The Hall of Records × The Craft Desk · The Protocols Editor's Pick
The Protocols: Ducked Space — Reverbs and Delay Throws That Don't Bury the Vocal
Modern records sound dry and enormous at the same time. That's not a reverb choice — it's a timing system: space that clamps down while the singer sings and blooms in the gaps. Here's the full mechanism, both methods, and the trick that makes it invisible.
Here is the paradox every producer hears and few can build: on a modern record the lead vocal sounds completely dry and hyper-present — every consonant legible, right at the front of the speaker — and yet the track around it sounds cinematic and enormous, full of space. Both are true simultaneously, and it is not a compromise reverb setting. It’s a system: the space is real and generous, and it gets ducked out of the way while the voice is working, returning to full bloom the instant a phrase ends.
Once you hear it, you can’t unhear it. Once you build it, you won’t mix without it.
The core mechanism
Send the lead vocal to a fully wet effect return — reverb, delay, whatever the record needs. Then, after that effect on the same return, insert a compressor whose sidechain listens to the clean, dry lead vocal. Now the effect return is being controlled by the voice itself: the singer sings, the tail clamps down; the phrase ends, the sidechain releases, the tail blooms into the gap.
The parameters that matter, and why:
- Attack: very fast. The tail must get out of the way at the syllable’s arrival, not after it. A slow attack lets the reverb collide with the very consonant it’s supposed to protect.
- Release: musical. Roughly a note value at the song’s tempo — fast enough that space returns inside short rests, slow enough that it doesn’t pump mid-word. Set release by listening to the gaps, which is the only place the effect is audible anyway.
- Ratio: firm. Enough to hold the tail down decisively; this is a dedicated control circuit, not a glue compressor.
- Sidechain high-pass. Filter the low end out of the key signal so plosives and body don’t trigger ducking that has nothing to do with intelligibility.
- Depth: a few decibels, not a mute. The goal is clearance, not disappearance. If you can hear the ducking as an effect, it’s too deep.
The upgrade: duck frequencies, not the whole tail
Broadband ducking works — but it lowers everything, so the reverb’s air and low warmth pump along with the mids. The modern refinement is a dynamic EQ on the return, keyed to the lead vocal, cutting only where intelligibility lives:
- A moderate dynamic cut in the low-mids, clearing chest-resonance mud from the tail.
- A deeper dynamic cut in the presence range, the band where consonants and vocal clarity live.
- Everything else untouched — the top-end shimmer stays open, the low warmth stays intact.
The result is a space that never appears to move: the tail keeps its size and color, and simply steps out of the frequencies the voice needs while the voice is using them. This is the single biggest upgrade available to most producers’ vocal spaces, and it costs nothing but routing.
The trick that makes it invisible
Duplicate the lead vocal, mute its output, and slide the copy slightly earlier on the timeline. Use that silent, advanced copy as the sidechain key. Now the ducking begins fractionally before the vocal transient actually arrives — so the space is already out of the way when the syllable lands, and the initial clash never happens. (Where a compressor offers lookahead, that’s the same idea implemented internally.) It is the difference between ducking you can hear and ducking you can only stop hearing when you bypass it.
Delay throws: the arrangement move, not the effect
The same architecture powers the other half of modern vocal space — the throw, where one word repeats into the distance while the rest stays dry.
Two ways to capture it. Automate the send: keep the throw send muted for the whole song and open it only on the target word, then close it. Or dedicate a track: copy the target word onto its own “throw” track routed fully wet to the delay return — visual, editable, and easier to manage in a dense session, which is why most assistants prefer it.
Then treat the throw like its own instrument:
- Hard-tune the repeats so drifting vibrato in the tail doesn’t smear against the track’s harmony.
- Pitch or formant-shift for character — an octave-up shimmer sitting well below the lead, or a darker shadow underneath.
- Micro-detune left and right for width, leaving the dead center open for the lead vocal.
- Band-limit hard — high-pass and low-pass the throw into a narrow midrange so it never competes with the lead’s clarity or the low end’s weight.
- Duck it to the lead, exactly as above, so it swells only in the gaps.
- Smear it with a short plate at the end of the chain, turning discrete repeats into an atmospheric cloud.
What to steal: build one ducked reverb return and one ducked throw bus, save them in your template, and stop rebuilding them. The professional advantage here isn’t knowledge of the trick — it’s that the trick is already routed when inspiration shows up.
The honest counterpoint
Ducked space is an aesthetic with an era attached. It exists to serve maximum-intelligibility modern production, and it can absolutely be wrong: a live-feel record, a jazz vocal in a real room, an ambient piece where the voice is supposed to dissolve into the space — all of these are damaged by a system engineered to keep the voice in front. And there’s a subtler cost worth naming: ducking lets you use far more reverb than a record actually needs, because the mechanism hides the excess while the voice sings. If your mix collapses into wash the moment you bypass the sidechain, the space isn’t ducked — it’s dependent. Build it so the record would still work with the trick turned off, then turn it on and let it be the polish it was meant to be.
The Protocols: Max Martin’s Vocal Architecture → · The Craft Desk: Your First Vocal Chain → · Mixing Yourself Without Lying to Yourself →