US2272638A - Method of color reproduction - Google Patents

Description

Feb. 10, 1942. A. c. HARDY METHOD 0F COLOR REPRODUCTION Original Filed Sept. 4, 1936 6 Sheets-Sheet 1 INVENTOR Rr//z/RZ /Awy ATTORNEY Feb. 10, 1942. 'A. c. HARDY METHOD OF COLOR REPRODUCTION 6' Sheets-Sheet 2 original Filed Sept. 4, 19:56

Feb. l0, 1942. A. c. HARDY METHOD 0F COLOR REPRODUCTION Original Fivled Sept. 4; 1956 6 SheetS-Sheet 3 v h@ @www Feb. l0, 1942. A. c. HARDY METHQD 0F COLORREPRODUCTION f original Filed sept. 4, 1936 i QN@ M N m Y Fem w, i942. A. c. HARDY METHOD OF COLOR REPRODUCTION Original Filed Sept. 4, 1956 6 Sheets-Sheet 5 wNw INVENTOR WA/H? ff/my Feb. 10, E942. A. c. HARDY METHOD OF GOLOKREPRODUGTION Original Filed Sept. 4, 1956 6 Sheets-Sheet 6 IFE ATTORNEY Patented Feb. 10, 1942 i f UNITED STATES PATENT OFFICE MTHOD F COLOR REPRODUCTION Arthur C. Hardy, Wellesley, Mass., assignor to Interchemical Corporation, New York, N. Y., a corporation oi' Ohio Original application September`4, 1936, Serial No. 99,415, now Patent No. 2,193,722, dated March `12, 1940. Divided and this application March 2c, 193s, serial No. 198,240

(ci. 17a-1.5.2)

4 Claims.

This invention relates to a method o f color reproduction and aims to provide amore exact reproduction by the three-color method than has heretofore been obtained. 1

The present application is a division of my ap..r

' is, by the proportionate strength of radiations of different Wave lengths which constitute the color. The spectral energy distribution of a color may be plotted as a curve whose abscissae represent diierent wave lengths and Whoseordinates indicate the relative strength of radiation at each wave length. The spectral quality of a color determines the stimulus required to produce the same color sensation. lThe stimulus may be dened by two factors, dominant wave length and purity, or by factors termed trichromatic coeiiicients.

The primaries of a color reproduction system are the colors of the colored lights which are mixed to give the observer a color sensation intended to duplicate the sensation which he would receive from colored light from the original subject.

A color-separation image is an image of the subject which is used to control one of the primaries in making a reproduction. In projection systems of color reproduction, the color-separation image may be a transparent positive which directly controls a colored light constituting one of the primaries. In systems of color reproduction used in the graphic arts, the color-separation images are formed upon or transferred to printing members, so that they control the primaries represented by the colored inks applied. Although color-separation images are not in themselves colored, they are frequently identified by the names of the colors of the primaries which they control. Thus the expression redcolorseparation image means a color-separation image to be used for controlling a primary Whose dominant wave length is in the red part of the spectrum.

A receptor is a material or device which undergoes some change when subjected to radiant energy in the form of light and thus serves to make a record of the amount of light which it then placed in three separate projection lanterns receives. The spectral .sensitivity of a receptor is its relative response to light of different wave lengths and may be indicated by a curve in which the abscissae represent diierent wave lengths' and the ordinates represent the relative extent to which the receptor is modified by radiation at each wave length.

To further clarify the meaning of these terms, I will describe the simplest form of three-color reproduction. This consists in 'making colorseparation imagesby photographing the subject three times, once througha red lter, once through a green filter and once through a blue iilter. In the taking of each photograph; the filter and the photographic plate combined constitute a receptor; and the spectral sensitivity of this receptor is determined by the spectral transmission characteristics of the filter and the spectral sensitivity of the photographic emulsion on each containing a White light shining through a red, green or blue lter. 'Ihe three images are superimposed on the same screen. The lights and the projection filters determine the primaries of the system. Thus, for example, the colored light coming through the red projection lter constitutes the red primary of the system. The spectral energy distribution of each primary is determined by the spectral energy distribution of the lamp in the lantern and the spectral transmission characteristic of the filter used with the lamp. In making the reproduction on the screen, the primaries are controlled by the color-separation images, that is to say, the three color-separation positives determine the proportions' of the mixture of the three primaries at each point of the reproduction on the screen.

