<--back--      home      --next-->

In order to determine if our face-based method is doing luminance or brightness matching, we looked in the photometry literature for previous methods of luminance matching. There are two main ones:

Heterochromatic Flicker Photometry exploits the difference in the temporal resolution of our perception of luminance versus saturation by alternating, about 20 Hz, between a color and a gray. If there's no visible flicker, the luminances match. But this is hard to reliably implement on any old monitor and graphics system, its pretty annoying to look at, and there's a very real risk of inducing a photo-sensitive epileptic seizures.

The other method is called Minimally distinct border, or MDB. In this method, the user looks at a sharp boundary where a gray and a color are adjacent, and then adjusts one side until the boundary between the two looks "minimally distinct". The boundary will be very distinct in this case, or this case, and somewhere in the middle it will be minimally distinct.

What's interesting about these methods is that the user is never subjectively comparing luminances-- that's impossible-- but with some care, a stimulus can be set up in such a way that it indirectly indicates luminance equality: the lack of a flicker, or the minimal distinctness of a boundary.

Or, perhaps the balance between which side the positive face appears on, in our face-based method, is another indicator of lumiance equality.