1、CIE 80 89 = 9006345 0003784 858 ISBN 3 900 734 17 8 C O M M ISS I O N I NTER NAT I O NALE D E L C LAI RAG E INTERNATI ON AL COM M ISSION ON I LLUM INATION INTERNATIONALE BELEUCHTU NGSKOM M ISSION SPECIAL METAMERISM INDEX: CHANGE IN OBSERVER Pub. No CIE 80 1 st Edition 1989 UDC: 159.937 Descriptors:
2、Perception of colours 535.66 Colours of objects 61 2.843.31 Colour vision - _- CIE 80 89 9006345 0003785 794 II This Technical Report has been prepared by CIE Technical Committee 1-07 of Division 1 Vision es ist zur Verwendung durch CIE-Mitglieder und durch andere Interessierte bestimmt. Es sollte j
3、edoch beachtet werden, da das Dokument eine Empfehlung und keine Vorschrift ist. Die neuesten CIE- Tagungsberichte oder das neueste CIE-Journal sollten im Hinblick auf mgliche sptere Anderungen zu Rate gezogen werden. Any mention of organisations or products does not imply endorsement by the CIE. Wh
4、ilst every care has been taken in the compilation of any lists, up to the time of going to press, these may not be comprehensive. Toute mention dorganisme ou de produit nimplique pas une prfrence de la CIE. Malgr le soin apport la compilation de tous les documents jusqu la mise sous presse, ce trava
5、il ne saurait tre exhaustif. Die Erwhnung von Organisationen oder Erzeugnissen bedeutet keine Billigung durch die CIE. Obgleich groe Sorgfalt bei der Erstellung von Verzeichnissen bis zum Zeitpunkt der Drucklegung angewendet wurde, ist es mglich, da diese nicht vollstndig sind. Q CIE1989 CIE BO 89 I
6、 7006345 0003786 620 m III Foreword The following members of T. C. 1-07 (Observer Metamerism) took part in the prepa- ration of this Technical Report. The Committee comes under CIE Division 1 (Vision and Color). Member E. Allen H. Hemmendi nger K. McLaren Y. Nayatani N. Ohta (Chairman) J. J. Opstelt
7、en D. Strocka K. Takahama Consul tant R. W. G. Hunt P. K. Kaiser C. Wys zech i U. S. A. U. S. A. Great Britain Japan Japan The Netherlands Germany, Federal Republic Japan Great Britain Canada Canada CIE 80 89 m 9006345 0003787 5b7 m IV Table of Contents Preface 1. Introduction 2. Procedure 2.1 Avera
8、ge Color Differences and Metamerism Index M.br 2.2 Range of Color Mismatches (1) Object Colors (2) Illuminant Colors 2.3 Age Effect of Color Mismatches Append ix References Table Page 1 8 9 13 14 Special Metamerism Index: Change in Observer Summary This report provides a method for evaluating the de
9、gree of color mismatch for a metameric color pair (object color or illuminant color) when an actual observer with normal color vision is substituted for the standard colorimetric observer. The method evaluates both the average values of, the color mi smatch. and the range of, Indice Special de Metam
10、erisme: Changement de lobservateur Rsum Ce rapport fournit une mthode pour valurr 1 cart de couleur dune paire de coul eurs mtamres (coul eur d un ob jet ou d un i 11 urni nant 1 , 1 orsqu un obser- vateur rel, ayant une vision normale des couleurs, est substitu lobserva- teur de rfrence colorimtriq
11、ue. La mthode permet dvaluer, la valeur moyenne de lcart de cou- leur et ltendue de celui-ci. la fois, Spezieller Metamerie Index: Unterschiede im Beobachter Zusammenfassung Dieser Bericht behandelt ein Verfahren zur Bestimmung des Maes der Farb- fehl abstimmung bei einem bedingt-gleichen Farbenpaar
12、 (Krperfarbe oder Farbe einer Lichtquelle) , wenn ein tatschlicher Beobachter mit normalem Farbseh- vermgen statt des farbrnetechnischen Normalbeobachters eingesetzt wird. Mit dem beschriebenen Verfahren lt sich sowohl der Durchschnittswert als auch der Gesamtumfang der Farbanpassung bestimmen. CIE
13、80 89 I 9006345 0003789 33T I 1 Preface Two specimens having identical tristimulus values, for a given illuminant and reference observer, are metameric if their spectral radiance factors differ from each other within the visible spectral range. This metameric color match is broken down either by a c
14、hange of the reference illuminant or the observer. The degree of color mismatch caused by substituting a test illurninant of differ- ent relative spectral composi tion for the reference illuminant is classified as the Special Metamerism Index : Change in Illuminant (illuminant metamerism). . The met
15、hod of evaluating illuminant metamerism has already .been formulated: the purpose of this report is to formulate a method of evaluating the degree of color mismatch for metameric color pairs (object colors or illuminant colors) caused by substituting a test observer with normal color vision for the
16、reference observer (CIE Standard Colorimetric Observer). This is classified as the Special Metam- erism Index : Change in Observer (observer metamerism). Color mismatch caused by observer metamerism is inevitable for real observers with normal color vision. The evaluation of observer metamerism thus
17、 gives a sound basis for establishing color difference tolerances for color matching using various metameric color pairs. It also gives an indication of the degree of i 1 luminant metamerism which can be considered al lowable in our everyday 1 ives. The methods of evaluating observer metamerism are