It has been customary to assume that, in a reproduction system such as that which has been described, the color filters used in making the three color-separation images should be the same as the color lters used in making the reproduction. Departures from this assumption in practice have, for the most part, been based on empirical attempts to improve color reproduction. A notable exception to these empirical attempts is the method described by F. E. Ives, U. S. Patent No. 432,530, Where it is proposed to base the colorseparation lters on Maxwells curves for the pri' I have ascertained that the common assumption that the color-separation filters and the color projection filters should be the same is cornpletely erroneous, and I have devised a method by which the spectral sensitivity characteristics of the receptors which should be used to make the color-separation images may be computed from either the spectral quality or the equivaient stimulus of each of the three primaries to be used in the reproduction.

This method consists in first determining either from the spcctrophotometric characteristics of the three primaries, or from the equivalent stimulus of each primary, the trichromatic coefficients of the primaries which maybe represented as rareza, :cor/ecc, and :mysan (Handbook of Colorimetry, Massachusetts Institute of Technology, 1936, pp. 9-13). I have found mathematically that the spectral sensitivity characteristics of the three receptors required to produce correct color-separation images for the control of primaries having the trichromatic coemcients above stated are determined from the following set of linear equations: i

wherein Sn, Se and Sa are the desired spectral sensitivities of the receptors, and rr, y and z are the basic data concerning the chromatic properties of the human eye, such as those published by the International Commission on Illumination.

By computations based on the above equations, I have ascertained that, in all cases where the primaries are real colors, the required spectral sensitivities of at least one of the receptors for making the separation images contains negative values. My invention gives physical effect to these negative values.

The method of color reproduction which I have invented consists in controlling one or more of the primaries used in making the reproduction by means of a. color-separation image in which light from one spectral region has had an eifect opposite to that of light from another spectral region. My invention includes also the method of making such a color-separation image by subtracting the eifect of light on one receptor Kfrom the effect of the same light on another receptor. The invention includes photographic and electrical methods of effecting this subtraction.

To clarify the nature of my invention, I will give a specific example of the use of the invention in the simplest method of color reproduction, that is, by three-lantern projection. In this example, I shall refer to the accompanying diagrams which are graphs of spectral energy distribution and spectral sensitivity characteristics plotted over the extent of the visible spectrum. In the diagrams,

Figs. 1, 2 and 3 show the spectral transmission characteristics of a red, a green and a blue projection filter;

Fig. 4 shows the spectral energy distribution of a lamp used in each projection lantern;

Figs. 5, 6 and '7 show the spectral energy distribution of the primaries of the system;

Figs. 8, 9 and 10 show the spectral sensitivities of the ideal or theoretical receptors required for making red. green and blue color-separation images to control the primaries whose spectral energy distributions are shown in Figs. 5, 6 and '7;

Figs. 11 and 12 show the sensitivities of two practical receptors used in making the red colorseparation image and Figs. 11a and 12a show the iilters forming part of these receptors;

Fig. 13 shows the result of subtracting the effect of light on the receptor whose sensitivity is shown in Fig. 12 from the eiect of the same light on the receptor Whose sensitivity is shown in Fig. 11;

Figs. 14, 14a, 15, 15a and 16 correspond to Figs. 11, 11a, 12, 12a and 13 and relate to the green color-separation image;

Figs. 17, 17a, 18, 18a and 19 correspond to Figs. 1l, 11a, 12, 12a and 13 and relate to the blue color-separation image; and

Fig. 20 is a diagrammatic View of an electrical apparatus for picture reproduction by scanning, which is used in one modification of my invention, and Fig. 21 shows a different electric connection which may be used with said scanning apparatus.

