Kyle Fakhoury, working as Olydian Musical Instruments, built Photosynthesis in 2025 — a digital synthesizer that runs on solar power. His next instrument is no less specific: the VC-01 Voicecrafter, a hardware synth built around nothing but the human voice.
Olydian’s first sound demo of the VC-01.
Two engines, and why both are necessary
The VC-01 runs formant synthesis and consonant-vowel sampling side by side. The developer has not yet explained how they interact, but the pairing is the design, and it follows directly from how speech is actually built.
Formant synthesis generates vowels. A vowel is, acoustically, a set of resonant peaks — formants — imposed on a periodic source by the shape of the vocal tract. Move the peaks and /a/ becomes /i/. This is tractable to synthesise and it is continuously controllable, which is exactly what you want for an instrument: you can glide between vowels, sit between two of them, or go somewhere no human tract can reach.
Consonants are not like that. Plosives are a closure followed by a burst of noise; fricatives are turbulence; affricates are both in sequence. They are transients and noise processes, and synthesising them convincingly is much harder than synthesising a sustained vowel. Sampling them is the sane answer — hence 20+ sample banks on the consonant-vowel side.
So the split is not two engines bolted together for variety. It is the continuous part synthesised and the transient part sampled, which is the correct division of labour for voice and is why singing synthesisers have historically been so much better at “aaah” than at “stop.”
The touch panel is the instrument
A large touch panel spans almost the entire centre of the unit, with individual vowels assigned to touch-sensitive keys.
This is the interface decision that determines whether the thing is playable. A vocal synth controlled by a menu of vowel presets is a sound module; a vocal synth where vowels are laid out spatially under your hand is an instrument, because you can move between them in time. Vowel transitions are most of what makes a vocal line read as a vocal line, and performing them requires the vowel selection to be a gesture rather than a parameter.
The physical layout is unusual and deliberate:
- A three-octave mechanical keyboard using Cherry-style switches in the lower section, with pitch bend on the left
- The vowel touch panel across the centre
- Knobs and sliders throughout
A mechanical-keyboard-switch keybed is a real choice rather than a cost saving. It gives no velocity nuance and a hard, defined actuation — which for a consonant-triggering instrument is arguably the right feel, since a plosive is an event, not a dynamic swell.
Sequencing
Each engine has its own programmable ADSR controlled by top-panel sliders and its own dedicated FX processor. Two LFOs with various shapes and a reverse option sit on the left. A global multi-FX on the right offers vibrato, reverb and three delay types, with the reverb exposing dry/wet and size. Stereo spread control and portamento are also on board.
The built-in sequencer supports looping and overdubbing, and the developer says it can store thousands of steps for long pre-written playback. Presets are storable in the instrument.
Looping with overdub, rather than step entry, is the right model for a voice instrument. Vocal parts are layered — a held vowel bed, a rhythmic consonant figure over it, an answering line — and building that by recording passes over a loop is how anyone actually works with voices. A grid editor would fight the material.
What it is and is not
Synth Anatomy’s own assessment is that this will be a niche instrument with a limited pool of interested people, and that is almost certainly right. Vocal synthesis is not what most people want from a hardware synth.
It is also why it is worth paying attention to. The things that get built for small audiences by one person tend to be the ones with an actual point of view, and there is very little hardware in this space — vocal synthesis has lived in software, in Vocaloid-lineage tools and in academic formant synthesisers, with almost nothing you can put your hands on.
Price and availability are TBA, and the instrument is not officially released.