Why Pro Vocals Sound Expensive (7 Small Moves)
Vocal mixing · 4 min read
Expensive vocals are not one plugin. They are about seven decisions of roughly 1 dB each — a level ride, one narrow cut, a touch of drive, a short predelay — all made at matched volume. Alone each one is inaudible. Together they are the entire difference.
The real formula
7 x 1 dB
Small moves · level-matched · in the right order
The 60-second answer
Ride the clip gain into a 3 dB window. Two compressors at 3 dB each with different attack times. One narrow subtractive cut. 3 dB of drive. De-ess 2–3 dB, then a 1–1.5 dB shelf at 10 kHz. A 1.2 s plate at 12–18% with 25–40 ms predelay. Then 0.5 dB fader moves per phrase.
Why one plugin never gets you there
Every stage has a ceiling before it starts announcing itself. A compressor is invisible up to about 3–4 dB of gain reduction and an effect beyond it. An EQ boost is a tone shape at 2 dB and a colour at 6. Saturation is warmth at 3 dB of drive and distortion at 10. Chase the whole result from one stage and you have to run it past its ceiling, and the listener stops hearing the voice and starts hearing the tool.
Why level matching is the actual skill
Turn anything up 1 dB and almost everybody calls it better. So every plugin that adds level wins its own A/B test whether or not it improved anything, and that is how cluttered chains get built, one convincing comparison at a time. Match the bypassed and processed level to within about 0.2 dB and the bias disappears. You will then reject about half the moves you would otherwise have kept.
The seven moves, in order
| # | Do this |
|---|---|
| 01 |
Ride the clip gain into a 3 dB window Before any plugin, pull the loud words down and push the quiet ones up until the waveform stops jumping. Target: +/-3 dB window |
| 02 |
Two compressors, 3 dB each A slower one first for levelling (attack near 10 ms), a faster one after for peaks (under 1 ms). Different attack times, or it is one compressor. 3 + 3 dB GR |
| 03 |
One narrow subtractive cut Boost 6 dB at Q 6, sweep 200 Hz to 5 kHz until something screams, then cut 2–4 dB there. -2 to -4 dB · Q 4–8 |
| 04 |
3 dB of drive Tube or tape model, mix 30–50%. Harmonics read as body where an EQ boost would read as a boost. Drive 3 dB · mix 30–50% |
| 05 |
De-ess, then add air 2–3 dB of de-essing at the frequency you swept out, then a 1–1.5 dB shelf from 10–12 kHz. That order, always. De-ess → +1.5 dB @ 10 kHz |
| 06 |
One space, filtered and short A 1.2 s plate at 12–18% send with 25–40 ms predelay, return high-passed at 300 Hz. Add a dotted-eighth delay at 10–15%. Plate 1.2 s · predelay 25–40 ms |
| 07 |
Automate 0.5 dB per phrase Per phrase, not per section, and on the sends as well as the fader so the space grows with the performance. 0.5 dB · fader + sends |
The trick: match the level or do not compare at all
Here is the habit that changes everything, and it costs nothing. Put a simple gain utility directly after every processor in the chain. Bypass the processor, watch the short-term loudness meter over one full chorus and note the number. Un-bypass it and trim the utility until the two readings match within about 0.2 units. Now flip a few times. If you cannot say instantly which one you prefer, the move did nothing and it should come out. That is why an engineer with three stock plugins beats a beginner with four hundred: not better moves, far fewer bad ones. Add one more rule — never approve a chain the day you built it — and you will delete two more plugins in the morning.
Stop guessing the chain.
The Vocal Chain Bible is 88 real vocal chains from Grammy-winning engineers — every plugin, every setting, in order, with free alternatives for each one. 5+ years, 100+ interviews, one PDF.
Get the Vocal Chain Bible — $48 →Mistakes that kill it
- Chasing a preset. A chain from someone else's session was built for their voice at their input level. The numbers mean nothing on your track.
- Making the biggest move first. A 6 dB boost at the front forces every later stage to work around it. Small moves, in order.
- No level matching. Every uncompensated comparison is a loudness test, and loudness always wins. You are not choosing, you are being fooled.
- Judging at one volume. Check quiet, check loud, check on a phone. Balance decisions move with playback level more than any plugin does.
Nothing here is exciting. That is the point: expensive is what boring, correct, level-matched decisions sound like.
Keep going
FAQ
What makes vocals sound expensive?
A stack of small correct moves: stable level, two gentle compressors, one narrow cut, a little saturation, controlled esses, one short space, detailed automation.
Do I need expensive plugins for an expensive vocal sound?
No. The order of the moves and the size of each one matter far more than the brand. A stock EQ used well beats a boutique one at 6 dB.
How do I level match when comparing plugins?
Put a gain utility after the plugin and trim until bypassed and processed read the same short-term loudness, within about 0.2 units.
Who's behind this
Ivan Meshcheriakov · IvanFromRnD
5+ years ago I moved to LA with zero connections and started building from nothing — 14-hour days, meeting people and interning at studios. It took me 5+ years to meet and interview 100+ Grammy-winning engineers — then I turned it all into the Vocal Chain Bible.
Read my full story →