taxonomy/MODALITY_INDEX.md

Master Modality Index

This document serves as the definitive index of all modalities in the taxonomy, with unique IDs and their locations.

ID Format

Modality IDs follow this format: {input}-to-{output}-{variant}

Examples:

• text-to-img - Text to Image (basic)
• img-to-vid-lipsync-audio - Image to Video with Lip Sync from Audio
• audio-to-text-transcription - Audio to Text (Transcription)

Unique ID Registry

Video Generation (Creation)

ID	Name	Input	Output	File Location
img-to-vid-no-audio	Image to Video (No Audio)	image	video (no audio)	video-generation/creation/modalities.json
img-to-vid-ambient-audio	Image to Video (Ambient Audio)	image + text	video (ambient audio)	video-generation/creation/modalities.json
img-to-vid-lipsync-text	Image to Video (Lip Sync from Text)	image + text	video (speech)	video-generation/creation/modalities.json
img-to-vid-lipsync-audio	Image to Video (Lip Sync from Audio)	image + audio	video (speech)	video-generation/creation/modalities.json
img-to-vid-lipsync-lora	Image to Video (Lip Sync with LoRA Character)	image + text + lora	video (speech)	video-generation/creation/modalities.json
text-to-vid-no-audio	Text to Video (No Audio)	text	video (no audio)	video-generation/creation/modalities.json
text-to-vid-with-audio	Text to Video (With Audio)	text	video (synchronized audio)	video-generation/creation/modalities.json
audio-to-vid	Audio to Video	audio + text	video (original audio)	video-generation/creation/modalities.json
multimodal-img-audio-to-vid	Image + Audio to Video	image + audio	video (original audio)	video-generation/creation/modalities.json
multimodal-text-img-to-vid	Text + Image to Video	text + image	video (no audio)	video-generation/creation/modalities.json
3d-to-vid	3D Model to Video	3d-model	video (no audio)	video-generation/creation/modalities.json

Video Generation (Editing)

ID	Name	Input	Output	File Location
(To be populated)				video-generation/editing/modalities.json

Audio Generation (Creation)

ID	Name	Input	Output	File Location
text-to-audio	Text to Audio	text	audio (general)	audio-generation/creation/modalities.json
text-to-speech	Text to Speech	text	audio (speech)	audio-generation/creation/modalities.json
text-to-music	Text to Music	text	audio (music)	audio-generation/creation/modalities.json

Audio Generation (Editing)

ID	Name	Input	Output	File Location
(To be populated)				audio-generation/editing/modalities.json

Image Generation (Creation)

ID	Name	Input	Output	File Location
text-to-img	Text to Image	text	image	image-generation/creation/modalities.json

Image Generation (Editing)

ID	Name	Input	Output	File Location
(To be populated)				image-generation/editing/modalities.json

Text Generation (Creation)

ID	Name	Input	Output	File Location
(To be populated)				text-generation/creation/modalities.json

Text Generation (Editing)

ID	Name	Input	Output	File Location
(To be populated)				text-generation/editing/modalities.json

3D Generation (Creation)

ID	Name	Input	Output	File Location
(To be populated)				3d-generation/creation/modalities.json

3D Generation (Editing)

ID	Name	Input	Output	File Location
(To be populated)				3d-generation/editing/modalities.json

ID Assignment Guidelines

1. Uniqueness: Each modality must have a globally unique ID across all files
2. Consistency: Use kebab-case format
3. Descriptive: ID should indicate input, output, and key differentiator
4. Stable: IDs should not change once assigned (treat as permanent identifiers)
5. Sequential: When adding variants, use descriptive suffixes (not numbers)

Next Available ID Prefixes

These prefixes are ready for new modalities:

Video

• vid-to-vid-* (video editing/transformation)
• music-to-vid-* (music visualization)
• multimodal-*-to-vid (other multimodal combinations)

Audio

• audio-to-audio-* (audio editing)
• img-to-audio-* (image sonification)
• vid-to-audio-* (video to audio extraction/generation)
• music-to-* (music transformations)

Image

• img-to-img-* (image editing)
• audio-to-img-* (audio visualization)
• vid-to-img-* (video to image extraction)
• multimodal-*-to-img (multimodal image generation)

Text

• audio-to-text-* (transcription, audio description)
• img-to-text-* (image captioning, OCR)
• vid-to-text-* (video captioning, transcription)
• multimodal-*-to-text (multimodal text generation)

3D

• text-to-3d-* (text to 3D model)
• img-to-3d-* (image to 3D reconstruction)
• vid-to-3d-* (video to 3D reconstruction)
• 3d-to-3d-* (3D model editing)

Change Log

• 2025-10-22: Initial master index created with existing modalities
• Future changes will be logged here