QVI Lab
Neyora
Neyora — DDSP Instrument Synthesis
Transform text, voice, or MIDI into realistic instrument performance using DDSP technology. Built for composers who want an instrument that understands phrasing, not another text-to-noise toy.
What DDSP changes
Classic samplers play back recordings. Pure neural vocoders often smear pitch and lose editability. DDSP sits between them: the network estimates musical controls, then a differentiable synth renders tone you can still reason about — breath, bow, reed, room. Neyora uses that idea for solo instrument performance from text, voice, or MIDI.
Demo
01
Condition
Start from a text prompt, a hummed take, or MIDI. Neyora reads intent, pitch contour, and timing — not a canned loop library.
02
Model
Differentiable DSP separates what should be learned (timbre, noise, expression) from what should stay physical (oscillators, filters, envelopes).
03
Perform
The engine renders a playable phrase you can drop into a score or session, with control instead of a one-shot sample dump.