In the specific example which I shall describe, the reproduction is to be made bythree images superimposed on a screen and projected from three lanterns provided with projection filters whose spectral transmission characteristics are shown in Figs. l, 2 and 3. The three filters are the red, green and blue iilters sold by the Eastman Kodak Company and identified as Wratten Filters, Nos. 24, 59 and 47. The spectral transmission characteristics of these three filters shown in Figs. 1, 2 and 3 are taken from Wratten Light Filters, 13th edition, revised 1934, published by the Eastman Kodak Company. The light used in each lantern is a tungsten lamp operated at a temperature of 2848" K. The spectral energy distribution of this lamp is shown in Fig. '4, which is based on data published by the International Commission on Illumination (Handbook of Coiorimetry, Massachusetts Institute of Technology, 1936, p. 16).

The first step in the use ofmy invention is to determine the spectral qualities of the primaries of the reproduction. In the present instance, the primaries are the colored lights emitted from the three projection lanterns. Their spectral energy distribution may be obtained hy multiplying the spectral energy distribution oieration images for use in controlling these three primaries. From these curves, the trichromatic coeililcients of the primaries are found to be as follows:

Red primary Green primary Blue primary z=0. (i749 1:0. 2944 z=0. 1369 y=0. 3218 y=0. 6069 v y=0. 0710 z=0. 0002 z=0. 0977 z=0. 7921 (id. pp. 8, 32, 33, 49. 50). By substituting these values and the values of cc, 'y, 2.7, published by the International Commission on Illumination (id. p. 7, p. 35, table XII) in the equations above given, the spectral sensitivities of the three receptors required to produce correct color-separation imltical receptor of Fig. 11.

and 10. From these figures, it is evident that the required spectralV sensitivities of the three receptors diier widely from the transmission characteristics of the projection filters shown in Figs 1, 2 and 3, and from the spectral energy distribution of the primaries shown in Figs. 5, 6 and 7, and that the required spectral sensitivity of each receptor is negative in a part of the spectrum.

The next step is to produce a red color-separation image in which light in the spectral region including wave lengths from 470 toI 550 (in which the curve shown in Fig. 8 is negative) has had an eiect opposite to that of light from the rest of the visible spectrum. This step may be carried out as follows:

A practical receptor having a spectral sensitivity corresponding to the positive part of the required sensitivity curve shown in Fig. 8 is provided. Such a receptor is made by the use of a panchromatic photographic plate, such'as that described by L. A. Jones and Otto Sandvik in the April, 1926, issue of the Journal of the Optical Society of America and Review of Scientilc Instruments, with the filter shown in Fig. 11a, which consists of a 98.2% sector of Wratten Filter No. 22 and 1.8% sector of Wratten Filter No. 47A. tained by making successive exposures of the plate through Wratten Filters Nos. 22 and 47A,

the relative length of the exposures corresponding to the relative size of the sectors shown in Fig. 11a.) The spectral sensitivity of this practical receptor is indicated in a. full line in Fig. l1. It will be seen that the spectral sensitivity of this practical receptor approximates the positive part of the theoretically correct red receptor which is shown in Fig. 8 and in a dotted line in Fig. 11 for the purpose of comparison.

A practical receptor having a spectral sensitivity approximating that of a mirror image of the negative part of the spectral sensitivity curve of the theoretical red receptor shown in Fig. 8 is provided. This practical receptor is made by combining the panchromatic plate above referred to with the lter shown in Fig 12a, which consists of a 4.7% sector of Wratten Filter No. 61, a 13.8% sector of Wratten Filter No. 29, and an 81.5% sector of Wratten Filters Nos. 22 and 53 superposed. The spectral sensitivity of this practical receptor is shown in Fig 12. It will be observed that', in a part of the spectrum, the sensitivity of this practical receptor approximates a mirror image of the negative part of the spectral sensitivity curve of the theoretical red receptor which is shown in Fig.-8 and shown in a dotted line in Fig. 12 for the purpose of comparison. It will be noted also that the sensitivity of the practical receptor is greater than that of the negative part of the curve for the theoretical receptor between the wave lengths 550 and 700,

-and that, in the same region, the curve of practical receptor shown in Fig. 11 vexceeds the sensitivity curve of the theoretical receptor.