18、generally grouped into two categories. For a given set of metamers, they are : (1) using actual observers with normal color vision to calculate the average of, and the range of, the color mismatches between metamers : (2) selecting a set of color matching functions representing the individual variat
19、ions among color normal observers, and using them to calculate the average of, and the range of, the color mismatches. Method 1 is straightforward, and work using this procedure has already been reported using the Davidson-Hemmendinger(D-H) color rule. l. - However, for actual metameric color pairs
20、other than the D-H color rule, method 1 requires subjective evaluation of color mismatches which is difficult and unstable. CIE 80 89 I 900bL45 0003790 051 2 Method 2 is therefore preferred. Once the suitable color matching functions are formulated, the calculation of color mismatches is as simple a
21、s the calculation of tristimulus values. In the early stages of formulating method 2, the use of 20 sets of color matching functions for normal color vision was recommended. . 5-6 However, it is rather cumbersome to use 20 sets of coIor matching functions in evaluating observer me- tamerism and effo
22、rts were made to reduce the number to only one and propose a Standard Deviate Observer “ who has color matching functions which differ from those of the standard observer by amounts equal to the standard deviations among the 20 sets of color matching functions. : these will be referred to as the 20
23、observers hereafter. The method employed was a statistical analysis of the var- iances and covariances of the color matching functions of the 20 observers. However. it was found that the Standard Deviate Observer did not give a satis- factory prediction of observer metamerism. An aiternative suggest
24、ion selected the best combination of two observers from among the 20 observers. I. These two observers give highly correlated results with those obtained by the 20 observers. A further proposal was to use a single Standard Deviate Observer by analyzing the same 20 observers but by use of a different
25、 statistical method known as a singular value decomposi t ion. This Standard Deviate Observer gives a good prediction of the average value of color differences among the 20 observers. A method of evaluating the range of color mismatches was also proposed. . la attempt to confirm these results a diff
26、erent approach using a nonlinear opti- mization of the 20 observers and actual metameric samples was performed. . I The results turned out to be much the same and the Standard Deviate Observer obtained well represented the original 20 observers. In an It was later shown that the age distribution of
27、the 20 observers was rather wide. ranging from about 20 to 60 years old and that apparent differences between the results obtained by method and 2 are simply because of the difference in scaling of color differences. I The same analysis has been carried out for the color matching functions of 53 CIE
28、 80 89 E 900bL45 0003793 T8 = 3 observers measured by Stiles et al. showing that the correlation between the ob- server metamerism indices by the 20 and by the 53 observers was satisfactory. However, the age distribution of the 53 observers is not uniform but significantly high in the twenties, and
29、the contribution factors of the decomposed components from the 53 observers were lower than those derived from the 20 observers. For these reasons. the color match basis of the Present method. The research work reported dea imetric Observer. It has also ng functions of the 20 observers are used for
30、the s with the CIE 1964 Supplementary Standard Color- been shown that observer variations in color matching functions are mainly caused by the variation of eye-lens pigmentation due to the change of observer age. I. Therefore, the present method is also applicable for the CIE 1931 Standard Colorimet
31、ric Observer. 1. Introduction A method of evaluating the average of, and the range of, color mismatches for metameric color pairs when test observers are substituted for a reference observ- er is presented. The reference observer is either the CIE 1931 Standard Color- imetric Observer or the CIE 196
32、4 Supplementary Standard Colorimetric Observer. The test observers are assumed to be a number of actual observers with normal color vision, and are represented by four deviation functions characterizing the variations of color matching functions of color normal observers. When metameric colors are e
33、valuated by the test observers, the resultant tristimulus values do not match each other. Instead they spread within a certain range in a three-dimensional color space. The range is characterized in a chromaticity diagram by a statistical confidence ellipse containing 95% of the spread. The followin
