How to Fix Nasal Vocals (800 Hz-1.2 kHz)
Vocal mixing · 4 min read
Nasal vocals have a peak between 800 Hz and 1.2 kHz. Sweep a narrow boost through 700 Hz–1.5 kHz to find it, then cut 2–3 dB at Q 4–6, never more than 4 dB. But first move the mic 10 cm and record four bars again, because half the time the honk moves with the mic.
Nasal peak
800 Hz–1.2 kHz
Cut 2–3 dB at Q 4–6 · 4 dB max
The 60-second answer
Record a test: four bars at 10 cm, four at 20 cm, four angled about 20 degrees off the mouth. If the honk changes level between takes it is placement. If it is identical it is the performance, so sweep a +10 dB bell at Q 8 through 700 Hz–1.5 kHz and cut 2–3 dB at Q 4–6.
What nasal looks like in the spectrum
When too much of the sound comes through the nose instead of the mouth, the nasal cavity adds a resonance of its own and cancels part of what the mouth is doing. The result is a narrow lift in the 800 Hz–1.2 kHz region with a slightly hollow feel above it. It is loudest on closed vowels, the ee and ay sounds, and on words with n and m in them. That is how you tell nasal from boxy: boxy is on every word, nasal is on specific ones.
Why EQ can only do part of the job
The resonance is in the performance, so it is on the lead, the double, the harmony and every adlib. Cutting 3 dB on the lead leaves five tracks honking. And because it only appears on some vowels, a static cut is wrong for most of the song: it takes 3 dB out of every word to fix eight of them, and 1 kHz is where a lot of a vocal's forwardness lives. Fix it at the mic first, then use a dynamic band.
Fixing nasality, step by step
| # | Do this |
|---|---|
| 01 |
Do the three-take test Ten centimetres, twenty centimetres, and twenty off-axis. Two minutes of recording tells you if this is a mix problem or a placement problem. 10 cm / 20 cm / 20 deg off |
| 02 |
Ask for an open vowel Have the singer drop the jaw and open the throat on the offending words. One retake usually beats any plugin. Retake the worst line |
| 03 |
Sweep 700 Hz to 1.5 kHz Narrow bell, +10 dB, Q 8. The honk is normally a single spot, most often between 900 Hz and 1.1 kHz. Bell +10 dB, Q 8 |
| 04 |
Cut 2–3 dB at Q 4–6 Small and narrow. Past 4 dB the vocal goes hollow and starts sounding like a phone call, because 1 kHz does real work. -2 to -3 dB, Q 4–6 |
| 05 |
Make it dynamic Threshold set so the band pulls on ee and ay words and sits still elsewhere. Range -3 dB, release 80–120 ms. Range -3 dB |
| 06 |
Fix the stack too Put the same cut on the vocal bus, or 1–2 dB more on doubles and harmonies where three copies of the resonance stack. Bus: -1 to -2 dB extra |
The trick: move the mic 10 cm before you open an EQ
A resonance that comes from the singer is identical in every take, at every distance, on every mic you own. A resonance that comes from the setup is not. So record the same four bars three times: at 10 cm, at 20 cm, and at 20 cm with the mic tilted 20 degrees so the singer aims past it. Line them up and loop the same word. If the honk is unchanged, it is the voice and you are going to EQ it. If it drops on the further take, it was proximity and a boundary reflection. If it drops on the off-axis take, it was the mic's presence peak firing down the singer's mouth. Any of those answers saves you from the 6 dB cut at 1 kHz, the move that turns a nasal vocal into a nasal vocal with no presence.
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
- A 6 dB cut at 1 kHz. The honk goes and so does the forwardness. The vocal ends up behind the beat, on a phone line.
- Confusing nasal with harsh. Nasal is 800 Hz–1.2 kHz on closed vowels. Harsh is 2–5 kHz on loud notes. Different fix.
- A static cut for a problem on eight words. Automate it or use a dynamic band. No reason to pay for it all song.
- EQing the lead and ignoring the stack. The double and both harmonies share the peak, and coherent layers stack it several dB higher.
Keep going
FAQ
What frequency is nasal in vocals?
A narrow peak between 800 Hz and 1.2 kHz on most voices. Sweep 700 Hz to 1.5 kHz with a boosted bell to find it.
How much can I cut for a nasal vocal?
2–3 dB, and 4 dB at the most. That region carries a lot of presence, so bigger cuts trade one problem for a hollow, distant vocal.
Is nasality the singer or the microphone?
Both happen. Record the same phrase at two distances and one off-axis angle: if the honk changes it is placement, if not it is the voice.
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 →