Fig 1.
(A). Opponent color chromatic responses y-b, r-g, over the spectrum, adapted from Reference 9. This shows r response as two spectral peaks about 440 and 610 nm. (B). Same as Fig (A) but with chromatic responses shown over the whole hue cycle, with nonspectral purples (beyond the effective spectrum extremes at 442 and 613 nm, wavelength limits to monochromatic optimal color stimuli [16]) given an arbitrary interval equivalent to some 70 nm on x-axis. This shows the r curve peak (colorimetrically calculated from the indicated two spectral r peaks of relative amounts 0.67:1 ratio) about 510 c, with unique Red (labeled *R) about 494 c from previous data. The nonspectral purples are indicated by their complementary wavelengths, e.g, 510 c.
Fig 2.
Test results for SO, one subject of 5 who were tested for both blueness (blue line) and redness (red line), graphed from Ref. 1 data.
The x-axis shows S cone excitation and dominant wavelength digitized from Fig 1 of Ref. 1 and CIE table of chromaticity coordinates for wavelengths. Peak blueness occurs near unique Blue (470 nm), where redness first commences.
Fig 3.
Hue naming and hue prediction data from Werner & Wooten 1979 [17] with permission.
Black solid lines: Hue naming data for 3 subjects (mean). Black dotted lines: Hue prediction for same 3 subjects using Hurvich & Jameson model [13]. Arrows indicate unique hue loci at 100% R or G (left ordinate), and B or Y (right ordinate).
Fig 4.
Traditional schema or wiring diagram encoding SML cone outputs to opponent color chromatic responses r—g, y—b.
Removing dashed grey line (contentious M input to y) transforms the schema to the simpler more symmetrical model in Ref. 18, wh+ere S = b, M = g, L = y and, as in tradition, S + L = r. In this version of wiring diagram [18] (with grey line omitted), the horizontal black links from SML cones to bgy chromatic responses represent cone primary outputs, and the two red lines represent cone secondary outputs where two cones S and L combine to form the redness chromatic response.
Fig 5.
Black curves: Cone and chromatic response curves from the cited sources, normalized at 1.0 response and with wavelength peaks aligned with Table 1 means.
Vertical dashed arrowed lines show identical wavelengths ± 2 or 3 nm for curve peaks in Fig (A) and curve peaks in Fig (B). (A). Black curves: Cone sensitivity estimates of Stockman & Sharpe [34]. Arrowed horizontal line from S and L curves’ intersection [equivalent to intersection of b and y chromatic response curves in Fig (B)] indicates null response of new chromatic response model at right y-axis. (This interval 0–1 is equal to 0–1 interval in Fig (B) below, allowing comparison of curves to the same shared scale). Cone sensitivities below intersection of S and L curves (arrowed line) are shown in gray, and are discounted from comparison. (B). Red lines: Spectral (bgy) opponent chromatic response curves from Hurvich in Fig 1A. Red curves: Spectral lobes for r chromatic response normalised at 0.5 response.
Table 1.
Wavelength peaks of human cone spectral sensitivities and of opponent chromatic responses from hue cancellation experiments (from Pridmore [18]).
Fig 6.
Curve fitting prediction of cone sensitivities and unique hue chromatic responses.
A. Three formulaic curves in blue [plotted from Eqns (4)-(6) of Ref. 18] predicting the SML cone curves and the bgy chromatic response curves. Their wavelength peaks are nominally 445, 535, 565 nm but can be shifted horizontally to any wavelength without changing curve shape. B. The formulaic curves (in blue) compared to SML cone response curves in black (from Fig 5 above) by overlapping the former’s wavelength peaks with the latter as labeled. C. The same formulaic curves compared to bgy chromatic response curves in red (from Fig 5) by overlapping the former’s wavelength peaks with the latter as labeled. The predicted curves (in blue) correlate very closely with cone response curves to a mean 0.987 correlation coefficient, and with chromatic response curves to a mean 0.967 correlation coefficient [18].