De-essing Without Destroying Your Vocal
Sibilance — the harsh “s” and “sh” sounds in a vocal — is one of the most irritating problems in recordings.
And over-de-essing — the cure that creates a lisp — is just as bad.
What causes sibilance?
Sibilance is caused by high-frequency energy in consonant sounds, particularly S, SH, T, and CH. Most human voices produce these at frequencies between 5kHz and 10kHz. Large-diaphragm condenser microphones — the standard for home studio vocal recording — are often very sensitive in exactly this range and can exaggerate sibilance significantly.
Room reflections also contribute. Bright, untreated rooms reinforce high frequencies, and what sounds fine through headphones in the vocal booth can become harsh in a dry, reflective space.
How a de-esser works
A de-esser is essentially a frequency-selective compressor. When signal in a defined high-frequency range (usually 5–9kHz) exceeds a threshold, the de-esser reduces gain in that range momentarily, taming the harsh transient of the consonant.
The critical settings are the frequency range and the threshold. Too high a threshold: the de-esser never triggers. Too low a threshold: it compresses everything, turning sibilants into lisp and making the vocal sound muffled and processed.
Finding the right threshold
Set your de-esser to the frequency range where the sibilance is most problematic — often somewhere around 6–8kHz. Then push the threshold down while playing back the vocal until you can hear the de-esser catching sibilant sounds. Now slowly raise the threshold until it’s just catching the worst offenders and leaving the rest alone.
The goal is selective action — the de-esser should only fire on the harshest sounds, not uniformly compress all high-frequency content. If the vocal starts sounding like it has a lisp, the threshold is too low or the frequency range is too broad.
An alternative: clip gain on individual sibilants
For vocals with occasional harsh sibilants — not consistent sibilance throughout — manual clip gain reduction on the specific consonants is often cleaner than a de-esser. Find the word, find the “s,” reduce its level by 2–3dB. Done. No processing artifacts.
This takes more time but gives precise control, especially for one or two problem sounds in an otherwise clean performance.
Jacob Korn at tailout.de addresses sibilance at mastering when needed, but the tools available — broad EQ, multi-band limiting — are coarser than surgical de-essing in the mix. A well-de-essed vocal arrives at mastering ready for final polish, not correction.
Ready to send your mix? Visit tailout.de — professional mastering with careful attention to vocal detail.