Displays 32 (2011) 17

Contents lists available at ScienceDirect

Displays
journal homepage: www.elsevier.com/locate/displa

Effect of light source, ambient illumination, character size and interline spacing on visual performance and visual fatigue with electronic paper displays
Der-Song Lee a, Ya-Hsien Ko b, I-Hsuan Shen c,, Chih-Yu Chao d
a

Department of Industrial Management, Oriental Institute of Technology, Taipei 220, Taiwan Department of Business Management, De Lin Institute of Technology, Taipei 220, Taiwan c Department of Occupational Therapy, Chang Gung University, 259, Wen-Hwa 1st Road, Kwei-Shan, Tao-Yuan 330, Taiwan d Department of Industrial Management, National Taiwan University of Science and Technology, Taipei 106, Taiwan
b

a r t i c l e

i n f o

a b s t r a c t
This study investigated the effects of light source, ambient illuminance, character size, and interline spacing on visual performance and visual fatigue in using commercial electronic paper displays. Regarding visual performance the results showed that display type, character size, interline spacing had signicant effects on search time. Electrophoretic electronic ink display had a shorter search time than chlorestic liquid crystal display. Searching time decreased as character size and interline spacing increased. Ambient illumination, display type, character size, and interline spacing had signicant effects on accuracy. Accuracy was highest for 1500 lx ambient illumination. Accuracy of electrophoretic electronic ink display was greater than chlorestic liquid crystal display. Accuracy increased as character size and interline spacing increased. Regarding visual fatigue, results showed that light source and ambient illumination had non-signicant effects on change of critical icker fusion (CFF) and subjective visual fatigue. Results could be able to provide some guidelines for consumers to choose a suitable electronic paper according to lighting condition and set appropriate character size and interline spacing. 2010 Elsevier B.V. All rights reserved.

Article history: Received 2 April 2009 Received in revised form 5 July 2010 Accepted 10 September 2010 Available online 19 September 2010 Keywords: Electronic paper Visual performance Ambient illumination Character size Interline spacing

1. Introduction With the advances of technology, the use of computer-based information systems to convert, store, process, transmit and retrieve information has increased immensely. Visual display units (VDUs) are the most convenient tools as human/computer interface. Because of the convenience and portability, pocket-sized displays will become the mainstream in the near future. Light weight, low power consumption, portable, and paper-like readability make electronic paper (E-paper) displays the ideal technology solution for reading-intensive handheld devices [1]. E-paper displays can also reduce paper consumption and thus is environmentally friendly. As E-paper has become the headline in the latest visual display topics, there are plenty of related products with different technologies. Among these technologies, electronic paper made of cholesteric liquid crystal (Ch-LC) and electrophoretic electronic ink (E-ink) are the two products which are available in the market. Ch-LC has two stable states: reective planar and focal conic texture. The planar texture reects a specic colored light in a certain angle according to the pitch length. So that the Ch-LC shows a certain color. E-ink comprises millions of tiny microcapsules where a mixture of positively charged white particles and negatively charged black
Corresponding author.
E-mail address: shenih@mail.cgu.edu.tw (I-Hsuan Shen). 0141-9382/$ - see front matter 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.displa.2010.09.001

particles suspended in uid. The black and white image is shown by applying an external electric eld to attract the charged particles on the surface according to the polarity. Ch-LC and E-ink displays use different mechanisms to display images, therefore they have different visual performance. Although they have been available in the market for a while, there are limited studies for their visual performance based on ergonomic considerations. Among these studies, Isono et al. [2] found no signicant differences in the level of visual fatigue between electronic reading and conventional reading. Jeng et al. [3] reported that legibility depends on the illumination intensity but not light source, and conventional paper has a higher visual comfort rating than electronic paper although they have similar performance in the letter-search task. Lee et al. [4] investigated the effects of character size under different level of ambient illumination and light sources on legibility of electronic paper displays and compared them with conventional paper. The results showed that searching speed depends on the luminance, not the light source. However, the effect of ambient illumination or light source on accuracy was not statistically signicant. Shen et al. [5] conducted similar study and found that search speed depends on the ambient illumination but not light source. Accuracy was greater signicantly for electrophoretic display and positive polarity. Ambient illumination is an important factor affecting visual performance and visual fatigue. With regard to level of ambient illumination, recommendations have been reported. For CRT

