De-Essing With Dynamic EQ: More Control Than Any De-Esser
Vocal mixing · 4 min read
A de-esser gives you one band. A dynamic EQ gives you as many as you want, and sibilance needs two: a narrow band at 5.5–7 kHz for the weight of the S, and a second at 8–10 kHz for the spit on T, Ts and Ch. Two bands doing 3 dB each beat one band doing 6 dB.
Two bands, not one
6 + 9 kHz
Q 4–5 · range -4 dB · release 30–60 ms
The 60-second answer
Insert a dynamic EQ after your corrective EQ. Band 1: bell at 6 kHz, Q 4, dynamic downward, range -6 dB. Band 2: bell at 9 kHz, Q 5, range -4 dB. Attack 1–3 ms, release 30–60 ms. Set each threshold so the band only moves on the sound it is aimed at.
Why one band always dulls something
Sibilance on most voices is not one hump. There is a lower peak that gives an S its weight, usually between 5.5 and 7 kHz, and a thinner, faster peak higher up that carries the spit on T, Ts, Ch and Sh. A single band wide enough to cover both has to move far enough to fix the louder of the two, which drags the quieter one down with it. That is the dull you hear: not too much reduction, reduction in the wrong place.
What a dynamic EQ gives you that a de-esser does not
Three things. Q, so you can work at Q 4–6 instead of a fixed band width. A range control, so the band can never pull more than the number you set. And a filter on the detector, so the band listens only above 4 kHz and stops being triggered by a belted vowel. That last one is why a dynamic EQ often needs less reduction than a de-esser to get the same result.
The two-band setup, step by step
| # | Do this |
|---|---|
| 01 |
Do the corrective EQ first Cut the mud and any 2–5 kHz resonance before you touch sibilance. Both change how loud the esses feel. EQ → dynamic EQ |
| 02 |
Band 1: find the weight of the S Solo the band, sweep 5.5–7 kHz until the S sounds thickest, set it dynamic downward with a hard ceiling on the movement. 6 kHz, Q 4, range -6 dB |
| 03 |
Threshold so it only moves on esses The band should sit still on vowels and drop on S and Z words. If it breathes on everything, raise the threshold 3 dB at a time. Reduction 3–5 dB |
| 04 |
Band 2: catch the spit Sweep 8–10 kHz for the thin hiss on T, Ts and Ch. Narrower than band 1, with less range. 9 kHz, Q 5, range -4 dB |
| 05 |
Set the timing to a consonant An S lasts roughly 60–150 ms. Attack fast enough to catch the front, release fast enough to be gone before the vowel. Attack 1–3 ms, release 30–60 ms |
| 06 |
High-pass the detector Set the sidechain filter above 4 kHz so a loud vowel cannot open either band. Then A/B against bypass on earbuds. Sidechain HPF 4–5 kHz |
The trick: one band for the S, one for the T, and never let them overlap
Set band 1 at 6 kHz with Q 4 and band 2 at 9 kHz with Q 5, then look at the two curves. If their skirts cross, a loud S pulls both bands at once and you are back to a wide cut with two names on it. Narrow them until there is visible daylight between the two, then set the thresholds independently and watch which words move which band: S and Z should move the low band only, T, Ts and Ch the high band only. Set up like that you can pull a full 5 dB out of the 6 kHz band on a very sibilant singer and keep every bit of the 12 kHz air, because nothing you did ever reached 12 kHz. That is the whole reason to do this instead of using the de-esser that came with your DAW.
Stop guessing the chain.
The Vocal Chain Bible is 88 real vocal chains from Grammy-winning engineers — every plugin, every setting, in order, with free alternatives for each one. 5+ years, 100+ interviews, one PDF.
Get the Vocal Chain Bible — $48 →Mistakes that kill it
- Q too wide. Below Q 2 a bell is basically a shelf, and a dynamic shelf at 7 kHz is a de-esser with worse timing.
- No range limit. Without a ceiling, one shouted word pulls 12 dB and the line lisps. Set range to -6 dB and forget about it.
- Release too slow. Over about 100 ms the band stays down into the vowel after the S, which dulls the word rather than the hiss.
- Skipping the detector filter. Left wide open, the sidechain hears the loudest thing in the vocal, and that is a vowel.
Keep going
FAQ
Is a dynamic EQ better than a de-esser?
For control, yes: you get Q, a range limit and a detector filter. A dedicated de-esser is faster to set up and often better at catching a whistle on a bright voice.
What Q should I use for de-essing with a dynamic EQ?
Q 4–6. Wide enough to cover the peak, narrow enough that the air above and the presence below stay untouched.
Can I de-ess with a static EQ cut?
Only as a last resort. A static cut is down for the whole song, including the 95% of words with no sibilance in them.
Who's behind this
Ivan Meshcheriakov · IvanFromRnD
5+ years ago I moved to LA with zero connections and started building from nothing — 14-hour days, meeting people and interning at studios. It took me 5+ years to meet and interview 100+ Grammy-winning engineers — then I turned it all into the Vocal Chain Bible.
Read my full story →