Comment by andrewla

Comment by andrewla 13 hours ago

3 replies

It seems that this would be well-suited to a simple online test -- show a square with one color and a square inside of that with a different color, and ask the user whether the inner square is brighter (or too close to call). Aggregate this across users and assess the fit to the CEILAB or other color spaces. It seems like you could get almost all hn users to take a stab at this for a bit before they get sick of it.

hatthew 13 hours ago

precise results would depend heavily on screen calibration

  • tshaddox 8 hours ago

    True, but the results would likely still be useful, particularly when choosing colors for web sites which will be displayed on the same distribution of screens.

  • andrewla 12 hours ago

    I think this is likely, but easily validated -- you'd probably see a couple of clusters or some latent variables, but the raw data would still likely be useful for getting average response characteristics.