Picture-Bound Sound
In plain words: Broadcast: picture-bound sound
How Broadcast designs sound tied to exact picture: effects, foley, room tone and ambience judged against a locked cut, including sound returned by video models.
- 1Each layer records event, place, distance, timing and rights.
- 2Model-returned sound needs Broadcast declared before execution and its own admission afterwards.
- 3A changed picture, cut or claim makes affected sound stale.
Words on this page, decoded click one for its card
A door closes in a small room. Its impact, distance, reverberant tail, and the silence around it belong to the meaning of that exact shot. Broadcast calls this picture-bound sound design: effects, foley, room tone, environmental beds, and ambience judged against picture, scene, cut, claim, or editorial timeline.
broadcast.design_picture_sound@2 starts with PictureSoundBrief@1, locked picture digest/timebase, event/space roles, source/rights references, target channel/loudness profile, and acceptance criteria. Recording, synthesis, licensed layers, processing, or a compound video-model result can propose candidates. Each layer retains source, implementation, transforms, rights, time relation, and digest.
Hear the scene, then admit the result
Describe how each candidate belongs to the scene in PictureSoundRevision@1 and optional PictureSoundCueMap@1: bind its event, place, material, distance, perspective, duration, continuity, cut relation, and uncertainty. Probes then measure boundaries, clipping, channels, loudness, and format. Editorial review uses that evidence to decide whether the sound communicates the visible event. Seal accepted selections as PictureSoundBundle@1 for Edit.
Riffmaw owns music; Language Edition owns spoken performance/captions against a locked source; EchoVoice owns speech lifecycle; Foundry may own game-event sound. A reusable standalone library would need its own proved CompositionApp records, judgment, finish, and recovery. Broadcast's scene does not create that owner.
When the video model also returns sound
A video model may return sound in the same compound result as its picture. Before execution, MediaFacetAuthoritySet@1 must declare picture_sound, broadcast.studio revision 4, and the exact PictureSoundBrief@1 digest. Core retains the technical attempt, and PrismVision & Media preserves the compound/child digests, timebase, and provenance.
After technical settlement, Broadcast independently issues SemanticFacetAdmissionReceipt@1. Successful demux, another owner's judgment, or generate_sound cannot substitute for that admission. Quarantine/delete unexpected sound; when media is inseparable, every necessary owner must accept the exact compound use.
A changed picture, cut, source claim, timebase, or brief stales affected sound. Recorder, generator, renderer, or destination acknowledgement loss remains unknown until reconciled. Corrections create successors; sound acceptance itself supplies no final mux or publication authority.