34、g procedure gives an average color difference and the shape of the ellipse. The two-dimensional ell ipse does not contain information about the Psychometric lightness. The ellipse is therefore not applicable to the intercornparison of the degree of observer aietamerlsrn between two different metamer
35、ic pairs. Rather it shows the chromaticity gamut in color matches of color normal observers for a me- tameric sample pair, and also evaluates the adequacy of the observer variations CIE AO 89 900bL45 0003792 924 4 of chromaticities in the color matches using a visual colorimeter for a number of colo
36、r normal observers. Four sets of deviation functions AFiA ), AYi(A), ATi(A),i= 1,2,3, and 4. are used in the method. These functions are defined and tabulated from 380nm to 780nm in 5nm steps. The first set of deviation functions in TabIe.l is the dif- ferences in color matching functions between th
37、e CIE Standard Colorimetric Ob- servers and the Standard Deviate Observer, and is used to evaluate the average value of color differences, or the metamerism index. The additional three sets of deviation functions are. together with the first one, used to define the shape of el 1 ipse. In the followi
38、ng procedure, the subscript shows the kind of deviation functions and the superscript the number of colors unless otherwise specified. This procedure is applied both for the CIE 1931 Standard Colorimetric Observer and for the CIE 1964 Supplementary Standard Colorimetric Observer. Though the metameri
39、sm index is Sig- nificant only for object colors. the variation of chromaticity coordinates, the confidence ellipse, is calculated both for object colors and for illuminant colors. 2. Procedure For a pair of object colors which is metameric with respect to a reference observer and an irradiating ill
40、uminant, their tristimulus values Xrct. Yret. 2,ef. j=l J J J and 2, are derived by the usual CIE recommendationl- I as (j = 1,2 (2. 1) where LI (2) is the spectral reflectance of the j-th object color of the metameric pair, S(A) is the spectral power distribution of the illuminant irradiating the o
41、bjects, and Y(A .Y(A ),T( A ) are the color matching functions of either the CIE 1931 Standard Colorimetric Observer or the CIE 1964 Supplementary Standard Colorimetric Observer. The distribution S(A 1 is normalized as C S(L) Y (A) AR=100. A (2. 2) CIE 80 87 ai 7006345 0003793 60 The o calcu 5 (2.3)
42、 The following relation holds by definition. I 2 Xrer = Xrer E Xrc t I 2 Yrsr = Yrei Yrei , Z rer = Z rcr Z rei . I 2 The four sets of deviation functions AFI(A 1. AYI(A . A?I(A 1. i=l.2.3,and 4. of Table 1 are used in the method. The first set is used to define the color matching functions Xdev( A
43、).Ydcv(A 1, z dav(1 ) of the Standard Deviate Ob- server as - - - xdev(A) =X(A +AYI(A, (2.4) - Y dcv(A ) =y(A + AYI(A )* zdev(A) =Z(A) +AZI(A). - - her three deviation functions are, together with th first one. used to ate the range of color mismatches. 2-1 Average Color Difference and Metamerism In
44、dex Mob. The tristimulus values Xds Ydev. Zder and Xdrv. Ydar. Zdev of the pair of meta- meric object colors are defined for the Standard Deviate Observer as I 1 I 2 2 2 The color difference AEobl of a pair of metameric object colors, when evaluated by the Standard Deviate Observer, is calculated by
45、 converting Xdev. Ydev. Zev and Xdcv.Ydev. Zde. into a uniform color space. A use of either the CIE 1976 (L u v* or (L a b ) space is recommended. I* In calculating L U* Y* or L a b values, the tristimulus values of the illuminant are obtained by putting p (A )=I in Es. (2.51. The metamerism index U
46、 Ob8 for a change in observer is I I l 2 2 2 defined as * Mobs ( u v or a bo) =AEob (2. 6) The uniform color space used to evaluate the color difference AE is included in the parenthesis. - CIE BO 89 900bL45 0003794 7T7 W 6 Note 1 When the samples are not exactly metameric, that is Xrei f Xrer, Yrer
47、 f Yrei. Zrei f Zref, then Xrei. Yrer.Zref are defined for convenience as 1 2 I 2 I 2 I I I Xre= Xref , Yrei= Yrei, Zrei= Zrc * (2.7) Then the tristimulus values obtained by Ea(2.5) are adjusted by the multiplica- tive method as follows. I 1 1 2 2 2 The corrected tristimulus values ( Xdev . (Ydev 1
48、, (Zdcr 1 are used in place of Xdev. Ydev. Zdev respectively in the above procedure. 2 2 2 Note 2 Each color normal observer shows a color difference AEl for a sample pair meta- meric with respect to a reference observer and an irradiating illuminant. About 95% of AEis for color normal observers are
49、 usually found within 2 AEo“ (U v or a b*). 8 2-2 Range of Co lor Mi sma t ches A set of tristimulus value deviations A2Xl. A2Yi. A2Zl. (i=1,2,3,4) for a pair of metameric object colors are defined by a procedure similar to the usual CIE recommendat ion. I as where Ap(A)=p2(A)- (1) and DI(A),D(A) are the spectral reflectance$ of the metameric object colors and S (A 1 is the spectral power distribution of the illuminant irradiating the objects . Then the two matrices D and V are d
