Quill
Categories: Software Prototypes/Non-Commercial Products
Submitted By: Chuling LI
Elevator Pitch

Quill
Product Description
Quill captures a sound and turns it into an instrument you can play. Record a few seconds of any instrument or voice, no perfect take required, and Quill recreates its tone and the way it is played on any MIDI keyboard or controller. Unlike a sampler, which stores the recording and replays a fixed clip, Quill learns the sound as a recipe and rebuilds it from scratch. The captured instrument stays natural well above or below its original pitch and sustains indefinitely, so it feels like playing a real instrument.
Quill also captures how an instrument is played. From a short performance clip it learns the source’s plucks, slides, legato and strums, then transfers them onto your MIDI keyboard: overlapping notes bring in its own legato or slide, quick chords trigger its strum. The captured instrument doesn’t just reproduce sound; it reproduces performance.
How It’s Innovative
Every tool that turns a recording into a playable instrument makes you give something up. A sampler will take any sound, but all it really does is play the recording back. Hit a key and you get the same clip, shifted in pitch. It can only reproduce techniques that were already in the recording, and it can’t truly sustain a note, since a held key either cuts off or loops. Purpose-built virtual instruments work the other way around: they play wonderfully, with real slides and sustain, but the tone is the vendor’s, not a sound you recorded. Existing tools do not combine a sound you captured yourself with the expression of the original.
Quill closes that gap by storing a recipe instead of a recording. It breaks a few seconds of a pitched sound into its harmonic and noise parts and rebuilds it from scratch, so you can play the captured tone high or low and sustain it for as long as you like. It also learns how the sound is played. From one short clip, without per-instrument datasets or per-instrument training, it picks up the slides, legato and strums and stores them as gestures it can reuse. A slide turns into a pitch curve you can replay at different speeds and intervals, and a strum into a timing pattern that fits different chords. Because the tone and the technique are kept separate, you can recombine them, so a voice can slide like a guitar. I have not found a tool that turns a single short clip into all three: captured timbre, extracted articulations, and MIDI-playable performance behavior. In the end, a sound you record becomes an instrument you can play on a MIDI keyboard you already own, with its tone and its feel intact.
Most Inspiring Use Cases
For a producer, the hardest part is often finding the right sound. You can hear it in your head, in an old record, or in the room around you, and still lose hours digging through sample libraries and presets for something close. Quill turns that around: instead of hunting for a sound, you capture it. I recorded forty seconds of a friend’s guitar, and minutes later I was playing it from my keyboard, with the slides landing just like his. Quill had caught his tone, and a little of his musicianship too. The same works for a wine glass, a voice memo, or a synth buried in an old record: a few seconds is all it takes to turn any of them into an instrument you can play. Every sound around you, and every player’s touch, becomes a starting point for new music.
Expansion Plans
Three directions stand out. The first is MIDI 2.0 and MPE support: per-note pitch bend and per-note controllers are a natural fit for captured slide and vibrato, giving each voice its own expressive channel. The second is a timbre layer I’m researching, built with AI methods and trained quickly per clip using accessible cloud GPUs, to push the resynthesis toward higher-fidelity, performance-aware replication. The third is a shared library, so people can share and remix captured instruments and playing styles, playing one person’s timbre with another’s articulation.
Commercialization
Quill is a working prototype today, built on a simple idea from my own music-making: when you hear a sound you want, capturing it on the spot should be easier than searching through sample libraries for something close.
The immediate plan is to open and document the analysis and synthesis core, so musicians, researchers and developers can test the approach, reproduce the results and help improve it. This gives Quill technical credibility and early community feedback while the product direction matures.
If the response is strong, the commercial path is a VST3/AU instrument plugin that lives inside the DAWs people already use. The natural users are music producers, sound designers, and game and film audio teams, anyone who needs a specific timbre quickly and wants to turn it into something playable.