A vocal chain works best in a fixed order: fix, then control, then colour, then place. That means cleanup and EQ cuts first, two stages of compression next, de-essing and saturation after that, and reverb and delay last, on sends. Reversing that order means you spend the whole mix amplifying problems you could have removed in ten seconds.
Before any processing, do the unglamorous work:
- High-pass at 80–100 Hz for most voices, 60–70 Hz for deep male voices. Set it by ear — sweep up until the vocal thins, then back off.
- Remove breaths and noise between phrases, or reduce them by 6–10 dB rather than deleting them. Fully silent gaps sound unnatural.
- Fix obvious pitch and timing issues at the source. Correcting a badly flat note after compression means the compressor was reacting to the wrong signal.
- Ride the fader or use clip gain to level out the loudest and quietest words before compression. This is the single biggest improvement most vocal mixes get, and it costs no plugin.
Cut before you boost.
| Problem | Range | Move |
|---|
| Rumble, proximity boom | 80–150 Hz | High-pass or shelf down 2–4 dB |
| Boxy, cardboard | 250–450 Hz | Narrow cut 2–4 dB |
| Honky, nasal | 800 Hz–1.2 kHz | Narrow cut 2–3 dB |
| Harsh, edgy | 2.5–4 kHz | Narrow cut 2–3 dB if needed |
| Sibilance | 5–9 kHz | Handle with a de-esser, not static EQ |
Find problem frequencies by boosting a narrow band 6 dB and sweeping until it sounds worst, then cut at that point by 2–4 dB with a Q of about 2–4.
Compressor 1, for level. Ratio 2:1 to 3:1, attack 10–30 ms, release 100–200 ms or auto, threshold set for 3–4 dB of gain reduction on average words. A slow optical-style compressor is ideal here. This evens out the performance.
Compressor 2, for character. Ratio 4:1, attack 3–10 ms, release fast (30–80 ms), another 3–4 dB of gain reduction. A faster VCA or FET-style compressor adds density and forwardness.
Two stages of 3–4 dB sound dramatically better than one stage of 8 dB. If the vocal still moves around after both, the problem is clip gain, not compression.
Put the de-esser after compression, because compression makes sibilance worse — the fast S sounds survive while the vowels get pulled down.
Target the range you identified above, aim for 3–6 dB of reduction on the worst S sounds only, and check that the vocal does not develop a lisp. If a single word is far worse than the rest, automate a clip gain reduction on that word instead of turning the de-esser up for the whole track.
Light saturation adds harmonics that let a vocal sit above a dense mix without extra level. Tape or tube-style saturation at a drive setting where you can only just hear the difference is usually right — if it sounds obviously distorted soloed, it is too much in context.
Then tonal EQ, the boosts:
- 150–250 Hz, a broad 1–2 dB lift for warmth if the voice feels thin.
- 3–5 kHz, 1–3 dB for presence and intelligibility.
- 10–14 kHz, a gentle shelf for air. A dynamic or shelving move here is safer than a static boost that also lifts sibilance.
Reverb and delay go on sends, not inserts. A plate with 20–40 ms pre-delay and a 1.2–1.8 s decay is a dependable pop vocal reverb. High-pass the return at 300 Hz.
Add a delay only if reverb alone leaves the vocal feeling short — an eighth or dotted-eighth delay filtered to 300 Hz–5 kHz, ducked under the dry vocal, adds depth without wash.
Automation. A vocal that has been ridden line by line, with pushes on the ends of phrases and pull-downs on plosives, beats any plugin chain. Do the compression to make the vocal consistent, then automate to make it musical.
Related reading: how to clean up a vocal recording, recording vocals at home, and reverb and delay basics.
Start making music in Veena
Free, browser-based, no downloads required.
Try Veena Free