Presentation is loading. Please wait.

Presentation is loading. Please wait.

Sub-Pixel Text Rendering – Preference, Legibility and Reading Performance Jim Sheedy Jim Sheedy Yu-Chi Tai Yu-Chi Tai Manoj Subbaram Manoj Subbaram Sowjanya.

Similar presentations


Presentation on theme: "Sub-Pixel Text Rendering – Preference, Legibility and Reading Performance Jim Sheedy Jim Sheedy Yu-Chi Tai Yu-Chi Tai Manoj Subbaram Manoj Subbaram Sowjanya."— Presentation transcript:

1 Sub-Pixel Text Rendering – Preference, Legibility and Reading Performance Jim Sheedy Jim Sheedy Yu-Chi Tai Yu-Chi Tai Manoj Subbaram Manoj Subbaram Sowjanya Gowrisankaran Sowjanya Gowrisankaran John Hayes John Hayes

2 Legibility of Fonts James E. Sheedy OD, PhD Manoj V. Subbaram MS John R. Hayes PhD Sponsored by Microsoft Corporation Sheedy JE, Subbaram M, Zimmerman AB, Hayes JR. Legibility of text. Human Factors 2005;47(4):797-815.

3 Font type Verdana Arial Georgia Times New Roman

4 Font Enhancements n Font smoothing –Aliased –Grayscale –ClearType

5 ClearType explained

6 Measuring relative legibility n Size threshold method –Or, back-up method (Tinker) n How small can it be and still be identified? n Essentially, make an acuity chart with each font design to be tested –Measure VA on group of subjects with each chart –Mean group performance represents relative legibility

7 Standardized measurements of visual acuity (ISO) n Landolt ring n Minimum angle of resolution –1/5 of character size MAR

8 Legibility Method - measure VA n Letters or words arranged in logarithmic size progression –100,80,62,50,40,32,25,20,16,12.5 –Or, logMAR steps are: –0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1, 0.0, -0.1, -0.2 n Cannot resolve small letters with normal computer displays n Can measure VA by keeping letter size the same - altering viewing distance in logarithmic progression

9

10 Clear Type – Preference and Performance Support Provided by Microsoft Corporation Sheedy J, Tai Y, Subbaram M, Gowrisankaran S, Hayes JR. ClearType sub-pixel text rendering: Preference, legibility and reading performance. Displays, 2008;29(2):138-151.

11

12 Experiment 1. Preference of CT Implementation Level n Method –30 subjects indicated preference between a pair of text displays (with different CT levels) on a scale of 0 ~ 100 mm. –For each pair, subjects also rated the reason for not selecting the non- preferred display. They rated each of 3 factors on a scale of 0 ~ 100 mm. ColorColor ContrastContrast ClarityClarity

13 Mean preference ratings Textured data are significantly different from preference rating of 50 (which represents no preference)

14 Clear Type preference summary n For smaller font (10) –Significant preference for L1, –Non-significant preference for L2 –Significantly selected against L3 and L4 n For larger font (12) –Quite neutral between CT levels 0-3 slightly higher preference for CT Levels 1 and 2slightly higher preference for CT Levels 1 and 2 differences are not significantdifferences are not significant –Significantly selected against L4

15 Reasons for non-preference Ratings on a scale from 1 to 100 on the contribution of color, contrast, and clarity on the reason for NOT selecting a particular CT level. * p<.05; ** p<.01

16 Reasons for non-preference – summary n For medium-high (L3) and high- level (L4) CT fonts, color is the main reason that one does NOT prefer higher-level CT implementation. n When lower CT levels are preferred, it is because of contrast and clarity

17 Visual acuity and color vision n Hypothesis: CT preference is related to the color vision or visual acuity capabilities of the subject n Subject testing –FM 100 hue color vision test –Visual acuity testing (from legibility) n Neither individual visual acuity nor color vision is significantly related to preferred CT level. n Both visual acuity and color vision bear a positive relationship with magnitude scaling of reasons for selecting one CT level over another.

18 Experiment 3. Central vs. Peripheral Vision n Purpose: –To measure subject’s perception of color at different CT levels n Methods: –A paragraph of 10 point font text was flashed for 200 msec – –Subjects indicated if they saw color (“yes” or “no”) – –Central vs. peripheral Central condition: A 3-line text in the center of the display Peripheral condition: A full-page text without the central 3 lines – –Naïve vs. Informed Naïve measures: Initial measures Informed measures: subjects were measured after being shown a comparison of CT 0 and CT 4 and told thatInformed measures: subjects were measured after being shown a comparison of CT 0 and CT 4 and told that the “color” in CT 4 was the color effect to which the “color” in CT 4 was the color effect to which they were to respond. they were to respond.

19 Central vs. Peripheral stimuli Central presentationPeripheral presentation Fixation point

20 Color Perception and CT – central Mean rating (SEM bars) of perceived color (“yes” = 1, “no” = 0) in central viewing condition as function of CT implementation level

21 Acknowledgment: These studies are supported by Microsoft Corporation. These studies are supported by Microsoft Corporation.


Download ppt "Sub-Pixel Text Rendering – Preference, Legibility and Reading Performance Jim Sheedy Jim Sheedy Yu-Chi Tai Yu-Chi Tai Manoj Subbaram Manoj Subbaram Sowjanya."

Similar presentations


Ads by Google