The effect of light from the subject on the practical receptor of Fig. 12 is subtracted from the effect of light from the subject on the prac- 'I'his is accomplished in the following manner: A negative image is made with a receptor of Fig. 11 using short exposures, so that the chemical eiect on the plate as measured by the transparency of the developed plate bears a linear relation to the (A similar result can, of course, be obamount of light received. In a similar manner, a negative is made with theV receptor of Fig. 12. From this negative a positive is made by contact printing. A third plate is then exposed successively through the negative made on thereceptor of Fig. 11, and the positive made trom the r v receptor of Fig. 12. In this instance also, the ex-sposures arefshort, so that the chemical effect bears approximately a straight line relation to the amount of light received. The resulting composite image on the third plate is a positive at each point of which the effect of vlight on the receptor of Fig. 12 is subtracted from the elfect of the same light from the same part of the subject on they receptor shown in Fig. 11.

The result of subtracting the effect of light on the receptor of Fig. 12 from the effect `of light on the receptor of Fig. 11 is indicated by a full line in Fig. 13. The separation image made in the manner described is, therefore, identical with the separation image which would be made by a theoretical receptor having the spectral sensitivity indicated by the full line in Fig, 13. It will be observed that, the curve showing the effect of the subtraction in Fig. 13 closely approxproximations of the practical receptors shown in Figs. 11 and 12 to the positive and negative parts i of the theoretical curve also shown in those iigures. This indicates another advantage of my method, since, as indicated in Figs. 11, 12 and 13, the departures of the practical receptors from the theoretical sensitivity required can be made to balance off so that the subtraction not only gives effect to the negative part of the theoretical curve but also eliminates` a large part of the departure ofthe sensitivities of the practical receptors from the vsensitivity theoretically required. a

The next step is to prepare a green color-separation image in which light in the spectral regions between wave lengths of 400 and 470 and between 610 and '100 (the regions in which the spectral sensitivity curve shown-in Fig. 9 is negative) has produced an eiect opposite from that of light in theregion between wave lengths of 470 and 61() (the region in which the sensitivity curve shown in Fig. 9 is positive). This is accomplished-in the manner already described, by subtracting the effect of light on a practical receptor whose spectral sensitivity is shown in Fig. 15 from the eiect of light on a practical receptor Whose spectral sensitivity is shown in-Fig. 14. The result of the subtraction, indicated in Fig. 16, closely approximates the required spectral sensitivity curve shown in Fig. 9. The practical receptors whose spectral sensitivities are shown in Figs. 14 and 15 consist of panchromatic photographic plates such as that already specied used with the lters shown in Figs. 14a and 15a.

The next step is to prepare a blue color-separation image in which light inthe spectral region between wave lengths of 550 and 620 (the region in which the sensitivity curve shown in Fig. 10 is negative) has had an eiect opposite from light in the spectral region between wave lengths of 400 and 510 (the region in which the sensitivity curve shown in Fig. 10 is positive). This is accomplished in the manner already described, by

ceptor whose spectral sensitivity is shown in Fig. 17. The result of the subtraction, indicated in Fig, 19, closely approximates the required spectral sensitivity curve shown in Fig. 9. The practical receptors whose spectral sensitivities are shown in Figs. 17 and 18 consist of panchromatic photographic plates such as that already specified used with the filters shown in Figs. 17a and The next step consists in using the red, green and blue color-separation images prepared in the manner above described to control the three primaries in making the reproduction. In the three-lantern method of color reproduction, this is accomplished by placing the red color-separation image upon which light from the subject has had the effect indicated in Fig. 13 in the lantern emitting the red light whose spectral energy distribution is shown in Fig. 5, placing the green color-separation image on which light from the subject has had the effect indicated in Fig. 16 in the lantern emitting the green light whose spectral energy distribution is indicated in Fig. 6, and placing the blue color-separation image upon which light from the subject has had the effect indicated in Fig. 19 in the lantern emitting the blue light whose spectral energy distribution is shown in Fig. 7. It is to be noted that, in the specific example given, each of the three colorseparation images is a photographic positive image adapted for use in the projector.

My invention is, of course, not limited to the simple three-lantern form of color reproduction. It may be applied to the color reproduction methods used in the graphic arts in`cluding four-color methods which merely use one black print in addition to three colored prints. In order that the invention may produce accurate reproduction, the spectral sensitivities required for the theoretical receptors for the separation images must be reproduction systems, as the subtraction can be accomplished at the same time that a half-tone image is made so that the color-separation image is made as a half-tone image adapted for use in engraving a half-tone plate and, in this way, controlling the amount of the primary of the system represented by the colored ink applied to this plate.

