Live-coded transitions: audio, visual, and paired downbeat moves

A section change in a live-coded AV set is a public act: the audience sees the code mutate and hears and sees the result simultaneously. This module treats transitions as a vocabulary — a set of named, rehearsable moves — so you can select the right technique for a boundary, execute it under pressure, and know whether it will need a follow-up save or clear itself automatically.

The module is split into three registers that culminate in the paired downbeat. Audio transitions are covered first because they are the primary signal the audience reads. The filter open-build (low-pass cutoff ramp upward) and high-pass breakdown (removing low-end weight without dropping the pattern) are slow-burn moves suited to building and releasing tension over a full phrase. The degrade thin-out (raising degradeBy across cycles) is a voice-shedding technique that preserves the timing grid while the voice fades. The drop-and-return (one cycle of near-silence, then the full groove snapping back) is the most aggressive audio transition and requires the most setup — the audience must have been trained by the build to expect the snap. The delay throw (adding feedback delay on a voice’s final cycle) is a self-clearing move whose echo tail carries the transition without any follow-up. Muting and unmuting a voice by commenting it in or out is the most legible and reversible move in the vocabulary, and it belongs in every performer’s toolkit for when a cleaner, more auditable change is needed.

Visual transitions are covered as counterparts. Feedback bloom (feeding the previous frame back at an FFT-driven amount) is the visual equivalent of an audio build — trails accumulate as the audio energy rises. Kaleidoscope symmetry stepping on a downbeat gives a visual drop that reads as a sudden structural change. Color inversion or posterize toggled on the drop downbeat is a single-frame hit — the briefest possible visual event, which makes it the right partner for a sharp audio snap. Scroll and scale speed tied to a rising FFT band mirror the audio’s ramping energy in the spatial domain. Fade-to-black empties the canvas gracefully for a full breakdown. Reactivity repointing (swapping the FFT index driving a visual parameter) aligns visual energy with the new section’s audio driver, and must be done as part of or immediately after the audio transition to avoid a disconnected visual.

The paired-downbeat atom brings the two registers together: the strongest section changes fire one audio and one visual transition on the same downbeat. This synchronisation is the whole task — it is what distinguishes a coherent AV performance from two independent streams running in parallel. The capstone enforces this by requiring four section changes with named pairs, at least one self-clearing pair and one multi-cycle pair.

Required atoms gate every move in the capstone vocabulary. Supporting atoms connect this module to the broader set-craft context: livecoding-edit-cadence-grid-anchor and livecoding-diff-size-ladder govern how transitions are timed and sized; set-arc-build-with-reserve explains why the biggest transition must be held until the boundary it has built toward; set-arc-ramp-then-resolve is the tension-resolution principle the filter/degrade ramps embody.