Abstract

The visual mechanism of a normal observer is so constructed that objects keep nearly their daylight colors even when the illuminant departs markedly from average daylight. The processes by means of which the observer adapts to the illuminant or discounts most of the effect of a nondaylight illuminant are complicated; they are known to be partly retinal and partly cortical. By taking into account the various fragments of both qualitative and quantitative information to be found in the literature, relations have been formulated by means of which it is possible to compute approximately the hue, saturation, and lightness (tint, value) of a surface color from the tristimulus specifications of the light reflected from the surface and of the light reflected from the background against which it is viewed. Preliminary observations of 15 surfaces under each of 5 different illuminants have demonstrated the adequacy of the formulation and have led to an approximate evaluation of the constants appearing in it. More detailed and extensive observations have been carried out in the psychological laboratories of Bryn Mawr College. and these observations have resulted in an improved formulation.

© 1940 Optical Society of America

Full Article  |  PDF Article

Corrections

Deane B. Judd, "Erratum: Hue, Saturation, and Lightness of Surface Colors with Chromatic Illumination," J. Opt. Soc. Am. 30, 296-296 (1940)
https://www.osapublishing.org/josa/abstract.cfm?uri=josa-30-7-296

Cited By

OSA participates in Crossref's Cited-By Linking service. Citing articles from OSA journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (4)

Fig. 1
Fig. 1

The 75 illuminant-sample combinations represented on the uniform-chromaticity-scale Maxwell triangle. The samples with Abbot-Priest sunlight as the illuminant are shown by circles; those for the four strongly selective illuminants (red, yellow, green, blue) are shown by crosses. The illuminants are indicated by squares.

Fig. 2
Fig. 2

A large-scale plot of a portion of Fig. 1 showing for the red illuminant (square) the locus of the achromatic points (dotted line) and two of the vectors defining hue and saturation by Eqs. (1), (2) and (3). The definition of hue according to angle by Eq. (1) is indicated in the upper left-hand part of this figure. The vector indicating a blue hue refers to a dark nonselective sample on a light ground; and that indicating a yellow-red hue, to a light nonselective sample on a light ground. Both vectors refer to momentary fixation of the samples by an observer who has been exposed to the field with red illumination for five minutes or more (Case 3, Section III, 2, e).

Fig. 3
Fig. 3

The loci of achromatic points (dotted lines) on the uniform-chromaticity-scale Maxwell triangle for the five illuminants studied. Note the varying lengths of the loci (determined chiefly by Df), and the curvature of the loci of achromatic points for the blue illuminant (determined by bf). Note also how the direction of each nearly straight locus is determined by the position of the point (r=0.36, g=0.30), indicated by a double hexagon, whose trilinear coordinates appear explicitly in Eq. (3). The definition of hue according to angle by Eq. (1) is indicated graphically by lines intersecting at the daylight point (indicated by a small single hexagon).

Fig. 4
Fig. 4

Comparison of estimated lightness, L, with that computed from Eqs. (4) and (5a). The agreement is generally good with the exception of dark samples on a light background for nonred illuminants. See Section VIII for discussion of this discrepancy.

Tables (12)

Tables Icon

Table II Spectrum stimuli for various hues.

Tables Icon

Table III Adaptation reflectance, A′.

Tables Icon

Table IV Spectral apparent reflectances for the 15 Munsell samples studied.

Tables Icon

Table V Spectral transmissions of the viewing filters.

Tables Icon

Table VI Trilinear coordinates (r,g) of the light reflected from each sample and the apparent reflectances (A) of the samples for each of the five illuminants used.

Tables Icon

Table VII Constants used in computing the trilinear coordinates (rn,gn,bn) of the achromatic points for the five illuminants studied (Eq. 3).

Tables Icon

Table IX Computed hues, lightnesses and saturations for 3 hypothetical nonselective samples for both light and dark backgrounds, for the 5 illuminants, and for 4 observing situations (Cases 1 to 4).

Tables Icon

Table X Comparison of observed and computed color descriptions. (Hue, lightness and saturation (HL/S) of the 15 selective samples studied. Illuminant: Abbot-Priest sunlight taken as representative of south daylight.)

Tables Icon

Table XI Comparison of observed and computed color descriptions. Hue, lightness and saturation (HL/S) of the 15 selective samples studied under the 4 chromatic illuminants; see Table V.

Tables Icon

Table XII Number of sample-illuminant-background combinations correct.

Tables Icon

Table XIII Effect on the color of a sample caused by substitution of a dark for a light background.

Equations (10)

Equations on this page are rendered with MathJax. Learn more.

H = f [ ( r - r n ) / ( g - g n ) ,             s g n ( r - r n ) ] .
S = 50 D = 50 [ ( r - r n ) 2 + ( g - g n ) 2 + ( b - b n ) 2 ] 1 2 .
r n = r f - D f [ 0.1 L ( r f - 0.360 ) - 0.018 b f A f ( L ) 2 log 10 2000 I ] g n = g f - D f [ 0.1 L ( g f - 0.300 ) - 0.030 ] } .
r n = r f - k ( r f - r w ) g n = g f - k ( g f - g w ) } ,
L = 10 ( L - L A = 0.03 ) / ( L A = 1.00 - L A = 0.03 ) .
L = 10 ( A - 0.03 ) ( A f + 1.00 ) ( 1.00 - 0.03 ) ( A f + A ) .
A f = ( A 0 + Ā ) / 2.
A f = ( n A + A 0 + Ā ) / ( n + 2 ) .
A f = A .
A f = ( 2 A + A 0 + Ā ) / 4.