Reshape any voice. Keep the pitch.
THROAT reshapes the vocal tract, not the performance. Move the formants, the gender, the tilt and the resonance, and the pitch comes out exactly as it went in. Not because it was corrected, but because the maths never touches it.
This is not a video and not a simulation. It is THROAT's own engine compiled to run in your browser, processing a real acapella live. Turn the knobs. The pitch will not move.
Press play
Drag a knob, or use the arrow keys. Double click to reset.
Audio used with permission. THROAT runs at 21 ms of latency here, exactly as it does in your DAW.
Every voice is a pitch and a tube. THROAT leaves the pitch alone and reshapes the tube, so a performance keeps its timing, its phrasing and its intonation while the character in front of it changes completely.
SHIFT lengthens or shortens the throat, GENDER scales the formants between male and female, FOCUS sharpens or softens the resonances, and TILT sets the brightness. Small moves flatter a take. Large ones make a different singer.
The engine warps the spectral envelope and nothing else. Bins never move, the hop never changes, the phase is left alone. There is no pitch tracking to drift and no grains to respace, so there is no wobble to tune out.
A true spectral envelope, tracked independently of the singer's pitch, then warped exactly where you ask and nowhere else.
Formant shift, ±12 semitones. Down for a bigger, darker throat, up for a smaller, brighter one. The note stays where the singer left it.
Scales the formants across the male and female range, ±100%. The fastest route from a voice to a different person.
Consonants and sibilants are guarded while the vowels move, so a big shift keeps its diction instead of lisping. Starts at 75%, and you will want it there.
FOCUS tightens or softens the resonances. TILT tips the whole spectrum, ±6 dB per octave, for brightness that follows the voice rather than an EQ curve.
Move the formants or the tilt on their own, 0.05 to 12 Hz. Slow for a voice that breathes, fast for something that should not be human.
DRIVE saturates the effect, not the source. MIX blends to a latency-matched dry, so a partial setting never phases against the original.
GRAB captures the vocal-tract signature of any voice you play it. STEAL moves your source toward it, MORPH travels between two of them.
Play a voice and capture its envelope. Two slots, so you always have somewhere to travel from and somewhere to travel to.
Blend your singer toward the captured timbre. At low amounts it is a family resemblance. High up it is somebody else entirely.
Ride between A and B. Automate it and the voice changes identity across a line, still on pitch, still in time.
At SHIFT 0, GENDER 0 and no vowel colour, the warp gain is exactly 1 and the engine hands back what you fed it. Line the output up against the dry, at exactly the latency THROAT reports, and the residual sits more than 75 dB down: far below anything you will ever hear. THROAT sitting on a channel doing nothing does nothing, and that is measured, not promised. SOLO hands you the difference so you can hear precisely what it added.
Where you stand depends on when you bought the bundle. Here it is plainly.
This drop only Buy the bundle this week and THROAT lands in your account, nothing to redeem. Own it already? Use your 50% code at checkout.
Free with this week's bundle · €35 on its own · 5-day trial, no card · AU · VST3 · AAX · macOS & Windows