color and color models - tu wien · color spaces. color metric spaces (cie xyz, lab) used to...

Einführung in Visual Computing186.822

Color and Color Models

Werner Purgathofer

Color

problem specificationlight and perceptioncolorimetrydevice color systemscolor ordering systemscolor symbolism

Werner Purgathofer 2

Color - Why Do We Care?

Visual Computing is all about the generation and the manipulation of color imagesproper understanding & handling of color is necessary at every step


Color - A Visual Sensation


eye brainlightstimulus

nervesignal

object

electromagnetic rays color sensation

realm of direct observables realm of psychology

What is Light?

“light” = narrow frequency band of electromagnetic spectrumred border: 380 THz ≈ 780 nmviolet border: 780 THz ≈ 380 nm


frequency(Hz)

102 104 106 108 1010 1012 1014 1016 1018 1020

AM ra

dio

FM ra

dio

and

TV

mic

row

aves

infr

ared

ultr

avio

let

X-ra

ys

1016 1014 1012 1010 108 106 104 102 100 10-2

wavelength(nm)

visible

… …

Light - An Electromagnetic Wave

light is electromagnetic energymonochrome light can be described either by frequency f or wavelength λc = λ·f (c = speed of light)

shorter wavelengthequals higherfrequency

red ≈ 700 nmviolet ≈ 400 nm


Et

λ

Light – Spectrumnormally, a ray of light contains many different waves with individual frequencies the associated distribution of wavelength intensities per wave-length is referred to asthe spectrum of a givenray or light source


Dominant Wavelength | Frequency

dominant wavelength | frequency (hue, color)brightness (area under the curve)purity


ED

ED – EW ED ... dominant energy densityEW ... white light energy density

white lightenergy

wave-length

700 nm400 nm

energy

wave-length

greenish light

dominantwavelength

ED

EW

The Human Eye

retina containsrods: b/wcones: color


rods

cones

lens

visual axisoptical axis

fovea

aqueous[Augenkammer]

cornea[Hornhaut]

iris [Regen-bogen-

haut]

macula lutea[gelber Fleck]

nerve

retina[Netzhaut]

vitreous humor[Glaskörper]

optic disc[Papille]

retina

opticalnerve

blindspot

The Human Eye

3 types of cones

differentwavelengthsensitivities:

redgreenblue


fraction ofabsorbed light

1%2%4%8%16%

400 440 480 520 560 600 640 680

λ

Color Blindness

red/green blindnessred & green cones too similar



1%2%4%8%16%

400 440 480 520 560 600 640 680

λ

Color Blindness


blue blindnessno blue cones



1%2%4%8%16%

400 440 480 520 560 600 640 680

λ

Color Blindness


blue blindnessno blue cones

monochromatismall cones missing



1%2%4%8%16%

400 440 480 520 560 600 640 680

λ

Color Blindness Tests


What do you see?



5 = normalnothing = red/green blind

2 = red/green weaknothing = normal



What do you see?



8 = normal3 = red/green weaknothing = red/green blind

8 = red/green blind12 = blue/yellow blind182 = normal

Color Blindness Example


normal vision



normal vision red/green weakness



normal vision red/green weaknessred/green blindness

Color Spaces

Color Metric Spaces (CIE XYZ, L*a*b*)used to measure absolute values and differences – has roots in colorimetry

Device Color Spaces (RGB, CMY, CMYK)used in conjunction with devices

Color Ordering Spaces (HSV, HLS)used to find colors according to some criterion

the distinction between them is somewhat obscured by the prevalence of multi-purpose RGB in computer graphics


What is our Goal?

to be able to quantify color in a meaningful, expressive, consistentand reproducible way

problem: color is a perceived quantity, not a direct, physical observable


Color - A Visual Sensation


eye brainlightstimulus

nervesignal

object

electromagnetic rays color sensation

realm of direct observables realm of psychology

Colorimetry

Colorimetry is the branch of color science concerned with numerically specifying the color of a physically defined visual stimulus in such manner that

stimuli with the same specification look alike (under the same viewing conditions)

stimuli that look alike have the same specificationnumbers used are continuous functions of the physical parameters


