Run a recording studio from a Python script.
Real VST3 amp sims, hardware drum machine samples, sample instruments — driven headlessly from code, rendered to finished multitrack songs. Start with a self-contained NumPy/SciPy sketch, then explore the studio rigs.
Two full albums came out the other end — have a listen.
Two Albums from One Python Studio
Distinct writing, source preparation, performed buses, and two albums with different musical identities.
Stems That Sum to the Mix — and What the Null Test Proves
Float-stem reconstruction, nonlinear buses, delivery dither, and the limits of a null test.
One Groove, Several Drum Kits
Compare sampled kits in the arrangement while preserving timing, source formats and listening level.
Test the Studio on a Clean Machine
What the repository CI checks, what it leaves to local renders, and how to avoid mistaking a passing job for musical quality.
Driving VST3 Effects from Python
An optional plugin stage for a studio built around saved performances and reproducible mixes.
Render Note Events into Audio with FluidSynth
Sample-position scheduling, separate instrument buses, and explicit SoundFont dependencies.
Run an Instrument in Its Own Process
Separate a plugin-specific runtime from the main studio with JSON events and a WAV result.
Build a Sample Instrument You Can Share — Zones, Velocity Layers, and Python Playback
Create a small instrument from synthesized tones, describe it with a JSON zone map, and play it through a Python sampler without a factory library.
A Complete Studio in a Python Script
Separate the score, performed sources, processed stems and delivery files so each change has a clear place.