AI Sequencer
The sequencer can run in three different modes. The first is rule-based — the deterministic probabilistic engine described on the Generative Sequencer page. The other two use Orbiter's neural sequencing model to generate notes, either on its own or in response to what you play.
Modes
The mode picker lives in the Sequencer panel under the SEQ tab. The three options:
- Rule-based — deterministic patterns driven by the standard sequencer parameters (per-instrument probability, burst rate, scene preset). The same seed always produces the same sequence.
- Orbiter AI — Autonomous — the model generates notes on its own, taking the current scale and tempo as context. Use this when you want the model to play unattended, like a generative musician.
- Orbiter AI — Call & Response — you play notes (on the pads, via MIDI, or from the on-screen controls); the model listens and responds with complementary phrases. Your trigger notes are silenced — only the model's response is heard, so the input acts purely as a prompt.
Switching modes is non-destructive: tempo, scale, and instrument settings stay where they are, only the source of the notes changes.
Autonomous controls
When Autonomous is selected, four extra controls appear next to the mode picker:
| Control | What it does |
|---|---|
| Variety | How adventurous the model's output is. Lower values stick closer to predictable, on-scale phrases; higher values explore more unusual choices. |
| Continuous / Motif | Continuous (~) keeps generating endlessly. Motif (>) generates short phrases with breathing space between them. |
| Scale / Chromatic | Scale (S) snaps every output note to the currently selected scale. Chromatic (C) lets the model use any note, including outside-the-scale colour notes. |
| Speed | Clock speed relative to the sequencer tempo: ¼×, ½×, 1×, 1½×, 2×. The 1× setting follows the sequencer's BPM exactly; faster or slower multiplies that. |
These controls also drive the model in the standalone scene presets — Sleepy / Calm / Melodic each leave the model in a configuration that fits the mood.
Call & Response controls
In Call & Response mode you get the same Variety, Scale / Chromatic, and Speed controls. Continuous/Motif doesn't apply because phrasing is determined by what you play.
A short note appears beneath the controls when this mode is active: Trigger notes are silenced — only AI responses are heard. That's the intended behaviour: play a few notes, the model responds; the trigger notes themselves don't sound, so the dialogue is between you and the response, not you and your own input.
Tips
- Start in Autonomous with Scale lock on and Variety around 0.4–0.8 — that's where the model tends to be most musical.
- Crank Variety higher if the output feels too predictable; back it off if it's wandering.
- Motif mode pairs well with the Sleepy scene preset — short phrases with quiet between them are exactly what that mood wants.
- Call & Response is most expressive over MIDI from a controller; the on-screen pads work too, but a keyboard or pad controller gives you more rhythmic and dynamic range to prompt with.