The current study compared reading performance of text rendered with five different fonts. Two of them, Eido and Maxular Rx, were designed specifically for readers with MD. The other three, Helvetica, Times, and Courier, are widely used fonts. For subjects with MD, significant font differences were observed for MRS, CPS, and RA. Compared with Helvetica and Times, Maxular Rx permitted both smaller CPS and RA, and Eido permitted smaller RA. However, the two new fonts did not present any advantage over Courier. Font variations had less influence on the reading performance of normally sighted subjects, affecting only RA.
We investigated the influence on reading performance of two physical properties that distinguish fonts: interletter spacing (Spacing) and perimetric complexity (Complexity).
Perimetric complexity is readily computable and depicts the “dispersion” of characters
30 and indicates how convoluted the characters are.
31 It has been shown to have a high correlation with other measurements of complexity (e.g., ink density and skeleton length).
32,39 Complexity was not a significant predictor of reading performance for subjects with MD, but it had a small but significant effect on RA for both normally sighted groups; higher values of complexity were associated with better acuity (smaller values of logMAR RA). This result is consistent with a previous study showing that higher complexity induced better recognition of flanked letters, possibly because more complex targets have more features to compete with flankers.
40 However, another study found that the acuity size of Chinese characters increased with complexity (defined as the stroke frequency).
41 One possible explanation for this discrepancy is that higher complexity resulted in larger crowding between components within Chinese characters.
Spacing was found to be a significant predictor for the MRS, CPS, and RA of the MD group. Recall that we define Spacing relative to the size of characters (specifically, relative to x-height), so this property of a font is independent of overall print size. Fonts with larger Spacing had smaller CPS and RA, that is, could be read with smaller characters. It is likely that this effect is due to reduced crowding in the fonts with larger Spacing. Crowding refers to interference with letter recognition by adjacent letters. It is known that crowding contributes to reading difficulty in peripheral vision
42,43 and can be alleviated by increasing the Spacing between letters.
10,23,43–45 For our age-matched and young control groups, RA was better than LA for all five fonts. This advantage of RA over LA is consistent with previous findings.
46,47 It may be due to a context benefit from meaningful sentence reading. However, for the MD group, this effect was absent with Helvetica and Times, possibly because the context benefit was offset by strong crowding in these fonts with smaller Spacing.
Although enhanced Spacing within a font contributes to an ability to read tinier letters, there is a trade-off. The LME model indicated slower MRS for fonts with greater Spacing. This means that the fonts that were most legible for the tiniest print were read more slowly for big print. This trade-off was more prominent for Eido and Maxular Rx than for Courier (
Table 3). One possible reason for this difference is that the subjects had more reading experience with the mainstream font Courier. It is also possible that Courier has some attributes that can compensate for this trade-off. Previous studies have compared reading performance with Courier and Times in individuals with low vision.
8,13 Our finding that Courier yields a smaller CPS and RA is consistent with these studies. However, Mansfield et al.
8 report a faster MRS for Courier than Times, whereas Tarita-Nistor et al.
13 and the current study report no MRS difference between Times and Courier. A meta-analysis that combines the findings from these three studies shows a weak (0.006 log wpm, 1%) MRS advantage for Courier over Times, but this difference is not significant (
P = 0.83, 95% confidence interval −0.044 to 0.055). Differences in the characteristics of the MD samples across the three studies might contribute to the discrepancies; binocular visual acuities for the three samples were as follows: 0.85 ± 0.35 logMAR (Snellen 20/148)
8; 0.47 ± 0.19 logMAR (Snellen 20/59)
13; and 0.99 ± 0.48 logMAR (Snellen 20/200) in the current study.
A salient property of Eido is its unfamiliarity. In a pilot study in which six normally sighted young subjects were briefly trained on the new fonts (26 sentences each), no significant improvement was found. However, whether more extensive training on unfamiliar fonts like Eido would yield better performance could be explored in future studies. A salient property of Maxular Rx is its boldness. In a previous study, Bernard et al.
17 measured reading speed as a function of stroke boldness for Courier text. They found that reading speed was invariant over a wide range of boldness, and decreased when gaps within letters or space between letters grew too small. However, Maxular Rx was designed with sufficient spacing between letters to mitigate any detrimental effects of increased boldness.
What font is best for readers with MD? To answer this question, the reader's goal task, magnification requirement, and display size should all be taken into consideration. When reading a lengthy text passage, readers would want to increase the letter size to at least CPS to achieve fluent reading. The number of characters that will fit on a line will depend jointly on the font's CPS, the font Spacing, and the display size. The best font should allow the most words in one line. Spot reading
48 is another common challenge for readers with MD when small print is used to squeeze a lot of important information into a small text area (e.g., price tags or medicine labels). The best font would allow the text to be read at the greatest viewing distance. Examples of an imaginary MD reader performing lengthy text reading and spot reading are shown in
Figure 6.
How general are our results? Although the designers of the two new fonts intended their designs to be helpful for anyone with MD, it is possible that the design features might be more helpful for people with specific clinical characteristics. In our current study we did not attempt to relate the font benefits to scotoma shape or size or the properties of preferred retinal loci. A more extensive study would be required to assess these relationships. Similarly, we focused on reading with high-contrast text. MD patients with low-contrast sensitivity might be more susceptible to low-contrast text and might benefit more from a bolder font such as Maxular Rx. For readers with very low acuity, it may sometimes be particularly important to extend the range of legible print to smaller size. Magnifiers can help with small print reading, but in situations in which magnifiers are not available or convenient for reading, fonts that yield better RA, such as Eido, Maxular Rx, and Courier, may be beneficial. A further study, taking into account the use of magnifiers, might shed more light on the merits of different fonts for MD reading.
Digital displays make it much easier to manipulate text properties (e.g., font type and size). Our results emphasized the importance of making Spacing, as a ratio of letter size, an adjustable property for digitized texts. Digital displays also make it possible to explore other stimulus dimensions for encoding alphabetic information. For example, Bragg et al.
49 recently introduced a font for low vision based on variations in color and temporal animation. Perhaps the ingenuity of font designers, working with traditional font properties or unusual stimulus features, can still further enhance reading by people with MD.