Keep the lead vocal mono and centered; create width around it. Best: record real doubles and pan them hard left and right, 3–6 dB under the lead. Good: fake a double by duplicating the vocal, shifting one copy 10–20 ms and ±5–10 cents, panning the pair. Quick: a subtle chorus or a stereo delay. Always check the result in mono — width that vanishes (or comb-filters) on a phone speaker isn't width, it's a problem.
- The lead stays mono and centered — width comes from what surrounds it, not from smearing the lead itself.
- Nothing beats real double-tracks: two more performances panned hard L/R, tucked under the lead.
- Fake doubles work when they differ from the lead in timing (10–20 ms) and pitch (±5–10 cents) — difference is what reads as width.
- Chorus, stereo delay, and Utility width are the quick tools; use them in moderation on sends or doubles.
- Every widening trick must pass the mono test: sum the mix and make sure the vocal neither vanishes nor hollows out.
Width is contrast, not a knob
The wide vocal sound on big choruses is almost never a widened lead — it's a centered, mono lead with different signals on the left and right around it. Stereo width in hearing terms is inter-channel difference: when the left ear gets something meaningfully different from the right, the image spreads. That's why the techniques below are ranked by how much genuine difference they create — and why the verse/chorus contrast (narrow verse, wide chorus) does more for perceived width than any plugin setting.
Tier 1 — Real doubles (nothing fakes this)
Record the same part again — twice more, ideally — matching the lead's phrasing as tightly as possible. Pan one take hard left and one hard right, and sit the pair 3–6 dB under the lead. The micro-differences in timing, pitch, and timbre between human performances are exactly the inter-channel difference stereo width is made of, which is why doubles sound huge where processed copies sound processed.
- Tight doubles (pop, rap hooks): match phrasing hard, tuck them low — felt as size, not heard as voices.
- Loose doubles (indie, punk): let the timing breathe for a rawer, roomier width.
- Keep doubles darker and less sibilant than the lead — de-ess them harder and shave some air so the esses don't smear across the image.
- Whispered doubles under a sung lead add width plus an airy texture pop producers lean on constantly.
This is the strongest argument for recording doubles while the mic is still up: thirty extra minutes at the session buys a chorus no widener can build later.
Tier 2 — Faking a double from one take
No double takes? Manufacture the differences a real double would have. Duplicate the vocal track, then on the copy: nudge timing 10–20 ms, detune ±5–10 cents (the clip's Detune field, with Complex Pro warp, keeps formants natural at these tiny amounts), and pan the original and copy apart. Some mixers run the copy through a slightly different EQ or saturation for timbral difference too.
The offsets matter: under ~10 ms of delay, the copies comb-filter into a hollow mess when summed, and it's the pitch offset that keeps the pair decorrelated in mono — a purely delayed copy still cancels (that's the Haas trade-off below). You're recreating human inconsistency — err toward more difference, tucked lower.
Tier 3 — One-device width: chorus, delay, Utility
- Chorus-Ensemble: a subtle setting (low rate, low amount, blended under 30% wet) modulates the two channels differently — instant, gentle spread. The Ensemble mode is thicker; watch that it doesn't wobble the pitch audibly.
- Stereo delay: the dotted-eighth-left / quarter-right setup widens by putting different rhythms in each ear — width and depth in one move.
- Utility Width above 100%: boosts the existing side signal. It can only exaggerate stereo content that's already there, and everything it adds disappears in mono — use it on already-stereo doubles or effect returns, never as the lead's only width.
- Haas panning (one copy 10–30 ms later on one side) sounds dramatic in headphones and collapses into comb-filtering in mono — know the trade before using it.
The mono test (non-negotiable)
Phone speakers, smart speakers, club PAs, and TikTok playback are mono or close to it. Drop a Utility set to Mono on your master, and listen: the vocal should stay full and centered, with the mix merely narrower. If the vocal gets quieter, hollow, or phasey, a widening trick is fighting you — usually Haas timing or aggressive width boosting. Real doubles and modulated width survive the sum; pure-difference tricks don't.
Width is also a measurable property: PhantomRack's analysis fingerprints a reference's stereo correlation and applies matched width settings in its generated racks — but real doubles are a performance, not a parameter, so when a reference is clearly stacked with them, the honest answer is the mic, and the rack handles making them sit. The full stock-device toolkit is covered here.
Frequently asked questions
How do I make my vocals sound wider without doubling?
Duplicate the take and manufacture a double: offset the copy 10–20 ms, detune it ±5–10 cents, pan the pair apart, and tuck the copy under the lead. Or use subtle chorus and a stereo delay send. All of these read as width while keeping the lead itself centered.
Should lead vocals be stereo or mono?
Mono and centered, almost always. A smeared stereo lead loses focus and risks collapsing in mono playback. Width belongs to the doubles, harmonies, and effects around the lead — that contrast is what makes the vocal feel big and stay intelligible.
Why does my widened vocal disappear on phone speakers?
Phone playback is mono, and mono playback sums left and right. Widening tricks built purely on channel difference — Haas delays, hard width boosts — cancel or hollow out in that sum. Check every width move with a mono Utility on the master; real doubles and modulation-based width survive it.
How loud should vocal doubles be?
3–6 dB under the lead for tight modern hooks — present as size, not audible as separate voices. Looser styles can run doubles hotter. Either way, de-ess doubles harder than the lead so stacked S sounds don't smear the image.
Hear it onyour voice.
PhantomRack turns everything in this guide into a finished rack: upload a reference, download a stock-Ableton .adg tuned to it. First rack is free — no card required.