Module plan 

What a voice tool has to do for one language, as a pure function of the dialogue, the scenes to voice, the cast and the manifest’s clips: the clips to make or make again (and why), the recorded clips a person must record again, the clips to remove and the lines nobody is cast to say. The tool reads and writes its files; plan only says what to do.

The tool brings three hooks:

• which variants a line needs, from its speaker and text (none for a line nobody says aloud, such as a reply);
• the words a line says as a variant (the hook VoiceManifest::is_stale takes);
• whether the voice of a generated clip changed, in the tool’s own words (“the sample changed”). The plan itself compares the clip’s voice with the cast’s.

Structs

Uncast

A clip a voiced line needs whose speaker the cast gives no voice.

VoicePlan

Everything there is to do for one language’s clips.

WantedClip

A clip a voiced line needs, with what the voice needs to say it.

Enums

Why

Why a clip is to be made, or recorded, again.

Functions

plan

Plans dialogue’s clips: the lines of the scenes in voiced need a clip for each variant variants(speaker, text) names, saying spoken(text, variant) (a variant whose words are None needs none). clips are the manifest’s. A generated clip whose voice is still the cast’s is made again if voice_changed(clip, cast voice) gives a reason.

recording_script

The clips of plan the voice under cast key key says, in the plan’s order (by scene, then script order): what that voice records in a session, each with the line before it.