How to Mix Vocals at Home: A Step-by-Step Guide for UK Artists

Your vocal is the part people hum on the way home, which is a roundabout way of saying it has to sit right. Most home mixes fall down in the same place: the voice is either buried, harsh, or sounds like it was tracked in a tiled bathroom. A repeatable process fixes most of that. What follows is a practical order of work — clean-up, EQ, compression, de-essing, reverb — that works in any DAW, with the plugins you already own.

Sort the recording before you reach for a plugin

Plugins flatter a good take and punish a bad one. Get the source right and every later decision becomes easier.

• Work 15–25 cm from the mic, slightly off-axis so plosives miss the capsule. A pop filter helps; so does not eating the grille.
• Sing into the soft end of the room. A duvet behind the mic and a rug underfoot beat most budget acoustic panels.
• Record at 24-bit with peaks between −12 and −6 dBFS. Leave headroom and never clip.
• Track flat. Compression and EQ printed on the way in cannot be undone later.

Then comp: pick the best take line by line, trim the gaps, fade the edges and tidy obvious clicks. Tune gently if the song needs it. Finally, deal with breaths — lower the loud ones with clip gain rather than deleting every single one. A vocal with no breaths at all sounds synthetic and strangely airless.

EQ: cut before you boost

Start with a high-pass

Most close-mic'd vocals carry rumble, desk thumps and plosive energy well below the fundamental. A high-pass filter at 80–100 Hz on a lead vocal clears that out without touching the body of the voice. Backing vocals can go higher, around 120–150 Hz, which helps them stack without turning to mud.

Find the boxiness

Sweep a narrow bell boost of about 6 dB between 150 and 400 Hz until the voice sounds hollow and cardboardy, then flip that boost into a cut of 2–4 dB. That is your mud band. Do the same around 800 Hz–1.5 kHz if the vocal sounds nasal, and near 3 kHz if it feels pinched.

Add presence carefully

A gentle shelf or broad bell at 3–6 kHz brings consonants and intelligibility forward. Two or three dB is usually plenty. Above that you are in harshness territory — 6–9 kHz is where bright condensers, cheap preamps and tired singers all show up.

Compression: keep the level steady

Set ratio 3:1 or 4:1, attack 10–20 ms so the first consonant survives, and release 60–120 ms so the gain recovers between phrases. Aim for 3–6 dB of gain reduction on the loudest lines, not constant movement. Then raise the make-up gain so the compressed and uncompressed versions match in loudness — it is the only fair way to judge whether it actually sounds better.

Ride the fader first

If one word is jumping out, fix it with clip gain or volume automation before the compressor sees it. Compressors smooth performances; they do not repair a bad balance.

Two gentle stages beat one heavy one

A slow 2:1 compressor for overall level, followed by a faster 4:1 for control, sounds more natural than a single unit doing 10 dB of work. Parallel compression — a heavily squashed copy blended under the dry vocal — adds weight without flattening dynamics.

If you can hear the compressor working, it is probably working too hard.

De-essing: tame the S sounds

Sibilance lives between roughly 5 and 9 kHz, though a harsh "sh" can sit nearer 3–4 kHz on some voices. Cheaper microphones and bright interfaces exaggerate the problem, so it is worth fixing rather than ignoring.

Use a de-esser to reduce 3–6 dB only on the offending syllables. If it is dulling every S in the song, you have gone too far. A dynamic EQ set to a narrow band works just as well and gives you finer control over frequency and width. For stubborn cases, clip-gain the individual S sounds by hand — slower, but completely transparent.

Order matters: de-ess before the compressor so the compressor is not reacting to sibilance, then check after it too. Gain reduction can exaggerate the very frequencies you just tamed.

Reverb and delay: create a space, then step back

Send, don't insert

Put the reverb on an aux or FX return and send the vocal to it. One reverb then serves the whole mix, and you can treat the return without touching the dry voice.

Use pre-delay

20–40 ms of pre-delay keeps the words clear before the tail arrives. On a plate or small room, a decay of 1.2–1.8 seconds is plenty for a pop or rock vocal.

Filter the return

High-pass the reverb at 200–300 Hz and low-pass it around 6–8 kHz. Reverb below 200 Hz muddies the low end; reverb above 8 kHz adds hiss and sibilance.

For depth rather than wash, try a filtered slap delay at 80–120 ms, panned slightly opposite the vocal and sitting 15–20 dB below it. Keep the dry vocal loud and upfront. Effects should be felt more than heard.

Check the mix before you call it done

Listen in mono — if the vocal disappears, something is out of phase or over-widened. Try a phone speaker, laptop speakers and the car; harshness and sibilance show up instantly on small drivers. Level-match against a reference track you know well, because louder always sounds better and will fool you every time. Low-volume checks reveal whether the vocal still leads the mix when the drums and bass are barely audible.

The order of work that keeps you sane

1. Clean and comp the vocal; clip-gain the loud breaths.
2. Corrective EQ: high-pass, mud cut, nasal cut.
3. De-ess before compression.
4. Compress in one or two gentle stages, then match levels.
5. Tonal EQ and light saturation for colour.
6. Reverb and delay on sends, both filtered.
7. Automate the vocal line by line so every word is audible.
8. Check in mono, on small speakers, and at low volume.

Save the chain as a template once it works — you will start every future vocal in a better place. Resist doing everything at once: make one change, listen, decide. If something sounds wrong and you cannot place it, sleep on it and come back with fresh ears. Half of mixing is knowing when to stop.

Photo: Reza Tavakoli / Pexels