Mixing Vocals

The Vocal Processing Chain

A common vocal chain, in order:

1. High-pass filter at 80–120 Hz (remove rumble and proximity effect)
2. Subtractive EQ for problem frequencies (boxy 300 Hz, nasal 1 kHz, harsh 3–5 kHz)
3. First compressor — fast, catching peaks (1176-style, 3–4 dB reduction)
4. Second compressor — slow, evening dynamics (LA-2A-style, 2–3 dB reduction)
5. De-esser targeting 5–8 kHz
6. Additive EQ for presence (shelf boost at 8–12 kHz for air, gentle 3 kHz boost for clarity)

Not every vocal needs all of these. Some vocals need more; some need less. A well-recorded vocal with consistent dynamics might skip the second compressor. A harsh recording might need the subtractive EQ to do more heavy lifting. The best approach is to use only what’s needed — if a plugin isn’t solving a problem, take it off. Listen and decide.

One thing worth considering: breaths. Some engineers remove all breaths from a vocal. Others leave them all in. Both are wrong as defaults. Breaths are part of the performance — they tell you when the singer is working hard, when they’re building to something, when they’re out of air. Removing all of them makes the vocal feel robotic. Leaving aggressive breaths in can be distracting. Listen to each one. Turn down the ones that stick out, leave the ones that serve the performance.

Vocal Level

The vocal level is the hardest thing to get right. It should sit on top of the mix without sitting above it — present, clear, and connected to the instruments.

Automate the vocal fader to ride phrases: pull it down slightly on loud sections, push it up on quiet moments. This is tedious but irreplaceable — it’s the difference between a vocal that floats and one that fights.

The best vocal rides follow the lyric sheet. Read the lyrics. Know which words carry the meaning of each line. Those words need to be heard. Other words can sit slightly lower. This is prosody — the relationship between the lyrics and the delivery — and respecting it in your level automation is what makes a vocal feel intentional rather than mechanical.

Prosody and the Lyric Sheet

Keep a lyric sheet open while you mix vocals. Read along. The lyrics tell you where the emotional peaks are, where the payoff lines land, where a quiet moment needs protection. A mix that serves the lyrics — that brings up the word “forgive” in the bridge, that lets the final line of the chorus ring — is a mix that serves the song.

If you don’t have a lyric sheet, make one. The act of transcribing the lyrics forces you to listen to the performance at the word level, which is where vocal mixing happens.

Dealing with Problem Vocals

Harsh Vocals

Harshness in the 3–5 kHz range. Use a dynamic EQ here rather than a static cut — the harshness may only appear on certain syllables or louder phrases. A static cut dullifies the entire vocal. A dynamic EQ only cuts when the frequency exceeds the threshold.

Muddy or Boxy Vocals

Too much 200–400 Hz. The proximity effect from close-miking adds bass boom that doesn’t belong. A high-pass at 100–120 Hz helps, but sometimes the boxiness is above the HPF range. A broad cut at 300 Hz cleans it up.

Thin Vocals

Missing low-mid warmth. A gentle boost at 200–300 Hz adds body. Or use a tube or tape saturation plugin to add harmonics that fill out the low end.

Inconsistent Dynamics

If compression isn’t enough, use clip gain editing before the chain. Go through the vocal phrase by phrase and normalize the clip gain so loud words and quiet words arrive at the compressor at roughly the same level. The compressor then works on a more consistent signal and sounds more transparent.

De-essing

Sibilance gets worse after compression. The compressor reduces the dynamic range, which brings the sibilant “s” sounds closer to the level of everything else — they become proportionally louder. Some engineers place the de-esser before the compressor — taming the sibilance before the compressor grabs onto it — while others place it after, cleaning up what compression brought forward. The chain above puts it after compression, but try both on the material you’re working with.

Listen for the threshold carefully. Too much de-essing and the singer lisps. Too little and the sibilance pierces through the mix. The target frequency is usually between 5–8 kHz, but it varies by singer. Sweep to find it.

Some engineers prefer manual de-essing: identifying each sibilant moment and reducing its gain with clip gain or automation. It’s slower but gives complete control.

Backing Vocals

Pan backing vocals away from center to create width. Process them differently from the lead:

• More reverb — push them further back in the mix
• Less presence — cut 3–5 kHz so they don’t compete with the lead
• Slightly quieter — they support, not compete
• More compressed — tighter dynamic range so they sit consistently behind the lead

Group them to a bus for shared processing and single-fader control. A bus compressor glues multiple backing vocal tracks into a cohesive bed.

Try the exercise from the stereo chapter: take a stack of backing vocals with the root, third, and fifth doubled. Pan each interval pair (root pair at narrow width, third pair wider, fifth pair widest) versus grouping each interval together in one spot. The results are dramatically different — and neither is wrong.

Effects on Vocals

• Reverb: Use sparingly on the lead vocal. A little plate adds polish. Too much pushes the vocal away from the listener. Use pre-delay (15–30ms) to keep the vocal’s consonants clear.
• Delay: A tempo-synced delay (quarter or dotted-eighth) fills gaps between phrases. Automate the send — more on held notes, less on fast passages.
• Saturation: A touch of tube saturation adds presence and warmth. Use subtly.
• Chorus: On backing vocals for width. Rarely on lead vocals (it smears the focus).

What to Practice

• Build a vocal chain: HPF → subtractive EQ → fast compressor → slow compressor → de-esser → additive EQ. Process each stage conservatively (2–3 dB each). A/B the chain on and off at matched loudness.
• Automate the vocal fader phrase by phrase through an entire song. Follow the lyric sheet. Bring up quiet words, pull down loud passages. This takes time — that’s the point.
• De-ess a vocal track. Push the threshold too far (the vocal will lisp), then back off until sibilance is controlled but the consonants still sound natural.
• Process backing vocals separately from the lead: add more reverb, cut the presence range, compress harder. A/B with unprocessed backing vocals and listen to how the processed version sits behind the lead.