No.236 · AI & Chat
Diffusion Preview
Image generation shown as it happens: four candidates start as coloured noise and resolve step by step, with the step count and seed alongside.
The prompt
Written for the variant you picked above: Paper
Build an image-generation card that shows denoising. A prompt row with a Generate again button; below, a large square canvas and a side column with four candidate thumbnails (a radiogroup), step/seed/sampler readouts and a thin progress bar. Each candidate has a target picture (here a generated canvas study per seed) and two noise textures from its seed — a 12×12 colour noise drawn up smoothly (the blotchy latent look) and a 96×96 noise drawn crisp (grain). Each step draws the target with a blur that shrinks from 22px to 0 and saturation rising, then the coarse noise fading out, then the fine grain fading out, with a = k^1.1 (coarse noise at up to 72%, grain at up to 36%) so shapes appear before detail. Thirty steps at 85ms; each candidate runs 1.5 steps behind the previous, so they resolve in a ripple. The large canvas mirrors the picked candidate; picking another after generation shows it large. Generate again bumps every seed and starts over; under reduced motion the finished images appear at once. Selected variant — Paper (paper): Apply the paper theme. Where the brief above gives no colour values for it, choose a restrained paper palette with text at 4.5:1 contrast or better. Keep the geometry, states and motion described above unchanged. This selection takes precedence over references to other themes or options in the general brief. Implement this selected appearance; do not add a variant picker or alternative appearances unless requested. Implementation requirements: Use React and TypeScript. Keep the component self-contained, with uniquely scoped CSS and no animation libraries. Include required imports, styles and font setup in the result. Interaction: Watch the four candidates resolve; pick one to see it large; Generate again for new seeds. Motion: 30 steps at 85ms; candidates staggered by 1.5 steps; blur 22px → 0 and noise fading on k^1.1. Accessibility: Candidates are a labelled radiogroup (with seeds in their names); the step count is announced politely and the large canvas is role="img" with the prompt. Reduced motion shows finished images. Responsive behavior: A card up to about 40rem wide that fills narrower screens; dense rows and side columns stack on phones. Touch: Everything is tap-driven; hover hints have tap or focus equivalents. Preserve keyboard access, visible focus and reduced-motion behavior. Keep fictional demo data and simulated service calls in a separate usage example; expose callbacks for real actions and recover from rejected requests. Keep usage-example links inside the demo: use local buttons or prevent navigation. Demonstration links must not navigate to Vitrine routes or open other pages; preserve real URL support in the reusable component itself. Verify the result at narrow and wide viewport sizes, with keyboard input and reduced motion enabled.
Notes
- Interaction
- Watch the four candidates resolve; pick one to see it large; Generate again for new seeds.
- Motion
- 30 steps at 85ms; candidates staggered by 1.5 steps; blur 22px → 0 and noise fading on k^1.1.
- Accessibility
- Candidates are a labelled radiogroup (with seeds in their names); the step count is announced politely and the large canvas is role="img" with the prompt. Reduced motion shows finished images.
- Responsive
- A card up to about 40rem wide that fills narrower screens; dense rows and side columns stack on phones.
- On touch
- Everything is tap-driven; hover hints have tap or focus equivalents.