Music Production Guide

Sound Recipe - UT Twin87 Vocal Recording

Target

A clean, expensive, "finished-sounding" lead vocal captured with the United Studio Technologies UT Twin87 (a twin-circuit U87-style condenser) into the UAD Apollo Twin. The goal is a detailed, present, but never harsh vocal that drops into a mix and sits like a record.

The capture should feel:

  • Detailed and airy (condenser clarity)
  • Present and intimate, not boxy
  • Controlled in the sibilance and lows
  • Big enough to need almost no "fixing" later

Follows the house method: intention → source (mic + room + preamp) → capture settings → plugin start → routing → settings → automation → taste checks.

The Mic: UT Twin87 Modes and Patterns

The Twin87 reconfigures its circuit on the fly. Pick deliberately:

Setting Use Going For
Vintage mode Brighter/sharper voices, sibilant singers Softer top, rounder U87-classic warmth
Modern mode Dark/dull voices, when you want air Extra shine, articulation, "expensive" sheen
Cardioid (true cardioid) Close lead vocals Best rejection + signal-to-noise; the default
Omni Treated room, doubling the room as part of the sound Natural, open, less proximity
Figure-8 Two singers face-to-face, or creative null-pointing at a noise source Stereo/duet tricks, killing a reflection
-10 dB pad Loud belting Avoid front-end overload
80 Hz HPF Close singing with heavy proximity/rumble Tame boom without an EQ later

Start: Cardioid + Modern mode for most pop vocals; switch to Vintage if the voice is harsh.

Setup and Capture

  1. Placement: mic at mouth height or slightly above, angled down at the mouth (reduces plosives and nasal honk). Distance 6–9 inches for a controlled, pro lead; closer = warmer + more proximity bass; further = more room.
  2. Pop filter: always, ~3–4 inches off the grille. The Twin87 has no built-in windscreen.
  3. Off-axis trick: have the singer aim just past the mic (sing slightly across it) to soften plosives and sibilance.
  4. Room: condensers hear everything. If the room is untreated, get closer, use cardioid, and put something absorbent behind the singer (or face them into a closet of soft clothes). If the room is nice, try Omni 12–18 inches back.
  5. Gain (Apollo Twin): set Unison preamp so peaks land around -12 to -6 dBFS. Use the mic's pad before pulling preamp gain too low.

Apollo Twin advantage

  • Load a UAD Unison preamp (e.g., UA 610, Neve 1073, or API Vision) on the input — it tracks the analog-modeled preamp into the recording.
  • Optionally print light compression on the way in (UAD 1176 ~2–3 dB GR) for a committed, finished feel. When unsure, track clean and add it later.
  • Use low-latency monitoring with reverb on the singer's headphone cue (a plate they like) so they perform better — but record dry.

Fast Path (capture + quick mix)

  1. Twin87 Cardioid + Modern, 7 inches, pop filter, HPF on
  2. Apollo Unison: UA 610 preamp, peaks ~-9 dBFS
  3. Insert: Pro-Q 4 (HPF 80–100 Hz, gentle 3–5 kHz presence), UAD 1176 (4:1, 4–6 dB GR)
  4. Send to a plate (UAD EMT 140) and a 1/4-note delay
  5. One pass of automation to ride the level

Produced Path (the full chain)

Capture clean, then build the channel:

  1. iZotope RX 12 (offline) — De-plosive, Mouth De-click, Breath Control, De-ess if needed, light De-noise for room
  2. Antares Auto-Tune Pro — transparent (fast retune only on problem notes) or hard-tuned as an effect
  3. FabFilter Pro-Q 4 — HPF 80–100 Hz; subtractive dips (boxiness 300–500 Hz, harsh 2–4 kHz); gentle air shelf above 10 kHz if Vintage mode is too soft
  4. UAD 1176 — fast FET catch: 4:1, fast attack, fast release, 3–6 dB GR (attitude/presence)
  5. UAD Teletronix LA-2A — in series after the 1176 for smooth leveling, 2–4 dB GR (the classic 1176→LA-2A vocal stack)
  6. RX De-ess / FabFilter Pro-DS-style move or Logic DeEsser 2 — only if sibilance remains
  7. Sends: plate (UAD EMT 140 / Eventide SP2016), room/hall (Lexicon PCM), and a tempo delay (Eventide H3000 / UltraTap). Filter returns darker than instinct.

Expensive one-box alternative: track/mix through UAD Manley VOXBOX or Avalon VT-737 (pre + EQ + comp + de-ess) instead of steps 3–6.

Routing

  • Vocal → insert chain (EQ → comp → comp → de-ess) → Vocal Bus
  • Vocal Bus → mix bus; sends to shared reverb/delay returns (never insert big reverb)
  • Keep a parallel "smash" duplicate (1176 all-buttons) blended low for density when needed

Automation

  • Ride the fader (or use sonible pureLevel) so every word reads
  • Open reverb/delay sends into the chorus; pull them back under busy verses
  • Brighten/De-ess dynamically rather than statically when only a few words are harsh

Adjustment Rules

Too harsh/sibilant: switch to Vintage mode, back off distance, sing more across the mic, then de-ess.

Too dull: Modern mode, add air shelf, get a touch closer.

Too boomy: HPF on the mic, more distance, less proximity, high-pass in Pro-Q.

Too roomy/boxy: get closer, cardioid, treat behind the singer; the room — not the mic — is usually the problem.

Sounds amateur: committed gain staging, the 1176→LA-2A stack, one shared reverb world, and disciplined de-essing fix 90% of it.

Common Mistakes

  • No pop filter (condensers are plosive-prone)
  • Recording in an untreated, reflective room far from the mic
  • Tracking too hot (clipping the Apollo front end)
  • Static de-essing that lisps the whole vocal
  • Inserting big reverb instead of using sends
  • Forgetting to pick a mode/pattern on purpose

Closest Tools I Own

Mic/pre: UT Twin87, UAD Apollo Twin (Unison: UA 610 / Neve 1073 / API Vision) Repair/tune: iZotope RX 12, Antares Auto-Tune Pro Dynamics: UAD 1176, Teletronix LA-2A, Manley VOXBOX, Avalon VT-737 EQ/de-ess: FabFilter Pro-Q 4, Logic DeEsser 2 Space: UAD EMT 140 / Capitol Chambers, Lexicon PCM, Eventide SP2016 / H3000 / UltraTap Level: sonible pureLevel

Practical Summary

Pick the Twin87's mode and pattern on purpose (Cardioid + Modern to start, Vintage to tame harshness), record 6–9 inches back with a pop filter through an Apollo Unison preamp at sane levels, and capture clean. Then build the channel with RX cleanup, transparent tuning, the 1176→LA-2A stack, surgical de-essing, and one shared reverb world. The mic is excellent — most of "expensive" comes from placement, the room, and disciplined gain staging.