Lecture #01

Images and Colors

Computer Graphics
Winter term 2019/20

Marc Stamminger
Computer Graphics
• is about the generation of images with computers

• So what is an image in this context?

• A file with extension .jpg, .png or similar ?
• A function mapping colors to a 2D-region ?
• A 2D table of color values ?

Computer Graphics 2019/20 - Images and Colors 2

Computer Graphics
• is about the generation of images with computers

• So what is an image in this context?

• A file with extension .jpg, .png or similar ?
• A function mapping colors to a 2D-region ?
• A 2D table of color values ?

Computer Graphics 2019/20 - Images and Colors 3

Images
• For now, an image is a 2D-array of pixels (Pixel = picture element)
• Each pixel can be
• a bit: black / white
• an integer value (typically 8 bits): grey value
• an integer triple (typically 3x8 bits): color value (red, green and blue)
→ we briefly discuss colors in a minute
• an integer quadruple (typically 4x8 bits): color value with transparency
• float instead of integer: high dynamic range images (HDR images)

w pixels

pixel at (x,y) → buffer[x+w*y]

h pixels
Computer Graphics 2019/20 - Images and Colors 4
Image Files
• Images can be stored in many different file formats
• uncompressed
• compressed
• lossless
• lossy, varying quality
• as movies

Computer Graphics 2019/20 - Images and Colors 5

Image Files
• Example:
• Image size 3888 × 2592
• Raw data: 3888 × 2592 × 3 ≈ 30 MB
• BMP-File: 30 MB
• PNG (compressed, lossless): 13.8 MB
• JPG (good quality): 6 MB
• JPG (low quality): 200 kB

• In a FullHD movie:
• HD-streaming 5 Mbit/s, 25fps:
5𝑀𝑀: 8: 25 = 25 kB !
By Octagon (Own work) [GFDL

• We don’t care about image file formats

(http://www.gnu.org/copyleft/fdl.html) or CC BY 3.0
(http://creativecommons.org/licenses/by/3.0)], via
Wikimedia Commons
in this lecture!

Computer Graphics 2019/20 - Images and Colors 6

Framebuffer
• Memory array storing the visible content of a window on screen
→ size varies with window
• display of windows on screen done by GPU (at least) one framebuffer
→ we don’t care for each window

Computer Graphics 2019/20 - Images and Colors 7

Framebuffer Example
• Directly writing colors to a frame buffer

Computer Graphics 2019/20 - Images and Colors 8

Color Models
• the human eye has tiny light-sensitive sensors in the retina: rods
and three different cones.
• Rods are sensing brightness independent of color
→ they deliver a black&white image
• Cones sense light varying with frequency
• the three types of cones have peak sensitivity at red, green, and blue
frequencies.
• they provide color, but with lower resolution and dynamic range
• Color as perceived by humans is three-dimensional
→ Tristimulus Theory
• Colors are represented as triples of red, green, and blue values

Computer Graphics 2019/20 - Images and Colors 9

Color Models
• There are many ways to define the meaning of these triples
• Technology-driven:
• RGB, CMY, CMYK, YUV, …
• Perception-driven:
• XYZ, HSV, HSB, HSL, Lab, Luv, …
• Standardized:
• XYZ, Lab, …

Computer Graphics 2019/20 - Images and Colors 10

Color Models
• RGB color model
• Red, Green and Blue primaries
• Used (internally) in every monitor
• Additive: adding primitive colors produces white.
• Black = (0,0,0), White = (1,1,1) (=Red+Green+Blue)
• The exact colors of RGB are not defined
Green Yellow
→ colors vary with displays G (0,1,0) (1,1,0)
• RGB is used in framebuffers Greyscale

Cyan White
(0,1,1) (1,1,1)
Black
(0,0,0) R
Red
B (1,0,0)
Blue Magenta
(0,0,1) (1,0,1)

Computer Graphics 2019/20 - Images and Colors 11

Color Models
• CMY color model
• Cyan, Magenta and Yellow primaries
• Used (internally) in hardcopy devices (printers)
• Subtractive (colors subtracted from white background)
→ Adding primary colors produces black.
• Black = (1,1,1), White = (0,0,0)
• Complementary to RGB : C = 1-R, M = 1- G, Y = 1-B
Blue
M Magenta

Greyscale

Red Black
White
C
Cyan
Y
Yellow Green
Computer Graphics 2019/20 - Images and Colors 12
Color Models
• CMYK color model
• Modification of CMY-model
• K for black, additional ‘color’ channel (for saving ink)
• No unique presentation: e.g. (1,1,1,0) and (0,0,0,1)
→ use as much black as possible (it’s cheaper)
• Standard conversion CMY -> CMYK
K = min(C,M,Y) , C = C-K ,
M = M-K , Y = Y-K

