Reflective display · color science · software
Color management for a six-state display
Can color conversion improve reproduction while keeping the same Pimoroni renderer? Experimental profiles reduced average error on a measured color grid.
Lower measured error on a 133-color grid
I tried color management to bring reproduced colors closer to the requested image while keeping the same Pimoroni renderer. An experimental ICC profile described the color response of that rendering-and-display setup under its characterization conditions. A color-management system used it to convert source sRGB colors into new RGB values for the fixed renderer. On a separate 133-color grid, the dense-profile route reduced mean measured error from 19.9922 to 11.2006 ΔE00.
Vincent van Gogh, Cypresses, 1889. The Metropolitan Museum of Art, 49.30, Rogers Fund, 1949. Public domain. The generated maps use measured native-state colors, scaled so the measured display white appears as screen white. They do not simulate absolute brightness, contrast or spatial interactions. The 133-color comparison below supplies the optical measurement of the two routes.
I could also see the difference when displaying the images on the physical panel. That is a visual observation; the numerical comparison below comes from a separate color grid.
See the same comparison on The Bedroom
Vincent van Gogh, The Bedroom, 1889. The Art Institute of Chicago, 1926.417, Helen Birch Bartlett Memorial Collection. Public domain. The generated maps use the same measured native-state colors, scaled so display white appears as screen white.
Lower average error on a separate 133-color grid
The generated pixel maps show which native-state choices changed. A separate test measured 133 source colors to compare the routes numerically. Lower ΔE00 means closer to the requested color. Under that campaign's conditions, the dense ICC-managed route had a lower average color error than the stock route. This 133-color comparison did not evaluate the project renderer.
| Rendering route | Mean error | Largest error |
|---|---|---|
| Pimoroni renderer | 19.9922 | 42.1941 |
| Sparse ICC + Pimoroni renderer | 12.2568 | 26.4508 |
| Dense ICC + Pimoroni renderer | 11.2006 | 23.7069 |
The optical comparison used a separate 133-color grid rather than the paintings shown above. Errors used measured-media-relative D50 Lab against each requested target, averaging repeated errors per source and then the 133 source values equally. The dense route had the lower overall mean while individual colors varied. This development comparison applies to the tested renderer, display and measurement condition.
How the two profiles were constructed
“Sparse” and “dense” describe the characterization grids used to construct the profiles. The dense version used a modeled grid built from measured native states and an optical mixture model. The artwork route used relative-colorimetric conversion before the unchanged palette separation.
The ICC-managed route reduced average measured error for the tested grid. To improve reproduction further, I investigated whether arrangements with the same native-color proportions would produce the same measured color. They did not always do so, which made spatial organization a further question for characterization and rendering.