A de-esser is essentially a specialized multiband compressor which specifically targets sibilance in a vocal. If you’re having an issue with sibilance in your vocals, let’s better explain what is a de-esser and how to use one to clean up your vocal.
What is Sibilance
Sibilance is the hissing sound associated with “S” sounds. It can crop up on a few other consonants such the “T” sound, but it’s especially prominent on “S” sounds (and OF COURSE there’s two “S” sounds in the world sibilance itself).
It commonly occurs somewhere in the 4-10k frequency range with the specific frequency range varying by singer and recording.
This sound is generally perceived as being unpleasant and harsh on the ears.
If you have a problem with sibilance in your vocal, it can be incredibly distracting for the listener to the point that it’s all you can pay attention to.
The causes of sibilance are many and range from certain voices or singing styles can lend themselves to sibilance more than others, but improper microphone choice or placement can lead to this, as well.
Thankfully even if you have sibilance on your record, you can effectively treat it with a de-esser.
What is a De-Esser
As I mentioned in opening, a de-esser is a type of multiband compressor which specifically targets sibilant sounds and that 4-10k region or so.
As sibilance occurs in a vocal, these frequencies experience a brief but prominent spike. A de-esser is designed to recognize and isolate these spikes, compressing them relative to everything else.
This attenuation smooths out that hissing sound, making it more subtle in the context of the rest of the vocal.
The goal isn’t to remove that “S” sound completely or overly aggressively smooth it out to the point that it doesn’t sound natural. Driving a de-esser too hard will yield an unnatural sound, so let’s cover how to use a de-esser.
How to Use a De-Esser
For this example, I’m going to be using FabFilter’s Pro-DS.
In the past, a de-esser was a more specialized plugin you could only get from a third party. Today, most modern DAWs feature a de-esser as a stock plugin.
Even if your DAW doesn’t have one stock, it likely has a preset on its multiband compressor like in my DAW of choice, Ableton Live. As I mentioned, a de-esser is essentially than a multiband compressor which focuses on that 4-10k or so range.
I prefer FabFilter’s Pro-DS because I find that it works much more transparently than a multiband compressor for specifically targeting that hissing sound.
The rest of the vocal remains intact, so it makes for the most transparent sibilance controller I’ve used, and I get that result with minimal effort of finding the sibilance. With a multiband compressor you need to manually find those sibilant frequencies whereas with Pro-DS most of that work is already done for you.
Getting back to how to use a de-esser, first let’s find a section on our vocal which has a lot of “S” sounds and specifically ones which could benefit from being tamed a bit. That last point is important, because not every vocal needs a de-esser. Depending on the vocal recording, there might not be a problem to address.
Drop the plugin on the track and set your range as the first order of business. Pro-DS targets the 7k-14k by default (with 6k-14k for male vocals).
This may seem like a large range, but again, it’s about finding the peak of those “S” sounds.
Step 1 – Find Sibilant Frequencies
The “Audition” button solos the frequency range we’ve selected. Virtually any de-esser or multiband compressor will have this feature so you can sweep in this range.
Play the most offending sibilant moment in the vocal on repeat and sweep until you find the frequency range where it’s loudest. This is the “fundamental” if you will of that sibilance and the frequency we especially want to target.
We can then narrow the band (which you should do with a multiband compressor) to specifically target the heart of that sibilance to pull it down.
As Pro-DS is specifically designed to locate sibilance (it’s the highlighted regions of the audio shown below), you don’t need to fine tune the range as much.
Step 2 – Set Threshold and Range
Now we need to set the threshold.
Just like with a regular compression threshold, this dictates how much of the signal is going to be compressed.
The only difference on a de-esser is that the input level is limited to the volume in the frequency range we selected in the first step.
As I mentioned earlier, sibilant frequencies have short but sharp peaks in volume. As such, we want to set the threshold to exclusively target these peaks in that frequency range. This leaves the rest of the vocal largely untouched to achieve the most transparent results.
To set your threshold, continue looping the most sibilant part of your vocal on repeat. You can use the “Audition Triggering” button on Pro-DS which looks like a pair of headphones. This exclusively plays the sibilant sounds which are being attenuated so you can hear what is being affected.
Pull the threshold down so that you hear a smoothing of the peak(s) of that sibilant part.
Note that you should do this in concert with the “Range” or the equivalent setting on your de-esser to determine the maximum amount of attenuation.
These two settings together will determine the amount of gain reduction you achieve on the sibilant parts of your vocal.
Generally I like to max out at roughly 6dB of gain reduction on those biggest spikes which come from the most sibilant parts.
The benefit of focusing on the most sibilant part of your vocal is that when you get that 6dB for a relatively natural attentuation, you know the rest of the less sibilant parts won’t trigger your settings quite as hard. They’ll still be treated for a few dB in gain reduction, but it will keep that treatment transparent.
Step 3 – Don’t Over Compress
This last step is more of an important check and reminder of that last note.
Over-compressing sibilance removes the transient of the “S” sounds. In addition to losing the punch of the transients, the hiss sounds are flattened into lisp sounds.
You can experiment with overly aggressive settings by dropping your threshold too low and raising the range to high.
Once you get into 10+dB on average of gain reduction in those specific frequencies, you’ll hear your vocalist instantly develop a lisp.
De-essing is all about subtle and transparent treatment, so aim for that 3-6dB range to just nudge that hiss down to a natural sounding level rather than removing it completely.
One last tip: de-essers don’t have to exclusively be used on vocals. While some of them like Pro-DS is specifically designed to hear sibilant vocal spikes, many can be used to attenuate harsh frequencies in general.
Occasionally when I’m working with cymbals with a bit too much cymbal, so much so that it bleeds into hissing territory, you can use a de-esser to attenuate them in a similar way because you’re working in a similar frequency range as sibilance on a vocal.
I mentioned using a de-esser as an option for how to fix a harsh mix, so don’t limit it exclusively to a vocal tool in order to get the most out of it!
- Sibilance is the hissing sound associated with consonants “T”, “Z”, and particularly “S”.
- The frequencies of sibilance can range anywhere from 4k to over 10k.
- A de-esser can be used to smooth out sibilance by attentuating those frequencies like a multiband compressor when the volume at those frequencies has a quick burst associated with a sibilant sound.
- Find your most grating sibilant instant on your vocal and use this to find the exact frequency range it occurs at and amount of gain reduction you want via the threshold and range.
- Aim for 3-6dB on sibilant frequencies at their peaks for the most natural and transparent treatment of the vocal.
- If you’re too aggressive with your de-essing, you’ll remove the transients and smooth the sibilance to a fault, replacing the hissing with a lisp sound.
- Try a de-esser on other elements of your mix which are too harsh in those upper frequencies.