I have proposed that three Fourier interactions (RGB) are defined within the eye following from light refraction ( the chromatic “aberration” of the eye) and a retina definitively defined as being sensitive to (or “tuned” to) only three wavelengths (RGB). The first transform is effected on the all-cone fovea and results solely from long wavelength (R) interactions (in fact, this long wavelength interaction precisely defines this end of the visible band). This interaction satisfies the Fourier equation (i.e, both light intensity and phase- encoding direction are detected). Thus it must be that this information is spatially correlated—the “image” itself is detected. This is the colorless (“color” has yet to be added) “outline sketch” that David Marr brilliantly deduced must be present and defined it (paraphrased) as “the first stage of the vision process”. I have referenced the optical (analogous to Fourier) transform of a black and white outline sketch showing that it is predominantly a small central “dot”… mimicking the fovea. This is what you see in the darkness of your bedroom… an outline sketch without color.
Then… how is the “color signal” added at increased light levels and spatially correlated with the R outline sketch to arrive at the beautiful color image perceived by the brain?
I have proposed that the central, geometrically-determined, mid-band or G transform occurs at some (small but real!) distance above the foveal R transform. Mid-band radiation is refracted to a ring at 7-8 degrees of retinal angle. I must stop here and remind that it is the brightness of radiation at this wavelength (G) that is detected on either side of the 7-8 degree radius—it is crucial that one grasp this!
It seems that two separate mechanisms must be addressed here—
- how is the the “color signal” itself determined? and…
- how is this signal spatially correlated (“overlayed on”) with the foveally-determined (R), outline sketch?
As to (1.), Edwin Land’s theory enters here..and it is almost uncanny how one can see on the retina what Land found in experiments and color simulations external to the eye, i.e., a precisely determined mid-band point on the retinal surface with brightnesses (his “lightnesses”) – NOT wavelengths – determined on either side! All that needs to be be worked out is the mechanism for “ratioing” these two brightness values… but I leave that to someone versed in retinal circuitry.
I note again that the G transform (that I will term the “Land color transform”) is temporally separated (it is a distinctly separate signal in time) from the R, fundamental image sketch transform! This again is exactly how Land simulated this externally using separate Retinex signals.
The G color signal is separated in time from the R fundamental outline sketch signal
As to (2.), it must be that the R and G signals are spatially correlated. this is logical as they derive axially from the same perceived image! It is really dramatic how the color signal is determined and then added to each pixel (picture element) of the fundamental R image… but this must be so!
The third… or B transform… I come to believe that this has two functions… (a) is determinative of the short wavelength end of the visible spectrum (precisely determining that wavelength!) and controlling pupillary constriction and overall light entrance into the eye. I do not think that it has any color-determinative function. Again, color is determined by brightnesses on either side of the G peak.
I leave out for the time being quantum considerations between the RGB distinctly temporally separated signals. I do not know the “order” of these signals… or if they have any order at all. They still connote to me, however, some idea of an “arrow of time”.
(I have dashed this off and will edit shortly)