AUD · Audio tools

Audio Doubler

Shareable link

Settings are written to the URL as you change them. Nothing differs from the defaults yet.

Turn one take into three. Two copies sit a few tens of milliseconds either side of the original, each drifting at its own slow rate, so the result reads as a performance that was tracked twice rather than as a processed single take.

Delay and spread

Delay, from 5 ms to 60 ms, sets how far behind the original the pair sits. Spread, up to 30 ms, pushes one voice earlier and the other later around that point.

The delay range matters more than it looks. Under about 10 ms the copies comb-filter against the original and the result is hollow rather than thick. Between 15 ms and 35 ms the ear fuses them into one wider event. Past roughly 40 ms it stops fusing and starts hearing a slapback.

Spread keeps the two voices from landing together. With spread at zero they arrive at the same instant and behave as one louder copy.

Drift

Drift, up to 6 ms at a rate from 0.05 Hz to 2 Hz, is what makes the copies sound like separate takes.

A delay that never moves produces a mathematically exact duplicate, and the ear identifies exact duplicates as the same sound rather than as two of them. Moving the read head slowly bends the pitch a few cents either side, which is the same small inaccuracy that makes two real takes sit together.

Each voice wanders on two rates that do not divide into each other, so the movement never settles into an audible cycle and the two voices never drift in step.

Rates above about 1 Hz start being heard as vibrato rather than as human imprecision.

Width

Width places the pair across the image, from centred at 0% to fully opposed at 100%. The gains are constant power, so moving it does not change the level of the pair.

At 100% the first voice sits left and the second right, with the original still centred between them. This is the setting that makes a doubler read as width rather than as depth.

Doubling vs chorus vs delay

All three put a delayed copy of the signal back over it, and the delay time decides which one you have.

Under 10 ms with continuous modulation is a flanger or chorus. Between 15 ms and 40 ms with slow random drift is a doubler. Past 60 ms it is a delay, and the copy is heard as a repeat.

Example: a single lead vocal that needs to sit wider

Delay 24 ms, spread 11 ms, drift 2.5 ms at 0.3 Hz, width 90%, mix 45%.

One copy lands at 18 ms on the left and the other at 29 ms on the right, both wandering slowly enough to read as timing rather than as modulation. The original stays in front because the mix keeps it there.

Frequently Asked Questions

Chorus modulates its voices continuously at a rate you set, which is heard as shimmer. A doubler places its voices at fixed offsets of tens of milliseconds and lets them wander slowly and unpredictably, which is heard as a second performance rather than as an effect on the first.

From the drift. Moving a delay line slides the read head, and a moving read head bends pitch the same way varying tape speed does. That is how hardware doublers have always worked, so the drift depth and rate are the honest controls rather than a cents value the effect does not really have.

Nothing. A single channel has nowhere to place the two voices against each other, so both sit centred and the control is inert. The tool reports this above the controls when the file it loads is mono.

The delay is too long. Past roughly 40 ms the ear starts hearing the copy as a separate event rather than as part of the same one. Bring the delay down toward 20 ms and the two fuse.

No. At 100% the original disappears and only the two copies are heard, which is thin because the copies are delayed. Somewhere between 40% and 60% keeps the original in front with the doubles filling in behind it.

Explore Our Tools

Browse all tools