Back To Schedule

Instrument Prototyping at the Speed of Phausto

What Happens When Building an Instrument Is (Almost) C++ Free?

What happens when building an instrument is (almost) C++ free? Audio Developer Conference 2026 —
Prototyping a performing physical instrument has always needed a heavy knowledge of C++. Development in the Bela platform with Pure Data or SuperCollider cut these barrier, but lowering performance significantly.

When writing C++ code, lighter cpu charge and performative sound has a tardeoff: e very change to a sensor mapping, every tweak to a resonance model, meant a compile–flash–tweak-listen loop measured. That time quietly burdens instrument design itself: you converged early, because iterating was expensive, and most exploratory ideas never got tried. This talk asks what happens when that cost drops to near zero. Live on a Bela, using the Phausto to Bela exporter , I build an instrument from scratch in real time: The Larsenito Pecker. Its only input is a microphone placed in front of the speaker: the instrument is the feedback loop that interact with Phausto sequencers and physical models made available by the extensive FAUST libraries , it can draw into. Phausto is a library and DSL for Pharo that makes DSP programming acessibleThe Phausto exporters gives you instantly ready to compile C++ code with entry point to Bela digital inputs/

Feedback is the perfect stress test for the idea. Phausto collapses that delay ears, hands, and the live signal in the same loop: which is what makes the instrument playable at all. I’ll try to build it on stage, tame the feedback by ear, and play it. Then I’ll use it as a case study for a larger question: when the loop between idea and sound disappears, do we explore more and make stranger, better instruments? or do we just arrive at the same instrument faster, having lost the friction that used to force good decisions early?