Colorimetric and spectral matching

Prague, Czech republic

June 29, 2017

Marc Mahy
Agfa Graphics
Overview
• Modeling color
• Color matching
• Process control
• Profile based color transforms
— Visual effects
— Light interactions
• Color matching
— Colorimetric matching
— Spectral matching
— Metameric colors
Modeling color
• Object has the color of the “light” leaving its surface
— Light source
— Object
— Human observer , )
 Modeling color
• Light source
— Electro Magnetic Radiation (EMR)
– Focus on wavelength range from 300 till 800 nm
— Different standard illuminants
– Illuminant E (equi-energy)
– Illuminant A, D50 , D65 , F11
Spectral Power Distribution
Modeling color
• Object colors: Classes object types
— Opaque objects
– Diffuse reflection: Lambertian reflector
– Specular reflection: mirror
– Most objects: diffuse and specular reflection
— Transparent objects
– Absorption, no scattering: plexi, glass, …
— Translucent objects
– Absorption and scattering: backlit
— Special effects
– Fluorescence: substrates
– Metallic surfaces: (in plane) BRDF
Modeling color
• Object colors: Characterizing object types
— Opaque objects
– Measurement geometry: 45°:0° or 0°:45°
– Colorimetric or reflectance spectra
— Transparent objects
– Measurement geometry: d:0° or 0°:d
– Colorimetric or transmission spectra
— Translucent objects
– In reflection or transmission mode
– For reflection mode: White backing, black backing or self backing
— Special effects
– Fluorescent substrates: colorimetric data or bi-spectral reflectance
– Metallic surfaces: BRDF based on colorimetric or reflectance data
 Modeling color
• Self luminous colors
— Displays
– Additive color mixing based on Red,
Green and Blue phosphors

• Characterizing self luminous colors

— Displays: colorimetric data or radiant
emission spectra
Modeling color
• Human observer
— Color vision: three types of cones
— Sensitive to EMR from 360 till 830 nm
• Tristimulus color spaces
— Based on color matching experiments
— Grassman’s laws (additivity)
 Color Matching Functions (CMF’s)
• Tristimulus values opaque objects
Color matching
• Colorimetric matching
— Two colors (test and reference sample)
– under the same viewing conditions
– under different viewing conditions
— Condition colorimetric match opaque colors
– Same viewing conditions: XYZ1 = XYZ2
– Different viewing conditions: Jab1 = Jab2

• Examples
— Contract proofing for a given illuminant
— Conversion from display (D65) to graphic arts (D50)

• Issues
— No match for changes in illumination
— No match for observers with different CMF’s
Color matching
• Spectral matching
— Two colors (test and reference sample)
– evaluated simultaneously for different illuminants
— Too strict condition spectral match opaque colors:

• Examples
— Conventional proofing systems
– Spectral color matching between proof and press
– Accurately simulating screening effects
• Issues
— Not possible to get a spectral match between display and hard copy
— Only supported for object colors
 Process control
• Normal distribution
— Continuous probability distribution 2

(μ=0, σ=1)
1.5
,
2 1

with x: variable area = 1

0.5
μ : average
σ : standard deviation 0
-5 -4 -3 -2 -1 0 1 2 3 4 5

— Properties 2
normal distribution φ , n=1
1.5
=> ̅ normal distribution φ , / n=5
=> Reducing uncertainty &
1 n=10
increasing accuracy 0.5 n=20
0
-5 -4 -3 -2 -1 0 1 2 3 4 5
Process control
• Normal distribution (cont’d)
— Confidence intervals
– +/- 1σ: 68.3 % 3170 errors per 10,000 measurements
– +/- 2σ: 95.5 % 460 errors per 10,000 measurements
– +/- 3σ: 99.7 % 27 errors per 10,000 measurements
 Profile connection spaces
Profile based transforms - XYZ, Lab
Exchange color data: Viewing conditions
• ICC.1 approach Colorimetric only - D50, 2° observer
- CAT