D.-S. Lee et al. / Displays 32 (2011) 17

workstations, an ambient lighting of 200500 lx is generally suggested. The choice of illumination level greatly depends upon the task (Helander and Rupp [6]). Ostberg [7] reported that a lower ambient illumination might be more appropriate for CRT work. Xu and Zhu [8] studied the effect of ambient illumination and found that performance deteriorated as ambient illumination increased. Chen and Lin [9] suggested that lower ambient illumination (200 lx) was slightly better than higher ambient illumination (700 lx) in terms of both visual recognition and subjective preference. Shieh and Lin [10] found that visual performance was better under 450 lx ambient illumination versus 200 lx for TFT-LCD screen but not for CRT screen. To sum up, ambient illumination about 500 lx or lower is recommended for CRT and LCD screens because they are back lighted. A higher ambient illumination can wash out the images on the screens and possibly cause glares that interfere with visual tasks. However, these recommendations work only for VDUs with backlight. E-Paper displays are VDUs that are reective, require no backlight, and have ultra-low power consumption. Reective paper or E-Paper requires higher ambient illumination to reect the messages on displays. This conjecture is supported in the ndings of several studies [4,5,11]. These studies found that a greater illumination (700 lx or higher) results in greater search speed and accuracy. These ndings were further investigated in this study. Legibility is the attribute of alphanumeric characters that makes it possible for each one to be identiable from others. It is dened as the visual properties of a character or symbol that determine the ease with which it can be recognized in ISO 9241-3 [12]. This means that stroke width, form of characters, and the amount of space between characters and font size can determine legibility [13]. There is an agreement over most international ergonomic standards that the minimum character size for good readability should be from 16 to 22 min of arc as [16]. According to the ANSI/HFS standard [14], font size is required to be a minimum of 16 min of arc and a maximum of 24 min of arc with a preferred range of 2022 min of arc. In ANSI/HFS standard [14] and Shurtleff study [15], character sizes of 1012 min of arc are the minimum size recommended to be used for legibility. Lee et al. [4] found that accuracy was about 90% when the character size was 2.2 mm (15.1 min visual angle) or greater for both conventional paper and E-paper. For smaller character size of 1.4 mm, accuracy decreased quickly. Character size is an important factor affecting visual legibility. The proper character size for electronic paper displays was examined in this study. Kruk and Muter [17] reported that single spacing produced reading that was 10.9% slower than that produced by double spacing. A series of studies were conducted to examine the role of horizontal word spaces on reading eye movements and reading speed [18 21]. When horizontal word spaces were removed from text, reading speed decreased, compared with normal text with spaces. The reading speed decreased approximately 50% [18,21] although smaller magnitudes of the effect were also reported [19]. More recently, Susana [22] showed that increased vertical word spacing, which presumably decreased the adverse effect of crowding between adjacent lines of text, would benet reading speed. Bailey suggested that distance between lines should be at least 50% of character height. A distance of 66% of character height was preferred. The two interline spacing were tested in this study [23]. In summary, many research addressed visual performance, comparing CRT with TFT-LCD [912]. Ergonomic studies related to electronic paper displays were still quite limited. Moreover, few research was found regarding interline spacing evaluation in electronic paper displays. Hence, in this study, we explored the effects of light source, ambient illumination, character size, and interline spacing on legibility of two reective-type E-paper displays by using the method of letter-search task. The ergonomic evaluation and comparison between commercial electronic paper displays could reveal what specications were good enough for reading.

