American cinematographer (Jan-Dec 1959)

GREEN (y) BEUE (-2 Violet Purple Magenta VALUES OF ^ ► z = I — jc highly sensitive color negatives are bal¬ anced for the lower (yellower) color temperature of 3200° K. To these films, tungsten light is “white.” Cer¬ tain brands of reversal color films are manufactured in two types, one bal¬ anced for 5000° K for outdoor use, and the other for 3200° K for indoor use with photoflood lamps. The two films may be interchanged (with a slight loss of speed) if appropriate cor¬ recting filters are placed over the cam¬ era lens — a bluish filter for the outdoor film used indoors, and a yellowish filter for the indoor film used outdoors. There are so many “whites” that it is perfectly proper to regard any and all of them as colors — colors to be re¬ corded as faithfully as red, yellow, lav¬ ender, or flesh-tint by the color film or camera. But how are so many colors faithfully recorded and reproduced? A color photographic film or a color TV camera works on the same simple theory which has often been used to explain human color perception, name¬ ly, that the retina of the eye contains three color receptors, each of which re¬ sponds to roughly one-third of the en¬ tire visible spectrum of colors. The “red receptor” sees the wavelengths of light from the longwave extreme in the red through orange, yellow, and char¬ treuse to the green region. The “green receptor” sees the wavelengths from FIG. 4 — The color gamuts resulting from certain arbitrary sets of primary colors plotted on the chromaticity diagram. The 700-525-460 millimicron set encloses the widest range of colors, but most of the vivid spectrum colors lie outside even this sotisfactory gamut. FIG. 2 — An “oversimplifted” color triangle. The colors produced by additively mixing the primaries, red, green, and blue are arranged along the sides of the triangle. Experiment proves, however, that most of the spectrum colors are more vivid than these additive mixture colors. FIG. 3 — The C.I.E. chromaticity diagram. By postulating three “supersaturated" pri¬ maries, the spectrum colors can be included in the color triangle, where they form a horseshoe-shaped curve known as the spec¬ trum locus. All real colors are contained inside this locus and the "purple line" which connects its two extremities. orange-red through yellow, chartreuse, green, and cyan to blue. The “blue re¬ ceptor” sees the wavelengths from green through cyan, blue, and blueviolet to the shortwave extreme in the violet (Fig. 1). The overlapping nature of these three bands of color perception — red, green, and blue — results in the inter¬ mediate hues named above; and a re-emergence of red sensitivity in the shortwave region of the spectrum, where blue is the predominant color, is responsible for the violet color of the shortest visible wavelengths. The purples and magentas are hues which do not occur in the spectrum, but they can be produced with maximum purity by combining violet light with red light in various proportions. Whether this “trichromatic theory” offers the true explanation of color vision or not (scientists are unde¬ cided), it stands us in good stead when we wish to devise a practicable color-reproducing system for photog¬ 400'“ 380 mu VIOLET ' raphy or television. The requirements for a workable color system are thus very simple in theory, though un¬ deniably beset by numerous “bugs” in practice. Briefly, in order to photograph a scene in full natural color, we must separately record the red, green, and blue components of the scene. This may be done by using a film having three color-sensitized emulsions (each of which records approximately onethird of the spectrum) or, more directly, by using a camera fitted with an optical beam splitter and three separate films. The beam splitter produces three identical images of the scene upon which the camera is focused. Over each image is placed a color filter whose transmittance is nearly identical with the response characteristics of the corresponding primary color receptor of the eye. This gives us red, green, and blue “records” of the photo¬ graphed scene in terms of black-andwhite values on a balanced, redsensitized panchromatic negative. Three positive transparencies are then printed from the negatives; and these are superposed on a projection screen by three identical projectors set up very close together. Over the lens of the projector showing the print of the red record is placed a red filter, and so with the projectors showing the green and blue-record prints. The original scene is reproduced on the screen in accurate natural color. Although the 3-projector process is only rarely employed, it illustrates the “additive” systems advocated for tele¬ vision and makes all other color pro¬ cesses more comprehensible. Continued on Page 568 SEPTEMBER • 1959 543