cornell cs465 fall 2004 lecture 2© 2004 steve marschner 1 displays

© 2004 Steve Marschner • 1Cornell CS465 Fall 2004 • Lecture 2

Displays

© 2004 Steve Marschner • 2Cornell CS465 Fall 2004 • Lecture 2

Framebuffer

[From Talton]

© 2004 Steve Marschner • 3Cornell CS465 Fall 2004 • Lecture 2

Cathode ray tube (CRT)

• First widely used electronic display– developed for TV in the 1920s–1930s

[H&

B fi

g.

2-2

]

© 2004 Steve Marschner • 4Cornell CS465 Fall 2004 • Lecture 2

Raster CRT display

• Intensity modulated to produce image• Originally for TV

– (continuous analog signal)

[H&

B fi

g.

2-7

]

© 2004 Steve Marschner • 5Cornell CS465 Fall 2004 • Lecture 2

CRT refresh images

© 2004 Steve Marschner • 6Cornell CS465 Fall 2004 • Lecture 2

Interlacing vs progressive scan

© 2004 Steve Marschner • 7Cornell CS465 Fall 2004 • Lecture 2

Interlacing vs progressive scan

© 2004 Steve Marschner • 8Cornell CS465 Fall 2004 • Lecture 2

Interlacing vs progressive scan

© 2004 Steve Marschner • 9Cornell CS465 Fall 2004 • Lecture 2

Vector vs raster scan

Arthur Clokey, the creator of Gumby, trying out NYIT CGL'sBBOP 3D keyframe animation system using an E & S vector display, 1984. Tempest

© 2004 Steve Marschner • 10Cornell CS465 Fall 2004 • Lecture 2

LCD flat panel or projection display• Principle: block or transmit light by twisting its polarization• Intermediate intensity

levels possible bypartial twist

• Fundamentally rastertechnology

[H&

B fi

g.

2-1

6]

© 2004 Steve Marschner • 11Cornell CS465 Fall 2004 • Lecture 2

LCD

© 2004 Steve Marschner • 12Cornell CS465 Fall 2004 • Lecture 2

Color displays

• Humans are trichromatic– match any color

with blend of three

• Additive color– blend images by sum– R, G, B make good primaries

[cs4

17

S0

2 s

lides]

red blue

green

yellow cyan

magenta

white

© 2004 Steve Marschner • 13Cornell CS465 Fall 2004 • Lecture 2

Color displays

• CRT: phosphor dot pattern to produce finely interleaved color images

• LCD: interleaved R,G,B pixels [H&

B fi

g.

2-1

0]

© 2004 Steve Marschner • 14Cornell CS465 Fall 2004 • Lecture 2

Triads and color mixing

SMPTE color bars

closeup on a Sony Trinitron monitor

Slide from Marc Levoy

© 2004 Steve Marschner • 16Cornell CS465 Fall 2004 • Lecture 2

DLP

© 2004 Steve Marschner • 17Cornell CS465 Fall 2004 • Lecture 2

Triads versus pixels

antialiased font(Adobe Acrobat)

subpixel font(Adobe Cooltype)

integral pixel font(IBM LCD)

integral pixel font(Sony Trinitron)

Slide from Marc Levoy

Kindle iPad

http://www.bit-101.com/blog/?p=2722

At 26x

Kindle iPad

http://www.bit-101.com/blog/?p=2722

At 400x

NewsPrint

http://www.bit-101.com/blog/?p=2722

© 2004 Steve Marschner • 25Cornell CS465 Fall 2004 • Lecture 2

How much spatial resolution (pixels) do we need?

© 2004 Steve Marschner • 26Cornell CS465 Fall 2004 • Lecture 2

How much temporal resolution (frames per second) do we need?

dark

bright

(fps)

© 2004 Steve Marschner • 27Cornell CS465 Fall 2004 • Lecture 2

Transfer function of display

• Say pixel value is 123– this means the intensity is 123. 123 what?

100%

00 255

like this?

(voltage)(frame buffer value)

(light)(photons)

© 2004 Steve Marschner • 28Cornell CS465 Fall 2004 • Lecture 2

© 2004 Steve Marschner • 29Cornell CS465 Fall 2004 • Lecture 2

© 2004 Steve Marschner • 30Cornell CS465 Fall 2004 • Lecture 2

Why nonlinear intensity?

~0.000.010.040.090.160.250.360.490.640.811.00

~0.00.10.20.30.40.50.60.70.80.91.0

• Closer to ideal perceptually uniform exponential

© 2004 Steve Marschner • 31Cornell CS465 Fall 2004 • Lecture 2

Checkerboard test

n = 64

n = 128

n = 192

I = 0.25 I = 0.5 I = 0.75

© 2004 Steve Marschner • 32Cornell CS465 Fall 2004 • Lecture 2

© 2004 Steve Marschner • 33Cornell CS465 Fall 2004 • Lecture 2

Gamma correction

[Phili

p G

reensp

un]

OKcorrected forg lower than

display

corrected forg higher than

display

© 2004 Steve Marschner • 34Cornell CS465 Fall 2004 • Lecture 2

8 bpp (256 grays)7 bpp (128 grays)6 bpp (64 grays)5 bpp (32 grays)4 bpp (16 grays)3 bpp (8 grays)2 bpp (4 grays)1 bpp (2 grays)

Quantization

[Phili

p G

reensp

un]

© 2004 Steve Marschner • 35Cornell CS465 Fall 2004 • Lecture 2

• You make a black and white printer. You don’t want your pictures to come out like this. Design a system for converting grayscale images to black/white that will look better than this. (You can only use black/white, what value goes in each pixel?)

© 2004 Steve Marschner • 36Cornell CS465 Fall 2004 • Lecture 2

Ordered dither example

• Produces regular grid of compact dots

[Phili

p G

reensp

un]

© 2004 Steve Marschner • 37Cornell CS465 Fall 2004 • Lecture 2

Diffusion dither

• Produces scattered dots with the right local density

[Phili

p G

reensp

un]

© 2004 Steve Marschner • 38Cornell CS465 Fall 2004 • Lecture 2

© 2004 Steve Marschner • 39Cornell CS465 Fall 2004 • Lecture 2

© 2004 Steve Marschner • 40Cornell CS465 Fall 2004 • Lecture 2