2. Methods 2.1. Experimental design The experiment evaluated ve independent variables: display type, light source, ambient illumination, interline spacing, and character size. Two types of E-paper, Ch-LC display (Kolin i-library) [24] and E-ink display (Sony LIBRIe) [25], were used as display types. There were two light sources: daylight D65 (6500 K), and orescent TL84 (4000 K). Ambient illumination had three levels: 300, 700, and 1500 lx. Character size was dened as the height of a lowercase x (x-height). Three character sizes used were 2.0, 2.5, and 3.0 mm. Interline spacing was dened distance between two adjacent lines (e.g. http://www.plainlanguagenetwork.org/ type/utbo350.htm) which was conveniently expressed as a ratio relative to x height. Two interline spacing were selected: 50% and 66% of the height of character size. Sixty participants were randomly assigned to each of the six treatments of the between-subjects factor (2 light sources 3 ambient illuminations) with 10 participants for each treatment. Each participant completed 12 combinations (2 display types 2 interline spacings 3 character sizes) of the within-subject factors. The method of letter-search task was used in this experiment and this task was found to be practical to evaluate the legibility of a display [26,27]. Four dependent measures; searching time, accuracy, CFF change, and subjective visual fatigue, were analyzed by the method of analysis of variance (ANOVA). Tukey HSD post hoc test was used for multiple comparisons. Accuracy was dened as the number of searched targets divided by the number of total targets. Critical icker frequency (CFF) was the frequency at which a ickering light appears steady. It was a measure method for assessing visual fatigue. CFF was always determined with the method of limits by which the ickering frequency progressively decreased (or increased) until the subject reported a change from fusion to icker (or icker to fusion) [28]. CFF change was the CFF difference before and after the experiment. Subjective visual fatigue was determined by the total score of subjective rating of visual fatigue. All calculations were made with the statistical analysis system (SAS). The level of signicance was a = 0.05. 2.2. Participants The participants were 60 college or graduate students, righthanded, with ages ranging between 18 and 28 (M = 24.32, SD = 2.54). All had corrected 0.8 or better visual acuity with normal color vision. 2.3. Apparatus A Topcon SS-3 screenscope and the Standard Pseudo-Isochromatic Charts were used to examine subjects visual acuity and color vision. The CIE chromaticity coordinates of color were measured with a Minolta chroma meter CS-100. A Kolin cholesterol liquid crystal e-Book Reader (resolution: VGA 640 480 dots, CIE color value foreground 3.2 cd/m2, 0.347, 0.380, background 12.9 cd/m2, 0.374, 0.451) and a Sony E-Ink e-Book Reader (resolution: SVGA 800 600 dots, CIE color value foreground 5.7 cd/m2, 0.323, 0.354, background 23.2 cd/m2, 0.325, 0.356) as shown in Fig. 1 were used to present the experimental material. The color assessment cabinet (VeriVide CAC 1205) was used to control light source and illumination. The illumination was measured with photometer LT Lutron (LX-103). The text was presented dark on light background. Target/Background luminance ratio (Lt:Lb) were set at 1:4 for Ch-LC and E-ink. Participants critical icker fusion frequency (CFF) was measured with Lafayette ick fusion control 12023.

D.-S. Lee et al. / Displays 32 (2011) 17

Fig. 1. e-Book reader: Sony LIBRI and Kolin i-library.

2.4. Conditions of workplace In the experiment, the displays were placed inside the color assessment cabinet with no interference from any external light source during the experiment. The E-paper display was placed on a 730 mm height table. The distance from the center of the display to the surface of the table was 100 mm and to the front rim of the table was 360 mm. The distance from the participants eyes to the screen was 500 mm. A chin supporter was used to ensure all subjects remained the same distance from the screen. The screen inclination was 105. These task set-up parameters were xed to all participants, who were free to adjust their seating positions to their comfort. 2.5. Task and procedure A series of letter-search task was conducted to evaluate the legibility [24,25]. A paragraph of alphabetic pseudotext with character size of 2.0, 2.5, and 3.0 mm of Times New Roman type was used for the task, where 812 targets of character q were embedded in a random strings of letters, digits, punctuation marks, and space. The text was arranged in 20 lines with 25 characters per line as shown in Fig. 2. There were three paragraphs for practice before the experiment started (data collection). Each participant was required to proceed with the following steps: (1) adjust seat to his/her best comfort and set his/her chin on the supporter with both eyes gazing at the center of the screen; (2) scan the text and identify the target q as accurately and quickly as possible and they were given three paragraphs to read under each experimental condition; (3) repeat steps 1 and 2 for all 12 (2 display types 2 interline spacing 3 character sizes) treatment combinations under the designated display type and illumination of a light source. A 1-min break was allowed after nishing step 2. Each participant was required to complete 36 paragraphs (3 12) in total. CFFs were measured before and after the experimental task mentioned before. In addition, a rating scale with six items was administered to collect subjective rating of visual fatigue (Heuer et al., [29]): (1) I have difculties in seeing; (2) I have a strange feeling around the eyes; (3) My eyes feel tired; (4) I feel numb; (5) I have a headache; (6) I feel dizzy looking at the screen. Each item was rated on a 10-point scale for severity of discomfort, with 1 representing not at all and 10 representing very much. Each participant required