Colorimetry Properties

Colorimetry only considers the visual discriminability of physical beams of radiationfor the purposes of Colorimetry a „color“ is an equivalence class of mutually indiscriminable beamscolors in this sense cannot be said to be “red”, “green” or any other “color name”discriminability is decided before the brain - Colorimetry is not psychology


observers had to match (monochromatic)test lights by combining 3 fixed primaries

test box: compare test light with combined lightWerner Purgathofer 27

Color Matching Experiments

0 10 10 1


observers had to match (monochromatic)test lights by combining 3 fixed primaries

R = 700.0 nmG = 546.1 nmB = 435.8 nm

goal: find the unique RGB coordinates for each stimulus

Color Matching Experiments

green test test

R+G

+B

0 10 10 1


Tristimulus Values

the values RQ, GQ and BQof a stimulus Q that fulfill

are called the tristimulus values of Q

in case of a monochromatic stimulus Qλthe values Rλ, Gλ and Bλ are called spectral tristimulus values

green test test

R+G

+B

0 10 10 1Q = RQ·R + GQ·G + BQ·B

Werner Purgathofer

Color Matching Procedure

(1) test field = 700 nm -red with radiance Pref

observer adjusts luminance of R (G=0, B=0)(2) test light wavelength is decreased in constant steps

(radiance Pref stays the same) observer adjusts R, G, B

(3) repeat for entirevisible range

400 450 500 550 600 650 700 nm35030

100

nomatch

possible!?!?

0400 450 500 550 600 650 700 nm350


Color Matching Result !?

observers want to „subtract“ red light from the match side...!?


for some colors observers want to reduce red light to negative values…!?but there is no negative light…!

Color Matching Experiment Problem

0 1

green test test

R+G

+B

0 10 1

?


“Negative” Light in a Color Matching Exp.

if a match using only positive RGB values proved impossible, observers could simulate a subtraction of red from the match side by adding it to the test side

green test

test

+ R

G+B

0 10 10 10 1


CIE RGB Color Matching Functions

350 400 450 500 550 600 650 700 nm

100

0

435.

8 nm

546.

1 nm

700.

0 nm

?

r(λ)

g(λ)

b(λ)

green test test

R+G

+B

0 10 10 1

CIE XYZ

problem solution: XYZ color systemtristimulus system derived from RGBbased on 3 imaginary primariesall 3 primaries are imaginary colorsonly positive XYZ values can occur!1931 by CIE(Commission Internationale de l’Eclairage)


X

Y

Z


RGB vs. XYZ

negative component disappearsy(λ) is the achromatic luminance sensitivity

350 400 450 500 550 600 650 700 nm

r(λ)g(λ)

b(λ)

350 400 450 500 550 600 650 700 nm0

x(λ)y(λ)z(λ)

1

RGB system XYZ system

amounts of RGB primaries needed to display spectral colors

amounts of CIE primaries needed to display spectral colors

CIE Color Model Formulas

XYZ color model C(λ) = X·X + Y·Y + Z·Z(X, Y, Z are primaries)normalized chromaticity values x, y

( z = 1 – x – y )

complete description of a color: x, y, Y


ZYXXx ++= ZYX

Yy ++=1

11 X

Y

Z

CIE Chromaticity Diagram

identifying complementary colorsdetermining dominant wavelength & puritycomparing color gamuts


spectral color positions are along the boundary curve

spectral colors

purple line

y

x

purple line contains all mixtures of red and blue

Properties of CIE Diagram (2)


representing complementary colors in the chromaticity diagram

C1

C2

C



determining dominant wavelength and purity with the chromaticity diagram

C1 → Cs

C2 → Cp?→ complement Csp

C1

C2

C

Csp

Cp

Cs



gamut of a typical RGB monitor

only the colors inside the triangle can be produced

R

G

B

Color Spaces

Color Metric Spaces (CIE XYZ, L*a*b)used to measure absolute values and differences - roots in colorimetry





