Abstract
Typeface design has long been considered an art, one guided by a long accumulation of best practices. Differences between typefaces can be obvious, such as the flourishes of a serif typeface versus the clean outlines of sans-serifs, or they may be minor, such as the variability of stroke width within a letter. Here we employ psychophysical techniques to investigate the legibility of two seemingly similar typefaces, a "humanist" style typeface and a "square grotesque". Subjects participated in a yes/no (2AFC) task in which they determined whether a briefly presented stimulus was a word or a pseudoword (a combination of letters that is not an English word, but is pronounceable). Each subject was exposed to four conditions: two typefaces by two color combinations (black text on a white background or vice versa). All text was shown using the computer operating system's standard text rendering algorithm. Stimulus presentation time was controlled via a 2-down, 1-up adaptive staircase procedure, and presentation time thresholds were calculated separately for the four combinations of typeface and color. Results indicate that presentation thresholds are significantly lower for the humanist typeface compared to the square grotesque, and are also lower for the black on white conditions compared to white on black. The typefaces chosen are both sans-serif with similar stroke widths, and were adjusted to equalize their optical heights. However, our results show that the deeper design characteristics of a typeface can substantially affect its legibility. We speculate that the more open and varied letterforms of the humanist typeface may reduce visual crowding, making it superior to the more uniform and ambiguous letterforms of the square grotesque. Whether the legibility advantage of black on white text arises from subtle optimizations in the text rendering algorithm or has a neurological basis is an open question.
Meeting abstract presented at VSS 2014