90 min to complete the experiment and was paid. NT$1000 (about 30 US$). 3. Results 3.1. Visual performance Table 1 shows the means and standard deviations of searching time and accuracy under each level of independent variables. The results of analysis of variance for searching time and accuracy are as follows. 3.1.1. Searching time Table 2 shows the results of analysis of variance for searching time. The effect of light source and ambient illumination were not signicant. The effect of display type was signicant (F(1, 54) = 159.14, p < .01). Searching speed for Kolin was signicantly slower than Sony. The effect of character size was also signicant (F(2, 108) = 71.28, p < 0.01). The results of paired comparison determined by Tukey HSD post hoc test (Table 3) indicated that searching speed increased as character size increased. Moreover, the effect of interline spacing had signicant effect on search time (F(1, 54) = 130.21, p < .01). Searching time decreased as interline spacing increased. The character size interline spacing interaction was also signicant (F(2108) = 11.09, p < .01). Fig. 3 shows that searching time decreased as interline spacing increased, and the decrease was greater for character size 2.0 mm than for bigger character size 2.5 mm and 3.0 mm. 3.1.2. Accuracy Table 4 shows the results of analysis of variance for accuracy. The effect of ambient illumination was signicant (F(2, 54) = 5.8, p < 0.01). Accuracy increased signicantly as ambient illumination increased. The results of paired comparison determined by Tukey HSD post hoc test (Table 5) indicated that accuracy was greater for the illuminance level of 1500 lx than 300 lx and 700 lx. There is no signicant difference between ambient illumination level of 300 lx and 700 lx. The effect of display type was signicant (F(1, 54) = 34.15, p < .01). Accuracy for Sony was signicantly greater than Kolin. The effect of character size was signicant (F(2, 108) = 25.97, p < 0.01). Accuracy increased signicantly as character size increased. The results of paired comparison determined by Tukey HSD post hoc test

D.-S. Lee et al. / Displays 32 (2011) 17 Table 1 Means and standard deviations of searching time and correct percentage under each level of the independent variables. Independent variables Searching time (s) Mean Light source D65 TL84 Ambient illumination (lx) 300 700 1500 Display type Sony Kolin Line space (%) 50 66 Character size (mm) 2.0 2.5 3.0 48.82 44.71 50.32 43.58 46.4 42.78 50.45 48.75 44.78 50.14 46.4 43.75 (SD) (13.24) (11.76) (14.85) (10.28) (11.58) (10.54) (13.39) (13.16) (11.88) (13.87) (12.20) (11.04) Correct percentage (%) Mean 85.22 86.94 86.2 84.82 91.8 89.34 85.21 86.54 88.01 84.48 87.88 89.47 (SD) (10.37) (5.59) (11.17) (10.73) (8.36) (8.78) (11.57) (11.05) (9.81) (11.61) (9.86) (9.20)

Table 2 ANOVA for searching time. Source Between subjects Light (L) Ambinet illumination (I) Ll Subject within group Within subjects Display type (D) LD ID LID D subject within group Character size (C) LC IC LIC C subject within group Line space (S) LS IS LlS S subject within group DC LDC IDC LIDC D C subject within DS LDS IDS LIDS D S subject within CS LCS ICS LlCS C S subject within DCS LDCS IDCS LIDCS D C S subject within
* **

Df 1 2 2 54 1 1 2 2 54 2 2 4 4 108 1 1 2 2 54 2 2 4 4 108 1 1 2 2 54 2 2 4 4 108 2 2 4 4 108

SS 9112.24 16509.84 683.54 202592.70 34325.44 124.66 1127.55 83.77 11592.33 214817.29 357.88 728.70 186.01 11224.89 8503.86 183.20 65.22 64.89 3526.52 386.83 258.11 552.73 470.35 7546.44 43.19 1.51 74.88 23.46 2611.62 1572.37 67.42 269.50 133.22 7654.79 121.85 65.82 725.83 262.58 9062.98

