
From a reference / Impruneta: Clay & Stone
What changes when colour is treated as material?
Learn from the mixture.
Explore the connection & keep your decisions
Clay and mineral-toned surfaces offer a starting question about colour relationships. The bath explores selected digital pigment models rather than deriving a formula from a photograph.
What this experiment representsThe bath uses simplified mixing and motion models. Study hex values are not measured pigment coefficients or paint recipes.
Read the study and its source notesTry one considered change.
- Choose two pigments and keep the vehicle fixed.
- Compare a restrained addition with a more saturated mixture; save an image of each.
- Decide which relationship serves your composition, and record the pigment selections separately from the resulting strokes.
Your judgement / optional
Keep the reason with the result.
Notes stay on this device when saved. Copying a study link includes them. The reference guides your thinking; choosing it does not apply a palette.
The link restores pigment and vehicle selections, not the live fluid field or every gesture. Save the plate image to keep the mixture you see.
Hold to deposit pigment, or begin with a measured two-drop mix.
Pigment & Emulsion Bath
Paint does not mix like light. This bath advects archival pigments through a stirred vehicle and blends them with two-constant Kubelka–Munk scattering — K/S inverted from each pigment’s masstone reflectance — so gold into Klein blue turns the muddy green of a real palette, not the gray of an RGB average.
Mix with a measure.
Start with two equal drops, then change the proportions. These swatches calculate a repeatable mixture; the fluid bath above remains a transient gesture.
The destination receives the computed colours, not pigment scattering or fluid state. Ratios: 1 : 1 : 0.
Keep one. Try another.
Pin a plate before changing the controls. Each pin keeps its own recipe and lighting context.
Carry this specimen into another tool
The image travels locally in this tab with its source recipe. This is a flattened plate: the destination does not recreate fluid motion or convert material coefficients.
Keep the reference and your decisions
Download the starting reference, current settings and your decision notes alongside your specimen. Add notes in the reference panel.
npx @temporal/chromologium add pigment-bathFurther reading
Questions & notes
A closer look at the ideas, materials and methods behind the work.
Material cues, responsive rendering, and the ergonomics of spatially grounded interfaces.
1 questionDiegetic Interface Architecture treats interface elements as parts of a coherent setting rather than detached controls. Material cues such as vellum tooth, lacquered borders, brass fittings, and responsive friction can clarify hierarchy, suggest how an object behaves, and give the interface a distinct sense of place. The method is one option among many: each cue should support the task, remain accessible, and earn its visual weight.
Related questions
CSS is the substrate: radial-gradient noise, composited backdrop-filters, and hardware-accelerated transform layers render the ambient aesthetic with zero layout shift. GPU shader instruments are opt-in overlays that mount only on visible, user-invoked specimen surfaces, share a single WebGL context, and unmount to static CSS under reduced-motion or frame-budget pressure.
Tactile Dwell Latency introduces intentional 120ms to 240ms micro-resistance to critical user interactions (such as chamber shifts or token downloads). This mirrors the physical tactile feedback of mechanical switches, dramatically reducing accidental double-clicks and reinforcing intentionality.
Developed by Fagerholt & Lorentzon (2009), the taxonomy categorizes UI along two axes: 3D spatial presence and narrative reality. Diegetic UI exists physically in the world; Spatial UI is rendered in 3D space for the user; Meta UI represents camera-lens artifacts (film grain, cracked glass); and Non-Diegetic UI is traditional 2D HUD overlays.
The workbench applies the Cauchy relation n(λ) = A + B/λ² across sixteen wavelength bands, then derives each band's prism deviation from δ = 2·asin(n·sin(α/2)) − α. Presets carry real coefficients for BK7 crown, SF10 flint, fused silica, and Bohemian lead crystal, with live Abbe number readouts and GLSL export.
The bath models paint as a turbid medium using two-constant Kubelka–Munk scattering: each pigment's K/S is inverted from its catalogued masstone via K/S = (1 − R)²/2R, and mixtures are concentration-weighted before reflectance is recovered. Gold stirred into Klein blue therefore yields palette green — never the gray of an RGB average.