Log inGet started
World

The House Wears a Face

Branchmain
Branchesmain
Commit8e7f3baa4a

README.md

FilmMix

The hundred seconds, as a mix.

One of these exists, on the Stagecraft entity. Every source in demo.stagecraft.SOURCES stands where its sound is made and is heard through the listener riding the camera, so most of the film's sound design is already done by walking the camera through the building. What this component adds is the part distance cannot say.

What it carries

Source envelopes. A list of { t, volume } per source, read off the film clock and written to the Audio component every frame. Rain that GETS IN at shot 6 and not before. A music box that stops mid-phrase. A calliope that is inaudible for the first half of the film and is the only thing left in it at the end, and a second, short-ranged copy of the same tune under the dressing-room floor, which is where it is first heard.

Channel envelopes. Three channels carry the act structure over the top of the sources: weather (outside, and what gets in), room (the building's own noises and the fall) and music (the box, and the calliope that does the laughing). The storm drops away for three seconds after the chandelier lands, because the ears are still ringing; it does not come all the way back; and it recedes for good under the last shot, so what is left when the light goes is one calliope in an empty building.

Because both are functions of the film's time and nothing else, the same second of the film sounds the same in every take.

Headroom

Twelve spatial sources, four thunderclaps and the chandelier sum into one master. The film is mixed under a master trim set in demo.stagecraft.reset(); without it the loudest second pins the output at full scale and the impact reads as a broken take rather than a heavy one.

Fields

FieldMeaning
activeOff freezes every source and channel where it stands.

Methods

CallReturns
mix:report()What every source is SET to beside what the mixer says it is DOING: its volume, whether a voice for it is audible, and if not, the reason (outOfRange, gainZero, noVoice). Plus the channel levels, the audible count and the listener. The one call that separates "asked for" from "heard".