MS 9112.24 8254.92 341.77 3751.72 34325.44 124.66 563.77 41.89 214.67 7408.65 178.94 182.17 46.50 103.93 8503.86 183.20 32.61 32.45 65.31 193.42 129.05 138.83 117.58 69.87 43.19 1.51 37.44 11.73 48.36 786.19 33.71 67.38 33.30 70.87 60.92 32.91 181.46 65.64 83.92

F 2.43 2.20 0.09

159.14 0.58 2.61 0.19 71.28* 1.72 1.75 0.45 130.21* 2.81 0.50 0.50 2.77 1.85 1.99 1.68 0.89 0.03 0.77 0.24 11.09* 0.48 0.95 0.47 0.73 0.39 2.16 0.78

Fig. 2. Pseudo-text is in the character size and interline spacing on the E-paper.

(Table 5) indicated that accuracy was greater for character sizes 2.5 mm and 3.0 mm than 2.0 mm. The effect of interline spacing had signicant effect on accuracy (F(1, 54) = 4.02, p < .05). Accuracy increased as interline spacing increased. The character size interline spacing interaction was also signicant (F(2108) = 6.82, p < .01). Fig. 4 shows that accuracy increased as interline spacing increased when character size was 2.0 mm. However, accuracy did not change signicantly as interline spacing increased for character size 2.5 mm and 3.0 mm. 3.2. Visual fatigue Table 6 shows the means and standard deviations of CFF and subjective visual fatigue under each level of independent variables. The results of analysis of variance for CFF and subjective visual fatigue are as follows. 3.2.1. CFF change Although the decrement of CFF under D65 was greater than TL84; the difference was not signicant. The effect of ambient illumination on CFF change was not signicant either.

Signicant at a = 0.05 level. Signicant at a = 0.001 level.

3.2.2. Subjective visual fatigue The effects of light source and ambient illumination on subjective visual fatigue were not signicant.

D.-S. Lee et al. / Displays 32 (2011) 17 Table 3 Tukeys multiple range test for searching time among the treatment means of character size. Independent variable Character size (mm) Treatment 2.0 2.5 3.0 Searching time (s)a 50.14 A 46.40 B 43.75 C Table 4 ANOVA for correct percentage. Source Between subjects Light (L) Ambinet illumination (I) Ll Subject within group Within subjects Display type (D) LD ID LID D subject within group Character size (C) LC IC LIC C subject within group Line space (S) LS IS LIS S subject within group DC LDC IDC LIDC D C subject within DS LDS IDS LIDS D S subject within CS LCS ICS LICS C S subject within DCS LDCS IDCS LIDCS D C S subject within
* **

Df 1 2 2 54 1 1 2 2 54 2 2 4 4 108 1 1 2 2 54 2 2 4 4 108 1 1 2 2 54 2 2 4 4 108 2 2 4 4 108

SS 0.0327 1.4144 0.1595 6.5802 0.9226 0.0034 0.1080 0.0001 1.5511 0.9349 0.0194 0.1556 0.0607 1.9437 0.1179 0.0004 0.0196 0.0457 0.9326 0.1021 0.0102 0.0844 0.0958 1.9495 0.0434 0.0090 0.0066 0.0416 1.2180 0.2294 0.0707 0.0949 0.0955 1.8194 0.1081 0.0101 0.0935 0.0348 2.2342

MS 0.0327 0.7072 0.0797 0.1219 0.9226 0.0034 0.0540 0.0001 0.0287 0.4675 0.0097 0.0389 0.0152 0.0180 0.1179 0.0004 0.0098 0.0229 0.0173 0.0510 0.0051 0.0211 0.0239 0.0181 0.0434 0.0090 0.0033 0.0208 0.0226 0.1147 0.0353 0.0237 0.0238 0.0168 0.0541 0.0051 0.0234 0.0087 0.0209

F 0.2683 5.8015* 0.6538

Means with the same letter are not signicantly different.