Source PCS Destination

• ICC.2 approach (iccMAX) Profile connection spaces

- XYZ, Lab or spectral
Exchange color data:
Viewing conditions (PCC):
Both Colorimetric and spectral - Illuminant, observer and
lighting surround

Source PCS Destination

Profile based transforms: Visual effects
• Visual effects
• Appearance based color matching
• Needed if viewing conditions source and destination are different
• For ICC.1: CAT’s are used
• For ICC.2: Support provided by modified CIECAM02
• Local effects such as lightness and color induction, crispening, ...
• Not simulated as they are present in the image
Profile based transforms: Light interactions
• Print substrates
— Large variety of substrates used in ink jet applications
– Plastics, backlit, plexi, glass, wood, …
– Special light objects interactions
 Not always easily to measure
Profile based transforms: Light interactions
• Measurement instruments
— Both available for reflectance and transmittance
— Support M0, M1 and M2 measurement conditions for reflectance
– M1 either for D50 or a custom illuminant
— Typical characteristics
– Different apertures to measure local non-uniformities
– Spectral data provided
– Black and white backing available
Profile based transforms: Light interactions
• Effect thickness substrate
— Plexiglass
— Reflectance measurements are not accurate or meaningfull
Detector
Light source
Expected
reflection

reflection
Actual
Air

Plexi

Lambertian reflector
Profile based transforms: Light interactions
• Transparent and translucent substrates
— Interaction light in substrate

ink
substrate

— Measurement dot gain based on ISO-12233 (Slanted edge method)

edge with Edge Spread Line Spread

a small angle Function MTF
Function
Fourier
Derivative Transform
Pre processing
 Profile based transforms: Light interactions
• Transparent and translucent substrates
— Based on edge spread function
– Patch size on translucent substrates 20 mm
– Patch size on transparent substrates 5 mm

Transparent substrate Translucent substrate

60000 60000
2.5 mm 8 mm
50000 50000
Edge spread function

Edge spread function

40000 40000

30000 30000

20000 20000

10000 10000

0 0
-10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10
D (mm) D (mm)
 Profile based transforms: Light interactions
• Transparent and translucent substrates
— Effect illumination size
– Effect white ink on transparent substrate
– Effect cyan ink on translucent substrate

Cyan on translucent substrate White on transparent substrate

1 1
0.9 Only boxaperture
Large 0.9
0.8 Medium
Mediumaperture
aperture 0.8
0.7 0.7
Transmittance

Transmittance
Smallaperture
Small aperture
0.6 0.6
0.5 0.5
0.4 0.4
unlimited size of
Large aperture
0.3 0.3
illumination
0.2 0.2
Small
small aperture
size
0.1 0.1
0 0
380 430 480 530 580 630 680 730 380 430 480 530 580 630 680 730
Wavelength (nm) wavelength
 Profile based transforms: Light interactions
• Transparent substrates
— Tunneling effect (inter-reflections in substrate)
Incident light

Interreflections
 Profile based transforms: Light interactions
• Textile
— Non-uniform surface
— Effect directional geometry

Proofing substrate Textile substrate

5 5
4 4
3 3
Z-value

2 2
1 1
0 0
-1 -1
-2 -2
-3 -3
-4 -4
-5 -5
0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360

Angle Angle
Colorimetric matching
• Conventional color management
— Supported both by ICC.1 and ICC.2 (iccMAX)
• Use cases
— Perceptual match=> typically viewing in isolation
— Proofing => side by side evaluation
• Object type combinations for ICC.2 (iccMAX)
Colorimetric Opaque Trans- Trans- Display Fluorescent
match parent lucent substrate
Opaque
Transparent
Translucent
Display
Fluorescent substr.
Colorimetric matching
• Handling mismatch in viewing conditions

