Chromatic Adaptation User Guide

The page is a single research workspace with six tabs: Lab for running the lab, Dynamics for the adaptation time course, Matrix for the adaptation matrices, Data for the datasets behind every figure, Export for outputs, and Reference, which holds the standards, formulas, citations, and research notes described below.

• Lab: controls on the left, image and chart output on the right.
• Export: JSON export, PNG frame export, share URLs, all-method comparison, and batch HEX analysis.
• Standards: summary notes for CIE 160, CIECAM02, CAM16/CAT16, ICC profile adaptation, Fairchild and Reniff, Brainard and Wandell, and color constancy literature.
• Formulas: math references for the exponential adaptation curve, von Kries style diagonal adaptation, CAT matrices, spectral power distributions, and spatial modulation.
• Citations: citations for CAT methods, psychophysics, standards, and SPD computation.
• Research notes: practical interpretation of the A(t) curve, state map, CAT comparison, CIECAM02 adaptation degree, and gamut behavior.
• Keyboard: Space plays or pauses the animation, R resets, C compares all methods, and G toggles the GPU checkbox state.

• Choose image: uploads a local image file and processes it in the browser. The source is scaled to fit a 480 by 320 processing canvas.
• Sample: generates a 360 by 240 HSL gradient image for quick tests without a file.
• Original Image: shows the loaded or generated input before adaptation.
• Adapted Image: shows the output at the current time value.
• Live: when enabled, slider and dropdown changes re-render automatically.
• Run: manually processes the current image when Live is off or when you want a deliberate refresh.
• Privacy: no upload service is used. File reading, image drawing, adaptation, charts, and exports stay on-device.

• Source illuminant: the original viewing white point. Default is CIE D65 for sRGB daylight.
• Destination illuminant: the target viewing white point. Default is CIE D50 for ICC PCS and print workflows.
• Supported illuminants: D65, D50, D55, D60, D75, A, B, C, E, F2, F7, and F11.
• Swap: flips source and destination to test the inverse adaptation direction.
• CAT methods: Bradford, Von Kries, CAT02, CAT16, Sharp, CMCCAT2000, HPE, and XYZ Scaling.
• Bradford: a strong baseline for ICC-oriented D65 to D50 workflows and large daylight shifts.
• CAT16: a modern option for new appearance-model work and high-chroma stability.
• XYZ Scaling: useful as a simple reference or lower-bound comparison, not as a production-quality perceptual model.

The lab drives the adaptation degree with the two-component time course measured by Fairchild and Reniff (1995): about half of the adaptation happens within a second, and the rest follows a slow phase with a 30-second half-life. In plain terms, the image begins at an initial adaptation value, then moves toward the degree of adaptation D along that curve.

• Dynamics tab: plots the two-component curve against the single exponential the page used to run (tau 0.6 seconds after a 0.4-second delay) from 0 to 120 seconds, with half-adapted and 90 percent times.
• Time-course controls: in the Dynamics tab, choose Two components or Single exponential, and set the fast time constant (0.2 to 4 seconds), its share of the adaptation (0 to 1), the slow half-life (5 to 120 seconds), and the playhead. These redraw the Dynamics plot only; the Lab image runs the measured model at the paper's defaults.
• Degree of adaptation D: final adaptation target from 0 to 1. Use 1 for full adaptation and lower values for partial adaptation.
• Initial adaptation: starting adaptation value at time 0, 0.1 by default. It has no slider; a share URL's init parameter sets it.
• Tau luminance scale: lets darker pixels adapt more slowly by increasing their per-pixel tau.
• Delay luminance scale: lets darker pixels begin adaptation later by increasing their per-pixel delay.
• Mix: blends global A(t) with per-pixel luminance-dependent A(t). Mix 0 is fully global; mix 1 is fully local.
• Time: scrubber from 0 to 5 seconds. Use it for still-frame inspection.
• Target FPS: animation target from 8 to 120 frames per second.
• HDR Tonemap: applies a Reinhard-style compression before output clamping when large illuminant shifts push values beyond display range.