32.1463* 0.1185 1.8815 0.0035 25.9722* 0.5389 2.1611 0.8444 6.8150* 0.0231 0.5665 1.3237 2.1800 0.2818 1.1657 1.3204 1.9204 0.3982 0.1460 0.9204 6.8274* 2.1012 1.4107 1.4167 2.5885 0.2440 1.1196 0.4163

Fig. 3. Character size interline spacing interaction for searching time.

4. Discussion This study was designed to investigate the effects of light source, ambient illumination, display type, character size, and interline spacing on visual performance and visual fatigue. Results of this study showed that light source had no signicant effect on visual performance and visual fatigue. These results were similar to the ndings of Lee et al. [4] and Shen et al. [5]. E-paper can be used under day light (D65) and orescent light (TL84) and perform as good as normal paper. With respect to ofce illuminance, an ambient illumination of 750 lx was suggested by Illuminating Engineering Society of North America [30], 500 lx by DIN [31], German Industry Standard. Most ofce lighting specied an illumination level of 300500 lx for VDT use [32,33]. On a screen, the additional or excess ambient illumination resulted in a reduction of contrasts within the area of the task. It was referred to as veiling reection. Grandjean [34] cited the work of Benz et al. [35], which showed that 40% of VDT users preferred illuminations between 200 and 400 lx, and 45% preferred illuminations between 400 and 600 lx. Lower ambient illumination could reduce the veiling of the message on the displays, and thus a greater degree of legibility on the displays. Therefore, the ambient illumination level could be a little lower, but not so low as to hinder other ofce tasks (such as typing and data-searching). However, the above recommendations work only for VDUs with backlight. E-Paper displays are reective, require no backlight, and have ultra-low power consumption. E-Paper requires higher ambient illumination to reect the message on displays. Results of this study indicated that

Signicant at a = 0.05 level. Signicant at a = 0.001 level.

an illumination of 700 lx was more appropriate than an illumination of 300 lx and an illumination higher than 700 lx up to around 1500 lx performed even better. Lee et al. [4] and Shen et al. [5] also had similar ndings. In summary, the proper ambient illumination for E-paper display seems to be greater than CRT or LCD, a level above 700 lx is acceptable. The comparison of E-paper types indicated search speed for Sony was signicantly faster than Kolin and accuracy of Sony was greater than Kolin. Shen et al. [5] found the similar ndings that accuracy for Kolin was signicantly lower than Sony and conventional paper. These differences may be related to screen characteristics. Sony has higher resolution (800 600) than Kolin (640 480). Furthermore, the advantage of Sony could be greater if the luminance contrast of Sony is set to its maximum 1:7 (in this study it was set to 1:4 to match the maximal luminance contrast of Kolin). As to the effect of character size, results showed that searching speed and accuracy increased signicantly with the increase of character size. Shurtleff [15], Snyder and Taylor [36] recommended that character sizes for VDTs should subtend a minimum visual angle of 11 or 12 min of arc. Character size can be translated to visual angle by the formula VA (min) = (3438 H)/D, where H is character height, and D is the viewing distance from the eye to display [13]. The three character heights, 2.0, 2.5, and 3.0 mm in this study

D.-S. Lee et al. / Displays 32 (2011) 17

Table 5 Tukeys multiple range test for correct percentage among the treatment means of the independent variables. Independent variables Ambient illumination (k) Treatment 300 700 1500 2.0 2.5 3.0 Correct percentage (%)a 86.20 A 84.82 A 91.80 B 84.48 A 87.88 B 89.47 B

Character size (mm)

Means with the same letter are not signicantly different.

