Carving Space for the Vocal: EQ the Beat, Not the Vocal
Vocal mixing · 4 min read
If the vocal is buried, turning it up makes it loud and still buried. Cut the instrumental instead: a bell of 1–3 dB at 2–4 kHz on the beat bus, plus a smaller dip at 250–400 Hz. Then sidechain that band to the vocal, so the hole only opens while somebody is singing.
Cut the beat
-2 dB at 3 kHz
Sidechained to the vocal · closed in the gaps
The 60-second answer
Put a dynamic EQ on the instrumental bus. Bell at 2.5–3.5 kHz, Q 1.5, range -3 dB, triggered by an external sidechain from the vocal bus. Attack 5–15 ms, release 100–150 ms. Add a static -1.5 dB at 250–400 Hz for the body clash. Never more than 3 dB in either band.
Masking is a ratio, not a level
Two sounds in the same band fight, and whichever is louder wins. When the vocal is masked at 3 kHz you have two ways to change the ratio: add 3 dB to the vocal or take 3 dB off the beat. Both work on paper, but only one keeps the vocal sounding like a person. Boosting 3 kHz on a voice is the fastest way to make it harsh, while pulling 2 dB out of a busy instrumental there is inaudible, because ten other things cover that range.
Why 2–4 kHz and 250–400 Hz
2–4 kHz is where consonants live, so it is where intelligibility lives. Hi-hats, snare tops, bright synths, guitars and sampled loops all pile into the same window, and a vocal that loses that fight sounds like it is behind the speakers however loud you make it. 250–400 Hz is the second clash: the vocal's chest against keys, 808 harmonics and pads. Fixing the first makes the vocal clear. Fixing the second makes it feel in front.
Carving the space, step by step
| # | Do this |
|---|---|
| 01 |
Find the clash by muting, not soloing Loop a chorus and mute the vocal in and out. Listen to what the beat does when it appears. What disappears is what you cut. Mute/unmute the vocal |
| 02 |
Count what lives at 2–4 kHz Hats, snare top, bright synth, guitars, loop bleed. If five elements share that window, no vocal EQ fixes it. List the offenders |
| 03 |
Static cut first, on the bus Bell at 3 kHz, Q 1.5, -2 dB on the instrumental bus. Judge it on the words, not on the beat. -2 dB, Q 1.5 |
| 04 |
Add the low-mid dip -1 to -2 dB at 250–400 Hz, Q 1.5. This is the one that makes the vocal feel closer, not just clearer. -1 to -2 dB |
| 05 |
Make the 3 kHz band dynamic Same band, dynamic downward, external sidechain fed from the vocal bus, range -3 dB. Now the beat is untouched between phrases. Range -3 dB, key: vocal bus |
| 06 |
Set the timing so it never pumps Attack 5–15 ms so it opens with the word. Release 100–150 ms so it closes between phrases but not between syllables. 5–15 ms / 100–150 ms |
The trick: sidechain the hole so it only exists while the vocal sings
A static cut in the instrumental is a compromise you pay for all song, including the intro, the bridge and the bars where nobody is singing. Everyone has heard a mix where the beat sounds slightly broken in the instrumental sections, and that is why. The fix is a dynamic EQ with an external sidechain input — FabFilter Pro-Q 3 and later, TDR Nova and most stock multiband dynamics plugins have one, so check yours. Put it on the instrumental bus, feed the detector from the vocal bus, and set a dynamic bell at 2.5–3.5 kHz with a range of 3 dB. The instrumental is now untouched until the singer opens their mouth, and then a 3 dB hole opens exactly where the consonants are, for exactly as long as there are words. It needs less reduction than a static cut and it beats a full-band ducking compressor, which turns the whole beat down and reads as pumping.
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
- Boosting the vocal at 3 kHz instead. You get a harsh vocal that is still masked, because the beat still has more energy.
- A 6 dB hole in the beat. The instrumental sections sound broken and the mix loses its bite. Two or three dB is all it takes.
- Carving with the vocal soloed. Masking only exists between two sources. If you cannot hear both, you cannot judge it.
- Release too fast. Under 60 ms the band opens and closes between syllables and the hats flutter.
Keep going
FAQ
What frequency should I cut in the beat for vocals?
2–4 kHz for clarity and 250–400 Hz for the body clash, 1–3 dB in each. Do it on the instrumental bus, not on individual tracks.
Is sidechain ducking the instrumental cheating?
It is how loud, clear records get made. A 3 dB dynamic dip at 3 kHz is inaudible as an effect and buys more clarity than 3 dB on the vocal fader.
Should I cut the whole instrumental bus or single tracks?
Bus first, because it is one move and easy to undo. Go to individual tracks only when one element, usually a hat or bright synth, is the problem.
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 →