Computer Graphics 2019/20 - Images and Colors 13

 Color Models
• The RGB/CMY cube

Blue Cyan

https://commons.wikimedia.org/

Magenta wiki/File%3ARGB_Colorcube_Cor
ner_White.png

White

Black
Green

Red Yellow
Computer Graphics 2019/20 - Images and Colors 14
 Blue
Color Models
Magenta Cyan
• HSV color model (sometimes HSB)
• More human-oriented color specification,
derived from RGB color model
Red
Green
• Look at RGB color cube along the diagonal
• Along the silhouette, we see the “rainbow” Yellow
colors → color hue, defined as angle
with red = 0°, green = 120° and blue = 240° RGB color cube
• White is at the center, from there to the
boundary the saturation increases
• We only see the brightest colors, darker ones are
inside the cube. The select these, we scale the
visible color by the value hue

saturation

Color hexagone
https://commons.wikimedia.org/wiki/File
%3ARGB_Colorcube_Corner_White.png
Computer Graphics 2019/20 - Images and Colors 15
Color Models
• Geometric model for HSV
• circular cone or cylinder

chroma = saturation

https://commons.wikimedia.org/wiki/File%
3AHSV_color_solid_cone_chroma_gray.png

https://commons.wikimedia.org/wi
ki/File%3AHsl-hsv_models.svg
Computer Graphics 2019/20 - Images and Colors 16
Color Models
• RGB → HSV (cone model)
• 𝑉𝑉 = max 𝑅𝑅, 𝐺𝐺, 𝐵𝐵 = 𝑚𝑚𝑚𝑚𝑚𝑚
𝑚𝑚𝑚𝑚𝑚𝑚−𝑚𝑚𝑚𝑚𝑚𝑚
• 𝑆𝑆 =
𝑚𝑚𝑚𝑚𝑚𝑚
𝐺𝐺−𝐵𝐵
60
𝑚𝑚𝑎𝑎𝑎𝑎−𝑚𝑚𝑚𝑚𝑚𝑚 𝑖𝑖𝑓𝑓 max 𝑅𝑅, 𝐺𝐺, 𝐵𝐵 = 𝑅𝑅
𝐵𝐵−𝑅𝑅
• 𝐻𝐻 = 60 + 120 𝑖𝑖𝑖𝑖 max 𝑅𝑅, 𝐺𝐺, 𝐵𝐵 = 𝐺𝐺
𝑚𝑚𝑚𝑚𝑚𝑚−𝑚𝑚𝑚𝑚𝑚𝑚
𝑅𝑅−𝐺𝐺 𝑖𝑖𝑖𝑖 max 𝑅𝑅, 𝐺𝐺, 𝐵𝐵 = 𝐵𝐵
60 + 240
𝑚𝑚𝑚𝑚𝑚𝑚−𝑚𝑚𝑚𝑚𝑚𝑚

• Color pickers often use HSL, which is a variant of HSV

(see definition here):

Computer Graphics 2019/20 - Images and Colors 17

Color Models
• HSV → RGB (cone model)
• 𝐶𝐶 = 𝑉𝑉 × 𝑆𝑆
𝐻𝐻
• 𝑋𝑋 = 𝐶𝐶 × 1 − 𝑚𝑚𝑚𝑚𝑚𝑚𝑚 −1
60°
• 𝑚𝑚 = 𝑉𝑉 − 𝐶𝐶
𝐶𝐶, 𝑋𝑋, 0 𝑖𝑖𝑓𝑓 0° ≤ 𝐻𝐻 < 60°
𝑋𝑋, 𝐶𝐶, 0 𝑖𝑖𝑓𝑓 60° ≤ 𝐻𝐻 < 120°
′ ′ ′ 0, 𝐶𝐶, 𝑋𝑋 𝑖𝑖𝑖𝑖 120° ≤ 𝐻𝐻 < 180°
• R , ,B =
0, 𝑋𝑋, 𝐶𝐶 𝑖𝑖𝑖𝑖 180° ≤ 𝐻𝐻 < 240°
𝑋𝑋, 0, 𝐶𝐶 𝑖𝑖𝑖𝑖 240° ≤ 𝐻𝐻 < 300°
𝐶𝐶, 0, 𝑋𝑋 𝑖𝑖𝑖𝑖 300° ≤ 𝐻𝐻 < 360°
• 𝑅𝑅, 𝐺𝐺, 𝐵𝐵 = 𝑅𝑅′ + 𝑚𝑚 × 255, 𝐺𝐺 ′ + 𝑚𝑚 × 255, 𝐵𝐵′ + 𝑚𝑚 × 255

Computer Graphics 2019/20 - Images and Colors 18

Color Models
• Windows color picker using HSL

hue

saturation

HSL
Computer Graphics 2019/20 - Images and Colors 19
Color Models

https://xkcd.com/1882/
Computer Graphics 2019/20 - Images and Colors 20
In This Lecture We Learn How To
• ... rasterize simple primitives
• ... transform these primitives
• … in 2D
• … in 3D
• … from 3D to 2D (camera!)
• … compute color values on 3D objects → lighting simulation
• … compute occlusion
• … model large scenes of many objects
• … to do all this using a GPU

• … use ray tracing to get shadows and reflections

Computer Graphics 2019/20 - Images and Colors 21

Next lectures …
• Simple Rasterization, Graphics APIs

Computer Graphics 2019/20 - Images and Colors 22