Every vocal is different: the singer, the mic, the room. So nothing here is a preset. These are the places I look first on a lead vocal and the moves I usually end up making. Start here, then let your ears decide.
The cheat sheet
| Move | Where | How |
|---|---|---|
| High-pass | 80-100Hz male 100-150Hz female 150-250Hz backing stacks | Gentle slope first. Push it up until the voice thins, then back off. |
| Mud / chest | 200-400Hz male 300-600Hz female | Cut 2-4dB, wide Q. Don’t take it all: that’s the warmth. |
| Boxy | 300-500Hz | Cut 2-3dB, medium Q. The “cardboard tube” sound. |
| Nasal / honky | 700Hz-1.5kHz | Boost a narrow bell 4-5dB, sweep until it sounds ugly, flip it to a 2-3dB cut. |
| Harsh / whistle | 2-4kHz 3.5-3.8kHz on condensers | Narrow cut, 2-3dB, before the compressor. |
| Presence | 2.5-5kHz male 4-6kHz female | Wide boost, 2-3dB. Or cut the guitars and keys there instead. |
| Sibilance | 4-8kHz | De-esser, last in the chain. |
| Air | around 10kHz | Wide bell, 2-3dB. Not a high shelf. |
Two rules sit over all of it. Corrective cuts go before the compressor. And every decision gets checked with the whole mix playing, not in solo.
Want to see where a vocal sits against everything else? It’s on the EQ frequency chart, with 23 other instruments.
The order I do it in
Plugin order on a vocal matters more than most people realise. The way I think about it: fix, control, flavour.
- Corrective EQ. High-pass, then cut the mud, the box, the nasal spot and any whistle. This goes before the compressor because a compressor reacts to whatever you feed it. Leave the low end and the resonances in and it wastes gain reduction on things you don’t even want.
- Compression. Levels the performance. (Compression is its own topic: lesson 01 of my compression course is free.)
- Character EQ. Now the boosts: presence and air. You’re shaping the controlled signal, not the raw one.
- Saturation. Just a touch, if the vocal wants it.
- De-esser, last. Compression brings the esses up, top-end boosts make them louder, and saturation adds new high-frequency harmonics. De-ess first and everything after it brings the sibilance straight back.
In a real session my chain often runs corrective EQ, two compressors, a touch of saturation, a Pultec-style EQ for the air, the de-esser, then a final EQ for last tweaks. Each plugin has one job.
Common vocal EQ problems
What each one sounds like, where it lives, and what to do about it.
Muddy or chesty (200-400Hz)
Thick and undefined, like the singer has a cold. Proximity effect makes it worse: the closer the singer is to the mic, the more low end builds up. Almost every male vocal needs some attention here, especially on a large-diaphragm condenser.
High-pass first (80-100Hz on a male voice, 100-150Hz on a female voice), then cut 2-4dB around 200-400Hz with a wide Q. On female voices the build-up usually sits a little higher, around 300-600Hz, and there’s less of it.
Next time, fix it at the source. Around 30cm between the singer and the mic is a good place to start.
Boxy (300-500Hz)
That singing-into-a-cardboard-tube sound. Really common with cheap condensers in untreated rooms, because the room resonances land right in the vocal’s power zone. Cut 2-3dB around 300-400Hz with a medium Q and you’ll hear it open up. If it only happens on certain words or notes, use a dynamic EQ so it only cuts when the boxiness shows up.
Nasal or honky (700Hz-1.5kHz)
The pinched-nose sound. Every voice has its nasal frequency in a different place, so don’t just cut 1kHz and hope. Find it:
- Set a narrow bell and boost it 4-5dB.
- Sweep it slowly through 700Hz-1.5kHz until the voice sounds its worst.
- Flip the boost to a 2-3dB cut, then widen the Q until it sounds natural again.
Some singers have it as high as 1.5kHz. You have to find it each time.
Harsh or whistling (2-4kHz)
Where a vocal goes from present to painful. Almost every vocal has a harsh spot somewhere between 3kHz and 4kHz, even on expensive mics. On a large-diaphragm condenser it’s often a whistle around 3.5-3.8kHz that gets worse when the singer pushes, and worse again once you add compression and saturation. Once you hear it, you can’t unhear it.
Sweep a narrow cut between 3.5kHz and 4kHz, find the worst spot, and dip it 2-3dB. Do it before the compressor. Don’t overdo it, or the vocal goes woolly.
If it comes and goes, a static cut is the wrong tool. A dynamic EQ only cuts when the harshness is there. I used to automate EQ word by word for this. These days I mostly let Soothe handle it and spend that time on the creative side of the mix.
Sibilant (4-8kHz)
Esses, tees and effs that jump out and spit. Female vocals often carry more energy up here naturally. A de-esser is the main tool, and it goes late in the chain, after compression and saturation, because both bring sibilance back up.
I use a de-esser in broadband mode by default: it ducks the whole vocal for a split second on each S. High-frequency-only mode can make the singer sound like they’ve got a lisp, because it carves out the top end that makes consonants clear. If sibilance still gets through, a dynamic EQ on the exact sibilant frequency catches what the de-esser misses.
On stacked vocals, de-ess each track before they hit the bus, since every layer adds its own sibilance. And filter the top off your vocal reverb and delay returns. A bright reverb spreads every S across the whole mix.
Buried in the mix
The instinct is to boost the vocal’s presence. Try the opposite first: cut 2-4kHz on the guitars, keys or synths by 2-3dB with a wide Q. The vocal steps forward without you touching it.
If it still needs help, a gentle wide boost of 2-3dB in the presence range: around 2.5-5kHz on a male voice, a little higher (4-6kHz) on a female voice.
And if the vocal sounds dry even though it has reverb on it, check the brightness before you turn the reverb up. A dark vocal doesn’t excite the reverb’s top end, so the reverb seems to vanish. A little air on the dry vocal can make it appear without touching the send.
Thin or brittle
Usually self-inflicted. The most common cause is EQing the vocal in solo and cutting too much warmth. What sounds muddy on its own often sounds right against a full band. Put back some of 100-250Hz on a male voice or 200-400Hz on a female voice, or pull the high-pass back down.
The other usual suspects: too much air, or a de-esser working so hard it sucks the life out of the whole top end.
One more thing. When someone asks for a warmer vocal, they rarely mean more bass. They usually mean less top end. Try taming 1.2-1.6kHz before you add any low end.
Male, female and backing vocals
Same problems, different places.
| Male | Female | Backing | |
|---|---|---|---|
| High-pass | 80-100Hz | 100-150Hz | 150-250Hz, higher on dense stacks |
| Mud | 200-400Hz | 300-600Hz | 200-500Hz |
| Presence | 2.5-5kHz | 4-6kHz | 3-6kHz, 1-2dB at most |
| Air | 8-14kHz (bell around 10kHz) | 10-16kHz | 10-16kHz |
| Watch for | Harshness, 2-4kHz | Sibilance, 4-8kHz | Competing with the lead, 2-5kHz |
Backing vocals
Backing vocals support the lead. They don’t compete with it. Every new layer adds more 200-500Hz weight, so high-pass harder than you would on the lead, then cut 2-4dB in that range. The thicker the stack, the more you cut.
If they sound as present as the lead, cut 2-3dB around 2-5kHz with a wide Q to push them back, and let a little air make them sound wide. A good test: if you can’t tell they’re there, but the mix sounds empty without them, they’re sitting right.
Process each track on its own before it hits the bus. Every take has its own dynamics, its own sibilance and its own resonances, and one plugin on the group can’t respond to all of them.