All couplings requiring onset, tempo, beat-phase, or per-instrument signal share one root blocker and are not achievable with the 4-bin bridge

The 4-bin FFT is display-rate (~60 Hz), smoothed, full-mix band energy. It contains none of the signals needed for: onset detection (transient per kick/snare), per-instrument isolation (kick-only channel), tempo/beat-phase (bar position, beat index), or downbeat/phrase position. This entire family of couplings — onset flash, kick-specific sidechain, beat-locked scroll, downbeat scene cuts, loudness (RMS) fade — is blocked for the same structural reason. All require an L0 analyser upgrade (onset/tempo/RMS/per-instrument taps plus a clock export) and an L3 perception bridge. The only current substitute for the whole family is an amplitude envelope proxy (a.fft[i]²), which is energy, not a detected event.

Examples

‘Flash on each kick’ (onset), ‘scroll one cell per beat’ (beat-phase), ‘duck the visual on the kick specifically’ (per-instrument onset), and ‘cut scene on the 1’ (bar phase) are all different-sounding requests blocked by the same root: no event-level signal on the bridge.

Assessment

A collaborator suggests four visual ideas: (1) flash the frame on each kick, (2) scroll one cell per beat, (3) hue-shift on every snare, (4) brightness follows overall loudness. Which are achievable now, which are blocked, and what is the shared root cause of the blocked ones?