Montage

The planner (Step 3) picks the treatment and the clips + anchors (WHAT); the frame-worker realizes it inside the frame file (HOW). Obey the frame’s pacing.

The three treatments

beat_cut — one clip per anchor (only on a beat_cut frame)

The asset-driven analogue of a per-onset typographic group: cut to a new clip on each anchor (the frame’s beats/onsets from the audiomap). Each clip is a class="clip" element (<img> for a photo, muted <video> for a motion clip) placed at its anchor with data-start/data-duration/data-track-index per the core clip contract. Between clips, crossfade the outgoing content to opacity:0 ending at the next anchor. Cut on the strong anchors; land a hero clip on a key_moment/downbeat.

ken_burns — slow push on one clip (fits a phrase_flow frame)

For calm frames: one clip held over the span with a slow scale/translate push (e.g. scale 1.0→1.08 + a small drift) eased across the whole span_sec — paced by the frame, not by beats. No hard cuts. Crossfade in/out at the frame edges. This is the right asset treatment when the beat grid is unreliable (calm music).

bg_under_text — clip dimmed behind a template/free group

A full-bleed clip dimmed ~30–50% as the background of a group whose foreground is a template or free-compose typographic treatment. The text rides on the same anchors; the clip is the bed. Use when the user wants their footage present but the message must stay readable.

Rules

• pacing decides the treatment: beat_cut only on a beat_cut frame; on a phrase_flow frame use ken_burns or a slow crossfade — never per-onset hard cuts on the (unreliable) calm grid.
• Clips are muted; the root owns audio. Mount each <video class="clip"> muted, as a direct child of the frame root (never nested in another timed element, or the renderer freezes it). The BGM is the only audio in v1.
• Crossfades animate opacity/autoAlpha, never visibility/display on a .clip (the framework owns clip visibility — that trips gsap_animates_clip_element).
• Backgrounds dim ~30–50% so any foreground text stays legible.
• Anchors are track seconds from audiomap.json; the worker subtracts the frame start to get frame-local time.
• Local staged assets only (assets/ via stage-assets.mjs); never remote URLs.

Deferred hook (not v1)

A clip that should play its own sound (interview cut, lyric clip) needs a sibling <audio> mounted at the root by the assembler, with the BGM ducked under it (a data-automation volume lane on the BGM, see creator-editing-recipes.md in hyperframes-core). The frame-worker mounts no audio. Keep clips muted in v1; wire clip-audio + ducking only when the user asks.