Representation of color

1
Representation of color

• CMYK

RGB
There are more, e.g. LAB, YUV, etc.
2
(No Transcript)
3
Comparison between RGB and CMYK
4
Color space conversion

• Useful scheme for photo
• Relative colorimetric - map white point
• Perceptual - preserver visual
• They retain the relationship between the color on
  the photo

5
Brightness and darkness
All channels
Midtones
Bright
Clipping
Dark
6
Redistribution of tone
Original
New
7
Problem after redistribution

• Posterization

8
Posterization

• Round off problem 61/2 60? 61?
• Edit photo in 16bits/channel mode if multiple
  changes are needed
• Use adjustment layer

9
Curve

• Same as level tool but more control
• Little changes in curve result in great change in
  appearance

• Output

• Input

10
White balance

• Which one is white?

11
White balance

• Difficult to judge by looking at only a small
  piece of photo
• Remember the situation when taking photo
• See a whole photo and use experience
• But keep the offset or correct it?
• That depends on what you want

12
Correction of white balance

• Color balance tool
• Add opposite color
• Curve (advance)
• Finely adjust each color channel

13
Sharpness

• Resolution

• Large change in adjacent pixels

14
Sharpen the photo

• Enlargement? No help
• Increase acutance by increasing contrast around
  the edge, i.e. small scaled region contrast
• 1. Smart sharpen
• 2. Unsharp mask (large amount, small radius)
• 3. High pass filter
• Sharpen too much results in halo!

15
Sharpen luminance

• Luminance contain most detail
• Edit a duplicated photo with layer option
  luminosity
• Edit LAB color mode L channel reduce color halo

16
Local contrast

• Increase global contrast reduce sharpness in
  bright and dark areas
• Increase local contrast, i.e. middle scaled
  region contrast
• Unsharp mask (medium amount, medium radius)

17
Noise

• Luminance noise
• Color (chrome) noise
• Happened in underexposed photo, High ISO, night,
  cheap digital camera, etc.
• Reduction in luminance noise reduce details
• Reduction in color noise reduce saturation
• No need to remove all noise

18
Overlap multiple photos to reduce noise

• Averaging several photos
• Produce good quality
• Only possible for static objects
• Be careful to the opacities of each layers

19
Reduce noise tool

• Strength luminance blur strength
• Preserve detail sharp change without blur
• Reduce color noise color blur strength
• Sharpen detail sharpen strength

20
Gaussian blur

• No direct threshold control like preserver
  detail in reduce noise
• Require multiple layers to preserver edge
• Provide greater flexibility in advanced methods

21
Smart blur

• Provide threshold control to preserve sharp edges
• Use multiple times with different threshold and
  radius with fade
• Easy to over do which produce oil painting
  looking

22
File format

• Lossless best quality but may be large size
• e.g. psd, bmp, tif, png
• Lossy good quality for human eyes and small
  size, very good for photos
• e.g. jpg

23
Reference

• Good website for more details
• http//www.cambridgeincolour.com/