Offset Digital print Digital

Offset profile PCS profile Print

Viewing conditions Viewing conditions

- D50 simulator - D65
- 2° observer - 2° observer

— ICC.1: CAT’s
– Perceptual: CAT’s used
– Proofing: undo CAT’s, recalculation destination profile is needed
— ICC.2: Profile Connection Conditions (PCC’s)
– Perceptual: mechanism in place to handle different PCC’s
– Proofing: inverse table cannot be reused
 Colorimetric matching
• Handling mismatch in viewing conditions (con’d)
D50
— Effect illuminants on gamut D65
— Spectral measurement file for Fogra51 D93
L* Illuminant A
a*

b*

D50-D65 L*=50
L*=50 Hue = 50°
C*
Colorimetric matching
• Closed loop

Offset Digital print

Offset profile
PCS profile
Digital
print

— Iteratively improving device model (forward transform)

— Closed loop approach limited by
– Accuracy measurements
– Stability printer
— Results in
– Increased profile accuracy w.r.t. behavior printer
Colorimetric matching
• Accurate linking
— Reduction interpolation errors for in-gamut and out-of-gamut colors
L* L*
DeltaE due to DeltaE = 0
interpolation
errors

Explicit
gamut
mapping
Desaturated mapping
out-gamut-colors
C* C*
Table based linking Accurate linking
Colorimetric matching
• Stability neutrals
— Proofing: Absolute colorimetric intent
— Neutral check target
— RGB target consisting of 11 different neutrals repeated per line (x26)
 Colorimetric matching: Stability neutrals
Conventional link Closed loop + Accurate linking

1.5 1.5
1 1
0.5 0.5
0 0
-0.5 -0.5
-1 -1
-1.5 -1.5
20 40 60 80 100 20 40 60 80 100

1.5 1.5
1 1
0.5 0.5
0 0
-0.5 -0.5
-1 -1
-1.5 -1.5
20 40 60 80 100 20 40 60 80 100

1.5 1.5
1 1
0.5 0.5
0 0
-0.5 -0.5
-1 -1
-1.5 -1.5
20 40 60 80 100 20 40 60 80 100
Spectral matching
• Not explicitly supported by conventional color management
— Dedicated support provided by ICC.2 (iccMAX) only
• Use cases
— Similar appearance prints from different CMYK presses
— Packaging: spot color processing
• Object type combinations for ICC.2 (iccMAX)
Spectral match Opaque Trans- Trans- Display Fluorescent
parent lucent substrate
Opaque
Transparent
Translucent
Display Identical
Fluorescent substr. Identical

— Focus on matching between opaque objects, CMYK printed output

 Spectral matching
• Spectral match preserving transformations
— Repurposing transformation
– From Fogra51 to CGATS TR 006 Original Repurpose ICC convert
– Preserves GCR from source
 Acceptable spectral match

K-only
— Identity transform
– Repurposing transform between
identical profiles

CMY
 Preserves spectral match
CMYK
Spectral matching
• Spot color mixing
— Dynamic rendering of spot colors
– Spectra step wedge on white and black background (CxF/X-4)
– Output profile defines substrate and viewing conditions
— Rendering characteristics
– Realistic rendering of ink combinations based on spectral mixing
– Continuity between different subsets of ink values
— Spectra substrate and process inks: spectral reflectance estimation
100 100
Spot Yellow

Spot Yellow
Spectral
75 75
mixing
50 50

25 25

0 0

0 25 50 75 100 0 25 50 75 100
Process cyan Process cyan
Spectral matching
• Spot color mixing (cont’d)
— Spectral reflectance estimation
– Given discrete version of the tristimulus values XYZ

- with reflectance spectrum object

spectral power distribution illuminant
color matching functions
N dimension spectral space
– Invert tristimulus values XYZ and solve for with constraints
» Reflectance: 0  Ri  1 i :1  N
» Smoothness criteria
» Typical aim curves
 Color matching: Metameric colors
• Illuminant discriminating metameric colors
— Idea:

CMYK Lab for D50

Lab for Illuminant A

— Calculation illuminant discrimination metameric color pair:

– Look for a set of CMYK values that
- are the same for a first illuminant
- have a maximum E*ab for a second illuminant
Thank you for your attention  Any questions?