On the Perception of Visuals

Prepared by Gitta Domik

Visualizations/Pictures and Visual Attributes

Visualizations/Pictures

• entirety of graphical objects and their visual attributes as result of visualization process

Visual attributes

• mode ("flavor") of presentation chosen, e.g. color, size, orientation
•  
• clever choice of visual attributes is paramount to visualization process
•  
• redundancy of visual attributes enhances interpretability

[BER67], [TUF83], [KEL93]

Interpretation of visual attributes

• Innate reaction to visual attributes
  • natural to interpret, usually simple
  •  
  • example: increasing brightness => increasing numerical values
  •  
  • preconscious/preattentive interpretation
  •  
• Acquired reactions to visual attributes
  • acquired through education, usually more complex
  •  
  • example: color ranking, street/travel signs, isolines, isosurfaces
  •  
• Illusory visual attributes
  • well documented illusions (not necessarily well understood!)
    •  A Web site for optical illusions
  •  
  • example: illusory triangle, color contrast, Machband effect and others

[GER94][FOS95]

Visual attributes discussed in depth

Color

• psychophysical process
  • physics: relates to wavelengths, spectral distribution and amount of light entering eye
  •  
  • psychology: perceived sensation with no linear relation to physics
  •  
• no complete theory (three types of cones: R,G,B; opponent theory)
•  
• variety of color spaces: geometric descriptions of color gamut
•  
• perceptual dimensions of color: hue, saturation, intensity
• • may be varied independently or in connection to each other
  •  
  • hue: "colors" of rainbow (relates to wavelength)
  •  
  • saturation: "paleness" of color is lack of saturation (relates to spectral distribution)
  •  
  • intensity: light/dark colors (relates to amount of light entering eye) - brightness

Hue

• effective use for nominal data types and ordinal data types (color scale!)
•  
• hints
  • small blue objects: disadvantage for short-wavelengths cones
  •  
  • blue (cool colors): farther away, cooler, lower or negative values
  •  
  • red (warm colors): nearer, warmer, higher and positive values; danger
  •  
  • shape of object displayed with rainbow scale may not be readily apparent
  •  
  • hues may change appearance on different backgrounds
  •  
  • "color-blindness"
  •  
  • ranking of hues not inherent
  •  
  • discontinuous color scales

Saturation

• effective use of saturation for ordinal data types
•  
• careful when interpreting saturation and brightness independently
•  
• 2-dimensional color scales for effectiveness

Brightness

• effective use of brightness for ordinal (and quantitative) data types
•  
• hints
• • bright objects on dark background look bigger than dark objects on bright background
  •  
  • fading brightness gives impression of distance/depth
  •  
  • absolute brightness not perceived linearly
  •  
  • change of brightness not perceived linearly (Machband)
  •  
  • brightness contrast influences perception of brightness
  •  
• 2-dimensional color scales for effectiveness

Texture

• effective use for nominal data types
•  
• hints
  • careful with overlapping textures
  •  
  • textures may give rise to other impressions, e.g. density
  •  
• include legend 

Orientation

• hints
  • familiarity of shape often connected to orientation
  •  
  • symmetry around vertical axis preferred
  •  
• use various orientations to assure correct view of objects

Depth attributes

Use depth attributes to enhance the perception of 3-d structures

• fading brightness to show increasing depth
•  
• perspective geometry to show increasing depth
•  
• occlusion to distinguish back/front
•  
• transparency/translucency to distinguish back/front
•  
• change of brightness (shading) to simulate surfaces
•  
• rotation/"rocking" to enhance 3-d perception
•  
• stereo effect: anaglyph, shutter glasses, VR

Motion

• frame update rate to perceive motion
  • at least 10 frames/sec [BRY94]
  •  
• Examples
  • animation
  •  
  • flick two or more images to depict differences, similarities

Perception Issues in Visualization

HyperVis Table of Contents

Last modified on March 29, 1999, G. Scott Owen, owen@siggraph.org