Mean Observer Metamerism and the Selection ofDisplay Primaries

Mark D. Fairchild & David R. WybleRIT Munsell Color Science Laboratory

A Conclusion...

“There is another reason why it may be desirable to use desaturated primaries in a television receiver. It has been found in direct colorimetry that observer differences can be minimized by making the color triangle of the primaries no larger than is necessary to include the variation of chromaticities to be measured.”

Wintringham, 1954

Observer Metamerism

A metameric match for one observer is likely to mismatch for another.

Terms

Mean Observer Metamerism

≠

Mean-Observer Metamerism

CIE TC1-36

• CIE 170-1:2006

• Cone Fundamentals (CMFs)

• Function of Field Size

• Function of Age

• Mean Functions

CIE 2006 Model

�

l (λ) = α i,l (λ) ⋅10−Dτ ,max,macula ⋅Dmacula ,relative (λ )−Dτ ,ocul (λ )[ ]

m (λ) = α i,m (λ) ⋅10−Dτ ,max,macula ⋅Dmacula ,relative (λ )−Dτ ,ocul (λ )[ ]

s (λ) = α i,s(λ) ⋅10−Dτ ,max,macula ⋅Dmacula ,relative (λ )−Dτ ,ocul (λ )[ ]

Ocular Media Density

f(age)

Cone Absorptivity

Spectra

f(field size)

Cone Fundamentals

Macular Density

f(field size)

Notation

• CIE2006(field size, age)

• e.g. CIE2006(2,32)

• 2-Degree Field, 32 Years Old

Examples: Field SizeL-, M-, & S-Cone Fundamentals (2- & 10-Deg. @ Age 32)

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

390 440 490 540 590 640 690 740

Wavelength (nm)

Rela

tive

Sen

siti

vity

l-bar (2)m-bar (2)s-bar (2)l-bar (10)m-bar (10)s-bar (10)

Examples: AgeL-, M-, & S-Cone Fundamentals (10-Deg. @ Ages 20 & 80)

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

390 440 490 540 590 640 690 740

Wavelength (nm)

Rela

tive

Sen

siti

vity

l-bar (20)m-bar (20)s-bar (20)l-bar (80)m-bar (80)s-bar (80)

CIE 2006 Code

Excel Spreadsheet & MATLAB Code

mcsl.rit.edu/online/cie.php

ComputationalExperiment

Select Original Colors24-Patch ColorChecker

CIE Illuminant D65

Use Matching Spectra to Compute CIELAB !E*ab

for CIE 1931 CMFs

Compute Display Matches

Compare Effectsof Primaries, Field Size,

and Age

Repeat forFS = 1°,2°,4°,7°,10°

Age = 20,32,40,60,80

Define CMFsCIE2006(FS,Age)

Define Simulated Display Primaries

Compute StatisticsAcross 24 Patches

Repeat for Broad- and Narrow-Band Primaries

Computational Match

CIE2006(fs,age)

fs = 1,2,4,7,10°age = 20,32,40,60,80y

Original Colors(D65, ColorChecker)

Broad-Band Primaries

Narrow-Band Primaries

Original Colors(D65, ColorChecker)

Evaluation of Matches

CIE 1931Std. Obs.

Broad-Band Primaries

Narrow-Band Primaries

CIELAB Color Differences

Display Primaries• 450, 540, 610nm Peak Wavelengths

0.1

1

10

100

390 440 490 540 590 640 690 740

Wavelength (nm)

Rela

tive

Sp

ect

ral P

ow

er

Broad-Band Primaries Narrow-Band Primaries

Matches to D65 White

Chromaticity Gamuts

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 0.2 0.4 0.6 0.8

x

y

Results

• Color Differences (CIELAB, 1931)

• Chart (D65) vs. Simulated Display Match

• Averaged Over 24 Patches

ResultsAge Field Size

(°) Mean (BB) SD (BB) Mean (NB) SD (NB)

2020202020

124710

1.211.022.183.243.89

0.720.751.251.922.37

1.911.574.007.6110.06

0.790.792.604.575.70

3232323232

124710

1.380.741.792.843.48

0.700.591.001.652.08

1.631.333.927.479.88

0.730.632.364.225.41

4040404040

124710

1.560.641.542.563.20

0.790.510.841.471.89

1.531.243.887.369.72

0.790.562.173.985.19

6060606060

124710

2.100.870.901.872.48

1.180.500.471.011.40

1.521.243.736.989.19

1.180.531.733.394.51

8080808080

124710

4.443.121.830.970.72

2.922.161.360.820.60

3.262.292.845.046.78

2.671.761.151.962.65

CIE-31CIE-64

210

02.53

01.68

04.30

02.39

32 Year-Old Observers

0.0

2.0

4.0

6.0

8.0

10.0

12.0

0 2 4 6 8 10

Field Size (Deg.) for 32-Years Age

CIE

LAB

Delt

a-E

*

Broad-Band Primaries Narrow-Band Primaries

10-Deg. Observers

0.0

2.0

4.0

6.0

8.0

10.0

12.0

10 20 30 40 50 60 70 80 90

Age (Years) for 10-Deg. Field Size

CIE

LAB

Delt

a-E

*

Broad-Band Primaries Narrow-Band Primaries

White Patch

-10

-8

-6

-4

-2

0

2

4

6

8

10

-10 -8 -6 -4 -2 0 2 4 6 8 10

a*

b*

20,1 20,2 20,4 20,7 20,1032,1 32,2 32,4 32,7 32,1040,1 40,2 40,4 40,7 40,1060,1 60,2 60,4 60,7 60,1080,1 80,2 80,4 80,7 80,101964

-8

-6

-4

-2

0

2

4

6

8

10

12

-16 -14 -12 -10 -8 -6 -4 -2 0 2 4

a*

b*

20,1 20,2 20,4 20,7 20,1032,1 32,2 32,4 32,7 32,1040,1 40,2 40,4 40,7 40,1060,1 60,2 60,4 60,7 60,1080,1 80,2 80,4 80,7 80,101964

Broad Band Narrow Band

Visualization

• Take Matching RGBs and Display on a Single sRGB Display

• Center: Broad-Band “Original”

• Surround: 4 Computed Narrow-Band Matches

• Illustrates Changes in Device RGB Needed to Compensate for Observer Variability

Visualization Key

Full ColorChecker

Pictures

Pictures

Conclusions

• Narrow-Band Primaries

• More Observer Variability

• Mean Observer Metamerism is Significant

• Individual Observer Metamerism is Even Greater

• Consider the CIE 170-1:2006 CMFs

Also ...

High-luminance desaturated primaries

appear more colorful than

dimmer monochromatic primaries...

Thank Youmcsl.rit.edu