Outputs
patch-press’s pipeline is deliberately format-agnostic. Everything up to the final export stage — trim, envelope, pitch, loop, normalize — produces a plain SampleSet that contains audio buffers, note assignments, and loop points, with no idea what device it’s going to end up on.
The --format flag on sample and batch picks a target-specific exporter that turns that SampleSet into files on disk in whatever layout that device expects.
The one place the target leaks upstream is how many notes get captured. Formats differ in how much of a melodic capture they can actually use — the Deluge maps every note to its own keyzone, a .pti has no keyzones and ships one repitched sample — so rendering a full note grid for the latter pays for renders the exporter discards. Each exporter declares its appetite via notes_used(notes), and runner/pipeline.py:notes_to_capture consults it before a VST/CLAP capture starts. The analysis stages stay untouched and format-blind; only the size of the note grid changes. A new exporter that omits notes_used simply gets the full grid.
Currently supported targets
--format |
Target | Docs |
|---|---|---|
deluge |
Synthstrom Deluge — XML preset + WAVs in the SD-card layout | Deluge output |
pti |
Polyend Tracker family — self-contained .pti instruments |
Polyend Tracker output |
Deluge is what shipped first because it’s the device patch-press was built for. The pipeline architecture is set up to add more targets over time — anything with a defined preset format is a candidate.