Vocal EQ cheat sheet: what to cut and boost

Here’s the chart first, since that’s probably what you came for. Below it is how we use it, and where it stops being useful.

The chart

Range	What lives there	Too much sounds like	Usual move
Below 80–100 Hz	Rumble, mic stand thumps, plosive “p” pops	A low wobble you feel more than hear	High-pass filter, but not above the singer’s lowest note
100–250 Hz	Chest and body	Boomy, especially when the singer was close to the mic	Small, broad cut if it booms
200–500 Hz	Room sound, boxiness	Muddy, like singing into a cupboard	Narrow cut where it rings
800 Hz–1.5 kHz	Nose and honk	Nasal, telephone-like	Narrow cut if one note sticks out
2–5 kHz	Presence, the consonants that make words clear	Harsh, tiring	Usually leave alone. Boost gently if the vocal sounds dull
4.5–10 kHz	Sibilance: “s”, “sh”, “t”	Spitty esses	A de-esser
10–16 kHz	Air, breath	Brittle, hissy	Gentle shelf or bell if the voice sounds closed in

Every singer and microphone puts these in slightly different places. A baritone’s chest sits lower than a soprano’s, and a cheap condenser in a bedroom has a different mud problem from a dynamic mic in a treated booth. Read the ranges as a map of where to start listening.

For the rest of the band, our free mixing cheat sheet has the same kind of ranges for kick, snare, bass, guitars, piano and overheads, plus compressor starting points, on one sheet to print.

Cut what you can hear, and keep it small

Most vocal EQ problems come from one or two resonances: a frequency that rings out every time the singer hits certain notes, usually because of the room or the mic. Sweep a narrow boost slowly through 200 Hz to 1.5 kHz and listen for the spot where the vocal suddenly sounds ugly. Cut there.

How much? Less than you think. When Mixroadie cuts a resonance on its own, it sizes the cut to how far the peak sticks out. That’s usually 2 to 4 dB and never deeper than 4 dB unless you ask, and never wider than twice the width the resonance measures. A 9 dB notch is a sign the recording needs fixing, or that you are cutting the singer’s actual note rather than a ring.

That second case catches a lot of people. If you sweep and find “a resonance” at 220 Hz on a song in A, that may simply be the singer singing A. A real resonance stays put while the melody moves. Mixroadie only cuts a peak that holds still while the notes change, for exactly this reason.

One range we leave alone by default: 2 to 5 kHz on a vocal. That’s where the consonants live, and it’s what makes lyrics understandable over a band. A harsh-sounding vocal often isn’t a 3 kHz problem at all. It can be sibilance higher up, or a compressor clamping down too fast. Cut presence to fix harshness and the vocal sinks into the guitars. If you really do hear one harsh ring in that range, cut it narrowly and check the vocal still reads in the full mix.

Boost less than the tutorials say

A vocal that sounds dull usually needs presence and air. The trouble is that each boost sounds better on its own, so they keep growing.

We cap our own presence and air lifts at 3 dB each, and only as far as vocals like this one usually go in finished records. A vocal that’s already bright gets nothing. We learned that number the hard way. An earlier version of Mixroadie added presence as a high shelf and air as another shelf on top of it, and on one vocal the two stacked up to +5.5 dB at the very top, close to double what either boost was supposed to add.

The fix was to make presence a broad bell at the singer’s own presence frequency instead of a shelf. That way presence and air each do their own job without adding up at 12 kHz. If you EQ by hand, the same trick works: a bell around 3–5 kHz with a wide Q, and a separate gentle shelf above 10 kHz.

De-ess after the EQ

Any presence or air boost also turns the esses up. So set your tone first, then deal with sibilance.

A de-esser beats a static EQ cut here because it only acts while the “s” is sounding. A fixed cut at 7 kHz dulls the whole vocal to fix a sound that’s there maybe 5% of the time. In REAPER, Mixroadie uses a de-esser that listens only to the ess range and turns the voice down for the length of each ess. If the esses barely stand out, it falls back to a narrow EQ cut of at most 4 dB, never centred below 4.5 kHz. Below that you start cutting the voice itself.

What a cheat sheet can’t tell you

The chart describes the vocal on its own. In a song, the vocal competes with guitars, keys and cymbals that sit in exactly the same 1–4 kHz range. You can EQ a vocal perfectly in solo and still lose it the moment the band comes in.

That’s a masking problem, and EQ on the vocal alone rarely solves it. We wrote about it separately in Vocal masking: why the band buries your singer.

The chart is still a good place to start listening. Just check its suggestions against what you hear in your own vocal.