might be more suitable. In general, with the increase of character size the space needed for the text become larger (see Fig. 2), and thus lengthened the time to visual scan the text. However, the results of this study showed that for character sizes 2.0 mm, 2.5 mm, and 3.0 mm, the advantage of greater size outweighs the disadvantage if greater visual scanning was needed. The effects of interline spacing on searching time and accuracy were signicant. Searching time decreased and accuracy increased as interline spacing increased. Increasing vertical word spacing in E-paper display could benet research speed and accuracy. Again, larger interline spacing increased the space needed for a text. But the results showed that for interline spacing 50% and 66%, the advantage of a larger interline spacing outweighs its disadvantage. Moreover, the signicant interactions of character size and interline spacing (Figs. 3 and 4) indicated that the inferiority of smaller character size worsened as interline spacing became smaller. That is, when character size was smaller, the interline spacing as measured by the percentage of character size should be greater. 5. Conclusions Results of this study showed that visual performance and visual fatigue did not depend on light source. Therefore, E-paper can be used under various light sources. In summary, our results are able to provide some guidelines for consumers to choose a suitable electronic paper according to their lighting condition (ambient illumination at 700 lx and greater ambient illumination such as 1500 lx may be even better for E-paper) and set character size at about 3.0 mm and retain appropriate interline spacing about 66% of character size. Better visual performance was shown for E-ink display (Sony LIBRIe) than Ch-LC display (Kolin i-library). With above information, designers and consumers will be assisted to make proper decisions regarding E-paper usage. Acknowledgments This research was funded by the grants from the National Science Council of the Executive Yuan, Taiwan (NSC972221E 161010). References
[1] Omodani, What is electronic paper? The expectations, SID04 Digest (2004) 128131. [2] H. Isono, S. Takahashi, Y. Takiguchi, C. Yamada, Measurement of Visual Fatigue from Reading on Electronic Paper, IDW04, 2004, pp. 16471648. [3] S.C. Jeng, Y.R. Lin, C.C. Liao, C.H. Wen, C.-Y. Chao, K.K. Shieh, Legibility of electronic paper, in: The 5th International Meeting on Information Display, Coex, Seoul, Korea, 2005. [4] D.S. Lee, K.K. Shieh, S.C. Jeng, I.H. Shen, Effect of character size and lighting on legibility of electronic papers, Displays 29 (2008) 1017. [5] I.H. Shen, K.K. Shieh, C.Y. Chao, D.S. Lee, Lighting, font style, and polarity on visual performance and visual fatigue with electronic paper displays, Displays 30 (2009) 5358. [6] M.G. Helander, B.A. Rupp, An overview of standards and guidelines for visual display terminals, Applied Ergonomics 15 (1984) 185195. [7] O. Ostberg, Accommodation and visual fatigue in display word, in: E. Grandjean, E. Vigliani (Eds.), Ergonomic Aspects of Visual Display Terminals, Taylor and Francis, London, 1980, pp. 4152. [8] W. Xu, Z. Zhu, The effects of ambient illumination and target luminance on colour coding in a CRT display, Ergonomics 33 (1990) 933944. [9] M.T. Chen, C.C. Lin, Comparison of TFT-LCD and CRT on visual recognition and subjective preference, International Journal of Industrial Ergonomics 34 (2004) 167174. [10] K.K. Shieh, C.C. Lin, Effects of screen type, ambient illuminance, and color combination on VDT visual performance and subjective preference, International Journal of Industrial Ergonomics 26 (2000) 527536. [11] K.K. Shieh, D.S. Lee, Preferred viewing distance and screen angle of electronic paper displays, Applied Ergonomics 38 (2007) 601608. [12] ISO, ISO 9241-3, Ergonomic Requirements for Ofce Work with Visual Display Terminals (VDTs) Part 3: Visual Display Requirements, Amendment 1: Annex C (Normative): Visual Performance and Comfort Test, 1992.

Fig. 4. Character size interline spacing interaction for correct percentage.

Table 6 Means and standard deviations of CFF change and subjective visual fatigue under each level of the independent variables. Independent variables CFF Mean Light source D65 TL84 Ambient illumination (k) 300 700 1500 1.30 0.27 1.10 0.73 0.53 Change (Hz) (SD) (2.01) (3.67) (2.66) (3.45) (2.80) Subjective visual fatigue Mean 21.33 22.23 25.65 21.00 18.70 (SD) (9.43) (12.11) (11.90) (11.01) (8.14)

subtended 13.8, 17.2, and 20.6 min of visual angle respectively, at a viewing distance of 500 mm. ANSI (Human Factors Society [15]) specied that under the situation where legibility is critical, the minimum character height for a capital letter should be 16 min of visual angle for reading tasks with a preferred character height of 2022 min of visual angle. Translating these into heights at our nominal reading distance of 50 cm gives a minimum height of 2.3 mm and a preferred height of 2.93.3 mm. Results of this study revealed that searching speed and accuracy increased as character height increased from 2.0 to 3.0 mm. Thus, for E-paper, the character height should be greater than 2.0 mm and character height of 3.0 mm that subtends a 22 min of arc of visual angle

