This component allows caption creators to define the "safe" transformation space for their captions. Creators:
- Upload videos with original caption files
- Set two anchor points (lower and upper bounds) that define acceptable ranges for:
- Level of Detail (minimal to elaborate descriptions)
- Expressiveness (neutral to creative/evocative language)
- Preview AI-generated transformations and manually edit if needed; can also lock certain captions
- Export a configuration file containing the original captions, transformation boundaries, and video metadata
The key idea is that creators maintain editorial control by establishing guardrails within which viewer customization can occur.
This component enables DHH viewers to personalize captions during playback based on their preferences. Viewers can:
- Adjust captions within the creator-defined boundaries using a 10×10 style grid (Level of Detail × Expressiveness)
- Select from three sound representation methods: source-focused, onomatopoeia, or sensory quality-based
- Toggle genre alignment to match captions to the video's tone and style
- Use either direct manipulation controls or a natural language chat interface to specify preferences
The system uses GPT-5 to transform captions in real-time based on viewer selections while respecting the creator's boundaries.