Processing turns a recorded vocal into a mix-ready vocal: clear, consistent, present and polished. Every engineer's chain is a little different, but almost all use the same core tools in a similar order.
Clean-up first
Before creative processing, fix problems:
- High-pass filter to remove rumble and low-frequency noise. Often somewhere between 70 and 120 Hz for a lead, higher for backing vocals. Listen as you raise it: stop before the voice thins out.
- Noise and breath control - see the editing chapter.
- Plosives that escaped the pop filter.
EQ
Subtractive EQ
Cut problems before boosting anything:
- Mud - a build-up around 200-500 Hz can make vocals sound boxy or muffled. A gentle cut here is common.
- Harshness - a sharp resonance somewhere around 2-5 kHz can make a voice fatiguing. Find it with a narrow boost sweep, then cut.
- Room resonances - narrow cuts to tame ringing frequencies from the recording space.
Additive EQ
Then enhance:
- Body - a gentle boost around 150-250 Hz for warmth, if the voice needs it.
- Presence - a boost somewhere around 3-6 kHz brings the vocal forward and improves intelligibility.
- Air - a high shelf from roughly 10-12 kHz adds sparkle and breathiness. High-quality EQs do this smoothly; cheap ones can sound brittle.
Every voice is different. Use these ranges as starting points, and make decisions with the vocal playing in the mix.
Compression
Vocals vary enormously in level - whispered lines, belted notes, quiet word endings. Compression evens them out so the vocal sits consistently in the mix.
- Ratio - around 3:1 to 4:1 is a common starting point for a lead.
- Threshold - set for a few decibels of gain reduction on average, more on peaks.
- Attack - medium (around 10-30 ms) lets the start of words through for clarity; faster attacks control more but can dull the vocal.
- Release - medium to fast, so the compressor recovers between words. Many compressors have an auto setting that works well on vocals.
- Make-up gain to restore level.
Serial compression
Instead of one compressor working hard, use two working gently: a fast one (often an FET-style compressor) to catch peaks, followed by a slower, smoother one (often an optical style) for overall levelling. Each does a few decibels of gain reduction, and the result sounds more natural than one compressor doing all the work.
Rides before compression
Automating the vocal level - riding the fader line by line - before it hits the compressor lets the compressor work gently and musically.
De-essing
Sibilance - harsh "s", "sh", "t" and "ch" sounds - is exaggerated by condenser microphones, presence boosts and compression. A de-esser is a compressor that only reacts to a narrow high-frequency band, typically somewhere between 4 and 10 kHz depending on the voice.
- Find the frequency - many de-essers have a listen mode to solo the band being detected.
- Set the threshold so it only acts on the sibilants.
- Do not overdo it. Too much de-essing makes a singer sound like they have a lisp.
- Place it after EQ and compression, or before and after if sibilance is severe.
For very sibilant vocals, manually turning down individual "s" sounds with clip gain is often cleaner than any plug-in.
Saturation
A little saturation - tape, valve or console emulation - adds harmonics that help a vocal cut through a dense mix and sound warmer and more expensive. Subtle is the word; you should feel it more than hear it.
A typical chain
- Clean-up (high-pass, noise, plosives)
- Subtractive EQ
- Compression (one or two stages)
- Additive EQ
- De-esser
- Saturation (optional)
- Sends to reverb and delay
The order is a starting point, not a rule. Try changes and listen.
Do this now
- Build the chain above on a lead vocal.
- Find and cut the harshest frequency with a narrow EQ sweep.
- Set up serial compression with two compressors, each doing 2-4 dB of gain reduction.
- Use a de-esser's listen mode to find the sibilance frequency, and set it to reduce only the worst "s" sounds.