D.-S. Lee et al. / Displays 32 (2011) 17 [13] M.S. Sanders, E.J. McCormick, Human Factors in Engineering and Design, McGraw-Hill, Singapore, 1992. [14] American National Standard for Human Factors Engineering of Visual Display Terminal Workstations (ANSI/HFS Standard No. 100-1988), Human Factors Society, Inc., Santa Monica, CA, February 4, 1988. [15] D. Shurtleff, Studies in television legibility: a review of the literature, Information Display 4 (1967) 4045. [16] W.J. Smith, ISO and ANSI Ergonomic Standards for Computer Products: A Guide to Implementation and Compliance, Prentice Hall, Upper Saddle River, NJ, 1996. [17] R.S. Kruk, P. Muter, Reading of continuous text on video screens, Human Factors 26 (1984) 339345. [18] A.B. Spragins, L.A. Lefton, D.F. Fisher, Eye movements while reading and searching spatially transformed text: a developmental examination, Memory & Cognition 4 (1976) 3642. [19] J. Epelboim, J.R. Booth, R.M. Steinman, Reading unspaced text: implications for theories of reading eye movements, Vision Research 34 (1994) 17351766. [20] K. Rayner, A. Pollatsek, et al., Reading unspaced text is not easy: comments on the implications of Epelboim et al.s (1994) study for models of eye movement control in reading, Vision Research 36 (1996) (1994) 461470. [21] K. Rayner, M.H. Fischer, A. Pollatsek, Unspaced text interferes with both word identication and eye movement control, Vision Research 38 (1998) 1129 1144. [22] T.L.C. Susana, Reading speed benets from increased vertical word spacing in normal peripheral vision, Optometry and Vision Science 7 (2004) 525535. [23] R.W. Bailey, Human Performance Engineering: Using Human Factors/ Ergonomics to Achieve Computer System Usability, second ed., Prentice Hall, NJ, 1989. [24] Kolin i-library. <http://www.i-library.kolin.com.tw>.

[25] Sony LIBRIe. <http://www.sony.jp/products/Consumer/LIBRIE/>. [26] M.C. Boschman, J.A.J. Roufs, Text quality metrics for visual display units: I. Methodological aspect, Displays 18 (1997) 3743. [27] M.C. Boschman, J.A.J. Roufs, Text quality metrics for visual display units: II. An experimental survey, Displays 18 (1997) 4564. [28] Y. Salib, G. Plourde, K. Alloul, A. Provost, A. Moore, Measuring recovery from general anaesthesia using critical icker frequency: a comparison of two methods, Canadian Journal of Anaesthesia 39 (1992) 10451050. [29] H. Heuer, G. Hollendiek, H. Krger, T. Rmer, Die Ruhelage der Augen und ihr Ein auf Beobachtungsabatandund visuelle Ermdung bei Bildschirmarbeit, Zeitschrift fr experimentelle und angewandte psychologie 36 (1989) 538566. [30] Illuminating Engineering Society, IES Lighting Handbook, New York, 1981 (Application vol.). [31] DIN 66234, Characteristic Values for the Adaptation of Workstations with Fluorescent Screens to Humans, Parts 19, German DIN Association, 1981. [32] Th. Laubli, W. Hunting, E. Grandjean, Visual impairments in VDU operators related to environmental conditions, in: E. Grandjean, E. Vigliani (Eds.), Ergonomics Aspects of Visual Display Terminals, Taylor & Francis, London, 1982. [33] L.W. Stammerjohn, M.J. Smith, B.G.F. Cohen, Evaluation of work station design factors in VDT operations, Human Factors 23 (1981) 401412. [34] E. Grandjean, Ergonomics in Computerized Ofces, Taylor & Francis, London, 1987. [35] C. Benz, R. Grob, P. Haubner, Designing VDU Workplaces, (German edition: Gestaltung von Bildschirm-Arbeitspltzen), Kln: Verlag TV Rheinland, 1983. [36] H. Snyder, G. Taylor, The sensitivity of response measures of alphanumeric legibility to variations in dot matrix display parameters, Human Factors 21 (1979) 457471.