Modalities

One page per modality. Sub-topics live inline because the ideas cluster.

Planned pages

  • Image — txt2img, img2img, inpaint/outpaint, upscaling, style transfer.
  • Video — txt2vid, img2vid, vid2vid, frame interpolation, consistency, clip-length limits.
  • Audio — dialogue and voice cloning, ADR, lip-sync, Foley/SFX, music/score.
  • Spatial — 3D generation, NeRF, Gaussian Splatting, camera tracking, plate reconstruction.