Vocal Mixing Signal Chain: EQ, De-essing, Compression and Space

In most songs the vocal is the first thing a listener follows and the last thing they forgive. A beat can be a little muddy and nobody notices. A vocal that's buried, harsh or jumping in level is noticed immediately.

This is the chain I start from on lead vocals in pop, hip-hop, R&B and rock. Adjust the order and amounts to taste, and skip any step the vocal doesn't need.

Before any plugins: editing

A lot of good vocal mixing happens before the first plugin:

• Comp the best takes into one performance.
• Clean up noises between phrases. Lower loud breaths by 6–10 dB instead of deleting them all; a vocal with no breaths sounds unnatural.
• Fix timing on words that drag or rush noticeably.
• Use clip gain to even out phrases that are much louder or quieter than the rest. This makes every compressor after it work better.
• Tune lightly if needed, unless an obvious tuned sound is part of the style.

Subtractive EQ

Start with a high-pass filter at 70–100 Hz (12 dB/oct) to remove rumble and handling noise. For deep voices stay lower; for a thin backing vocal in a busy mix you can go higher.

Cut only where you hear a problem. Every voice and microphone is different, and the table is a map, not a prescription.

Serial compression: fast FET into smooth opto

Vocals have a wide dynamic range, and one compressor trying to control all of it usually sounds squashed. The classic studio approach, an 1176-style FET compressor feeding an LA-2A-style optical compressor, splits the job into two stages:

• Stage 1, FET (peak control): ratio 4:1, a fast attack (in 1176 terms, roughly the middle of the attack range, still well under 1 ms), fast-to-medium release around 50–100 ms. It catches only the loudest syllables, 2–4 dB of gain reduction on peaks, adding a bit of energy.
• Stage 2, Opto (leveling): program-dependent attack around 10 ms and a two-stage release. Set it for 2–3 dB of steady reduction. It smooths the overall level without obvious grabbing.

Any FET-style and optical-style plugin works for this, including stock DAW compressors set to similar behavior. The compression guide covers the topologies in more detail.

De-essing sibilance at 5–8 kHz

Compression brings up the quieter parts of a vocal, and that includes "s", "sh", "ch" and "t" sounds. A de-esser reduces those frequencies only when they appear. Most sibilance sits between 5 and 8 kHz, a little lower for deep male voices and a little higher for bright female voices.

Find the frequency by listening to the de-esser's monitor or "listen" mode, then set the threshold for 3–6 dB of reduction on the worst "s" sounds. Don't overdo it: too much de-essing makes the singer sound like they have a lisp. You want the "s" sounds softer, not gone. In bright mixes, a second, lighter de-esser on the reverb send helps a lot, because reverb exaggerates sibilance.

Tone EQ: presence and air

After compression and de-essing, shape the vocal so it sits in the mix:

• A gentle boost around 3–5 kHz for presence, if the vocal is getting lost.
• A high shelf from around 10–12 kHz, +1 to +3 dB, for air.
• A small boost around 150–250 Hz for warmth on thin voices.

Boosting after dynamics control means you're brightening a controlled signal, which sounds smoother than brightening the raw vocal.

Saturation

A little saturation adds harmonics that help a vocal cut through on small speakers without simply making it louder or brighter. Tape and tube-style saturation tend to sound warm; transistor or console-style saturation sounds a bit edgier. Keep the mix knob low. If you can clearly hear the distortion, it's probably too much, unless that's the sound you're after.

Spatial depth: plate reverb and slap delay

Put effects on aux sends rather than directly on the vocal. You can then EQ and de-ess the effects separately and share them with backing vocals.

• Plate reverb: bright, dense and flattering on vocals. Decay 1.2–2.0 s for most pop; shorter for fast rap. A pre-delay of 20–60 ms separates the dry voice from the reverb, keeping the vocal upfront while the space blooms behind it.
• Slap delay: a single repeat at 80–120 ms, little or no feedback, filtered (HPF 300 Hz, LPF 5 kHz). It thickens the vocal and adds depth without sounding like an obvious echo.
• Tempo delay: a 1/4 or 1/8 note delay low in the mix, or automated as a "throw" on the last word of a line.
• EQ the returns: high-pass reverb and delay around 200–300 Hz and low-pass around 6–8 kHz.

The delay and reverb calculator gives tempo-synced times and pre-delay values for any BPM.

Automation: the final 10%

Even after two compressors, some words still get lost and others poke out. Riding the vocal level with automation, sometimes word by word, is what makes a vocal sound finished. Lower your monitoring level and listen: at low volume, any word that disappears is obvious.

Backing vocals

Backing vocals should support the lead, not compete with it. They usually take more high-pass filtering (often up to 150–250 Hz), heavier compression for a consistent pad, stronger de-essing because stacked sibilance adds up quickly, wider panning and more reverb than the lead, and a lower level than you first think.

Common vocal mixing mistakes

• Making the vocal too bright to help it stand out, which just makes it harsh.
• Too much reverb, which pushes the vocal back instead of making it sound bigger.
• Judging the level only at loud volume.
• Skipping the editing stage and expecting compression to fix wildly uneven phrases.
• Never checking the vocal in mono, where wide doubles and effects can collapse.