How to De-Ess Vocals Without Making Them Dull

To de-ess vocals without making them dull, reduce the harsh consonants only when they jump out, then check that the rest of the words still sound clear. Start with the recording and the existing vocal chain before adding more processing.

This workflow is for producers and home-recording artists whose rap or sung vocals sound sharp on “S,” “SH” or “T” sounds. Use a short loop that includes both an offending word and a normal phrase. Keep an unprocessed copy so every decision is reversible.

1. Check whether the problem starts in the recording

Listen to the raw vocal, then the processed version at a similar listening level. If the sharpness is already there and you can record again, test a small microphone-position change before committing to the next take. Shure recommends trying a slightly lower microphone position when pronounced consonants are a problem. Compare the whole vocal tone, not just the “S.” See Shure’s vocal-recording guide.

Make a quick test recording of the same line in both positions. Choose the take that balances natural tone and clear words; a position that removes brightness but makes the performance sound distant is not automatically an improvement.

2. Find the actual trigger instead of copying a frequency

Add a de-esser to the individual vocal track. If it has a detector-listen or audition control, use that briefly to find the consonants you want it to catch, then return to normal monitoring. Different plugins label these controls differently.

For example, FabFilter Pro-DS separates its detection frequency range from Threshold and Range controls. Threshold determines when processing triggers; Range controls the amount of reduction. Its Single Vocal mode is intended for an individual voice. These are useful functions to understand, not a requirement to buy that plugin.

Lower the threshold gradually while listening. If normal vowels keep triggering processing, revisit the detection settings. There is no single frequency or threshold that fits every singer, microphone and recording level.

3. Judge the words, not a gain-reduction target

Use enough reduction to make the distracting consonant sit comfortably in the phrase. Stop and back off if the singer develops a lisp or loses clarity. A meter can show activity without telling you whether the result is musical.

Waves explains that excessive reduction or an overly broad affected range can make a vocal muted or unnatural. Its de-essing guide also notes that EQ and compression can accentuate sibilance. Compare the vocal before and after those processors: a bright EQ move may need revisiting instead of compensating with stronger de-essing.

Check a quiet verse and a louder hook before accepting one setting for the song. The goal is intelligible lyrics with controlled sharpness, not a perfectly flat meter.

4. Fix the few words that need separate attention

If only one or two consonants remain distracting, use clip gain or volume automation on those moments instead of making the whole track's processing more aggressive. For example, lower just the sharp opening consonant in one word, then replay the entire line.

Keep the transition smooth and listen for clicks at edit boundaries. iZotope’s vocal-editing guide describes manual de-essing with clip gain and recommends crossfades where edited regions meet. Preserve the original audio so you can undo an edit that changes the delivery too much.

5. Check what reaches the reverb

If the dry voice sounds controlled but the reverb still sprays a long “S” into the mix, inspect the effects path. A send taken before the de-esser may still feed the original sharp consonant into the effect. Test de-essing the signal feeding that effect and compare the result.

iZotope’s de-essing guidance suggests experimenting with de-essing before reverb or distortion on an effects send. Verify your routing rather than assuming every send receives the same processed signal.

A final listening check

• With the beat: Can you understand the line without the consonant jumping forward?
• With effects: Does the reverb or delay bring the sharpness back?
• Across sections: Does a setting that works in the hook damage the quieter verse?
• Bypassed: Is the change useful, or is it simply making the whole vocal quieter and darker?

Use a comfortable monitoring level and take a break if you keep making stronger changes without a clear improvement. If the vocal remains unpleasant between consonants, stop treating every problem as sibilance and reassess the recording and processing.

Want help finishing the whole mix?

Book MixedByEl’s Mix & Master service for your vocals and instrumental. Choose Beat & Vocals for a stereo instrumental, or Beat Stems & Vocals when you have separate instrumental parts.

Keep your dry recordings and include a rough mix with timestamps for the words that bother you. Use the vocal file-prep checklist to prepare the handoff without locking unwanted processing into your files.