and make our vocals sound smoother and more pleasant.
What Does a De-Esser Do?
The most commonly used tool for taming sibilance these days is a dedicated de-esser plugin. A dynamic EQ like FabFilter’s Pro-Q 3, or even a multiband compressor can be used to get the same effect as a de-esser—however, plugin developers design de-essers to operate specifically on mid and high frequencies where sibilance lives.
A de-essing plugin reduces a frequency’s amplitude by a chosen amount when it surpasses the threshold. In the case of harsh sibilance, anywhere between 4 and 10 kHz will often be where the problem lies. For example, you can choose to tame 7 kHz by reducing it by 3 dB when it surpasses the threshold.
Over de-essing becomes very apparent, though, audibly warping transients and making a vocal sound unnatural. It’s best used sparingly!
A de-esser isn’t just for sibilance.
Remember, all it’s doing is compressing a specific frequency when it gets too loud. It’s up to us to decide which frequency that is!
Most de-essing plugins let you set frequencies within a certain range—usually something like 2 to 10 kHz. The most offensive sibilance is typically between 4 and 7 kHz—but what’s happening lower than that?
One of my favorite tricks is to set up a de-esser on my vocal bus with the frequency set around 2 kHz. This isn’t acting on sibilance specifically, but rather on the sometimes overly harsh and unpleasant sounds in the 2 kHz range!
Control overall vocal brightness with a de-essing plugin.
If you’re going for an in-your-face, exciting sounding vocal, you need that extra air in the highs. The problem is that there’s a fine line between a pleasant bit of brightness and harshness. A de-esser works brilliantly to keep things controlled and musical.
What you can do is boost the top end with an EQ (think anywhere between 6 and 12 kHz) and tame it with a de-esser. You’ll get the brightness and clarity from boosting with the EQ, minus the exaggerated “s” and “t” sounds thanks to the de-esser.
Don’t rely on a plugin for the most natural results.
While it’s quick and convenient to slap a de-esser on our vocals and call it a day, sometimes they need a little more care to be their absolute best. This requires a bit of manual labor.
Manual de-essing means grabbing all the sibilant parts of a vocal and clip-gaining them down. If this sounds ridiculous and annoying, it is, but it’s totally worth it the hassle. When you take the time to go over edits like this, the result can sound much more natural “automatic” de-essers.
What you would do is look for the “ess” in the waveform, which usually looks something like a solid (American) football shape. Then you’d separate the ess into its own clip and pull the gain down on it.
Manual de-essing has other benefits, too, in that it lets you decide how hard sibilance hits other parts of the signal chain. Nothing’s worse than setting up a great sounding vocal chain and realizing every time the singer hits an “ess,” the compressor noticeably clamps down and sounds weird. Everything was fine up until then!
Of course, manual de-essing involves a lot of time and patience, but if it makes your vocals sound incredible, it’s very much time well spent!
De-ess in stages.
This is a common tip for approaching compression as well. The idea isn’t to slam one processor for major amounts of gain reduction, because that’s when compression gets unnaturally audible. Instead, you get a dB or two here, and a dB or two there.
It’s the same with de-essing. Your results will be much more natural if you de-ess a little bit here, and a little bit there, using more than one in your chain.
For example, you might try manually de-essing a dB on the biggest offenders while still having a de-esser plugin active in the signal chain. This way you aren’t relying on the plugin for all of the dynamic control you need.
Automate your de-esser settings.
It’s very tempting to set it and forget it, right? In fact, lots of engineers do! However, this isn’t maximizing the potential of the performance for the best possible sound.
Realistically, there are countless variations in a vocalist’s recording making sibilance sound different throughout the song. You can find a solid middle-of-the-road frequency that catches most of the worst offenders, but taking the extra time to automate when a different “ess” slips through the cracks can really help a performance shine.
Check out our other Tips for Mixing Vocals!