• Info chip: reports readiness, image load status, animation status, and reset state.
• Params chip: shows the time-course model, the current time, and the current degree of adaptation D.
• Adapted Image: the transformed image for the selected method, illuminants, and time.
• RGB Histogram: red, green, and blue distributions for the adapted frame.
• Adaptation Curve A(t): plots the full five-second curve and the active time marker.
• State Map: grayscale map of local adaptation. White is more adapted; black is less adapted.
• LMS Cone-Channel Analysis: compares source and destination cone-response channels for the active matrix.
• SPD Spectral Overlay: plots source and destination illuminant spectra: the D series from the CIE daylight basis, A from Planck's law, and E as a flat spectrum. B, C, F2, F7, and F11 have no spectrum held, so the plot says so instead of drawing a curve.
• Matrix tab: shows M, diag(d) with the L, M, S scale factors, and the full adaptation matrix at 4, 6, or 9 decimals, a round-trip check, copy formats, and all eight transforms compared with Bradford.

• Export JSON: downloads state parameters, the full matrix, LMS scale factors, determinant, and current timing values.
• Export Frame: downloads the adapted canvas as a PNG at the current time.
• Copy Link: copies the current page link.
• Generate share URL: writes a URL with source, destination, CAT method, tau, delay, strength, initial adaptation, and time query parameters.
• Compare All Methods: runs all eight CAT methods against a standard color set and reports mean, median, min, max, determinant, and matrix norm diagnostics.
• Comparison chart: visualizes method difference scores so method behavior is easier to compare quickly.
• Batch Colour Analysis: accepts up to 50 HEX values and reports adapted HEX, Delta E 2000, L*, C*, and LCh hue for each color.

• Standards: use this section when you need context for CIE 160, CIE 159, CAM16, ICC profile adaptation, time-course studies, asymmetric color matching, and color constancy.
• Formulas: review the exponential curve, diagonal adaptation matrix, CAT matrices, CIELAB and Delta E 2000 logic, SPD generation, and spatial adaptation equations.
• Citations: cite source literature and standards when preparing reports or technical notes.
• Research notes: interpret half-life, 90 percent adaptation time, state maps, CIECAM02 D, CAT method caveats, and gamut clipping behavior.
• Production warning: the tool is an educational and research simulator. It is not a substitute for measured device profiles, calibrated displays, spectrophotometer data, or full ICC workflow validation.

• JSON export: best for reproducible research notes because it includes the active method, illuminants, timing, spatial parameters, matrix, scale factors, and determinant.
• PNG export: best for slide decks, issue comments, and quick visual review.
• Share URL: best for sharing parameter state. It does not embed uploaded image pixels.
• Library save: the universal Library captures this tool as an adaptation preset with form inputs and source metadata.
• Restore: saved form values can be pushed back into the page; the tool also exposes a page API that can restore the internal state object.
• Image note: exported and shared parameter states do not persist the uploaded source image unless you separately keep the file or exported frame.

• D65 to D50 print check: load a sample, keep source D65, destination D50, start with Bradford, then compare CAT16 and CAT02.
• Lighting transition demo: choose D65 to A, set the degree of adaptation D to 1, and animate to show the white-point shift gradually.
• Partial adaptation study: lower D below 1 and compare output against full adaptation to illustrate incomplete constancy.
• Local adaptation study: increase tau luminance scale, delay luminance scale, and mix, then inspect the state map for dark-region lag.
• Method comparison: keep time fixed, run Compare All Methods, and use the mean Delta E 2000 chart to choose a method for a report.
• Brand color audit: paste brand HEX values into Batch Colour Analysis, then export JSON and save the preset to Library.

• Nothing changes: load a sample or image first, enable Live, or click Run after changing parameters.
• Animation feels fast: scrub manually with the Time control, or open the Dynamics tab to read the full 120-second time course.
• Output clips or looks harsh: enable HDR Tonemap, use a smaller illuminant shift, or compare CAT16 and Bradford before judging.
• Batch output is empty: enter valid six-digit HEX values, one per line. Invalid rows are ignored, and analysis is capped at 50 colors.
• Share URL is missing image content: share URLs encode parameters only. Send the source image separately or export the PNG frame.
• GPU checkbox does not change output: the current renderer is a CPU pipeline; the checkbox is stored as state for future acceleration experiments.
• Some spatial controls change nothing: local radius, local strength and tile factor are saved with the settings but not applied to the image yet. The state map follows tau luminance scale, delay luminance scale, mix, and the current time.