RGB Color Model

primary colors red, green, blueadditive color model (for monitors)


C(λ) = R·R + G·G + B·B

white(1,1,1)

yellow(1,1,0)

red(1,0,0)

green(0,1,0)

cyan(0,1,1)

magenta(1,0,1)

blue(0,0,1)

black(0,0,0)

RGB Color Model Images


3 views of the RGB color cube

Gamuts of RGB Monitors

Werner Purgathofer 45Werner Purgathofer 45

monitor gamuts can be very differentno monitor can display all colors

CMY Color Model

primary colors:cyan, magenta, yellowsubtractive color model(for hardcopy devices)

C = G + B, using C “subtracts” R


−

=

BGR

YMC

111yellow

(0,0,1)

red(0,1,1)

green(1,0,1)

cyan(1,0,0)

magenta(0,1,0) blue

(1,1,0)

black(1,1,1)

white(0,0,0)

CMY Color Model Images


3 views of the CMY color cube

Gamuts of CMY(K) Printers

Werner Purgathofer 48Werner Purgathofer 48

printer gamuts can be very differentno printer can display all colors

Color Spaces

Color Metric Spaces (CIE XYZ, L*a*b)used to measure absolute values and differences - roots in colorimetry





Color Ordering Systems (COS)

primary aim: enable theuser to intuitively choose colour values according to certain criteriachoice can yield single or multiple colour valuesexamples: HSV, HLS,Munsell, NCS, RAL Design, Coloroidused in bottom-up parts of a design processsometimes physical samples are provided


HSV Color Modelmore intuitive color specificationderived from the RGB color model:

when the RGB color cube is viewed along the diagonal from white to black, the color cube outline is a hexagon


RGB Color Cube

G

B

white

B

Color Hexagon

HSV Color Model Hexcone

color components: hue (H) ∈ [0°, 360°]saturation (S) ∈ [0, 1]value (V) ∈ [0, 1]


HSV hexcone

HSV Color Definition

color definitionselect hue, S=1, V=1add black pigments, i.e., decrease V add white pigments, i.e., decrease S


Shades

S

cross section of the HSV hexcone showing regions for shades, tints, and tones

HLS Color Model


HLS double cone

color components: hue (H) ∈ [0°, 360°]lightness (L) ∈ [0, 1]saturation (S) ∈ [0, 1]

Color Model SummaryColorimetry:

CIE XYZ: contains all visible colors

Device Color Systems:RGB: additive device color space (monitors)CMY(K): subtractive device color space (printers)

Color Ordering Systems:HSV, HLS: for user interfaces


Color Symbolism: Some Aspects

6 to 11 basic colorscategories, hierarchiesdependent on context / applicationlarge variation in use

what is red? what is blue?what is white? !


Color in Religion

Islam: greenBuddhism:yellow, orange,red & purpleHinduism:orange, blue& blue-violetChrists:liturgical colors withouttheological connex


Political Symbol Colors

partiesrevolutions / movementsflags


at homewater pipeselectrical wireswaste separation

traffictraffic signstraffic lightsparking conceptspublic transport

...

Color Labeling


Color Labeling

technologyresistorsthermochrome colors

naturecourtship [Balz]warning colorsprotective mimicry[Tarnfarben]

…


Color Effect: BLUEdistancefaithfulness [Treue]loyalitydesirephantasymaledevinepeacecold…


Color Effect: RED

bloodenergylovefemalerich, noblelabor movementwarmcorrections…


Color Effect: GREEN

profityoung lovehopeprematurity, unripepoisonnatureneutralenvironment protection…


Color Effect: YELLOW

sunoptimismenlightenmentjealousy [Neid]

stinginess [Geiz]

warning colorwarm…


Color Effect: BLACK

end, deathsadnessnegative emotionsbad luckeleganceemptinesscold…


color and color models - tu wien · color spaces. color metric spaces (cie xyz, l*a*b) used to...

Documents

color and color models - tu wien · color spaces. color metric spaces (cie xyz, lab) used to...