In accordance with my invention, the subtraction may be carried out by the scanning appara tus illustrated in Fig. 20. This apparatus is like the scanning apparatus for making half-tone reproductions shown in U. S. Patent No. 1,649,309 issued to H. E. Ives on November 15, 1927, ex-

' cept that it contains two picture drums A and B mounted on the same shaft as the lm drum C, instead of only one picture drum. To make a half-tone red color-separation image for use with the primaries shown in Figs. 5, 6 and 7, positives or negatives made with the receptors of Figs. l1 and 12 are placed on the two picture drums A and B, and are scanned by lights from two identical optical systems AI and BI. The scanning lights modifled by the two images fall on photoelectric cells A2 and B2 which, as shown in Fig. 20, are so connected that the electric currents which they control are opposed and the resulting current is passed into an amplier. The output current of the amplifier is used to control a light valve apparatus C which controls the effect of a light C2 upon a film of the film computed from the spectral qualities of the primaries of the reproduction system. In casesl where the color of the primaries is not readily apparent as in most systems used in the graphic arts, the spectral qualities of the primaries to be used in the calculation may be determined by the method described in my co-pending application, Serial No` 99,416, led September 4, 1936.

The invention is not limited to the particular photographic method of subtracting the eiect of light on one receptor froml the effect of light on another receptor which has been described in the specific example. A still simpler photographic method of subtraction, which may advantageously be used in forming positive colorsep aration images for usel in the projection system, is the following: An exposure of the subject is made on the receptor of Fig. 11 and is developed as a negative. An exposure of the subject is `then made on the receptor of Fig.' 12. Before the plate which forms part of the receptor of Fig. 12 is developed, this plate is exposed to white light through the developed negative plate of the receptor of Fig. 11. All three of the exposures are short for the reasons above explained. The plate of the receptor of Fig. 12 is then developed and will be found to be a positive image identical with that made by the three-plate method before described. Other photographic methods of subtraction may be adopted.

My invention is not limited to photographic subtraction, as the subtraction may be accomplished in other` ways. A feature of my invention consists in making the subtraction electri cally. This is of particular advantage in graphic drum C. The image produced on the film on the drum C is therefore, in effect, the result of subtracting the image on the drum B from the image on the drum A.

An alternative, but less desirable, method of effecting the subtraction consists in using a negative from one of the receptors on the drum A and a positive from the other receptor on the drum B and connecting the two photoelectric cells as shown in Fig. 21 in such a way that the currents controlled by them are added.

In the specific examples, I have described the carrying out of my method by means of known photographic materials and filters and electrical apparatus of a known type. It will readily be understood by those skilled in the art that receptors containing photographic plates and filters especially prepared for the practice of my invention may be made on the basis of the sensitivity characteristics desired, and that, by such special means, the' carrying out of the invention may be to some extent simplified. The invention is, of course,- not limited to any particular means for carrying out the method.

What I claim is:

1. The method of making a color-separation image for use in color reproduction, which comprises passing light emitted by a colored subject to photographic receptors having different spectral sensitivity characteristics to make photographic images of the subject whose point-topoint variation records the point-to-point variation in intensity of two different spectral components of light from the subject, modifying an electric current modulated by the variation in tone frompoint-to-point of one of said images by an electric current modulated by the variation in tone from corresponding point to corresponding point of the other of said images, and recording the modified electric current.

2. The method which comprises making different color-separation photographs from a subject, directing light through the different color-separation photographs to separate photo-electric receptors, modifying an electric current whose variations constitute the response of the iirst receptor to the variation in tone of the iirst color-separation photograph by an electric current whose variations constitute the response of the second receptor to the variations in tone of the second color-separation photograph. and making a record oi the modified current.

3. The method which comprises taking two different color-separation photographs of the same subject, modulating an electric current in accordance with the point-to-point variation in tone of one of said photographs, simultaneously modulating another electric current in accord ance with the variation from the same point to rents in accordance with the variations in tone` of said diierent color-separation photographs, and utilizing all said electric currents to .control a single recording device. i

ARTHUR C. HARDY.

Images