Using letter identifiability to predict readability of
transparent text on textured backgrounds

Lauren F. V. Scharff
Stephen F. Austin State University, Nacogdoches, TX

Albert J. Ahumada
NASA Ames Research Center, Moffett Field, CA

Introduction

Scharff and Ahumada (2002 JOV) measured text readability for two types of transparent text: additive (as occurs in head-up displays) and multiplicative (which occurs in see-through LCD virtual reality displays). Text contrast (30% and 45%) and background texture pattern were also manipulated.

Their masking index, combining text contrast and background RMS contrast, predicted search times well (r = 0.91) and correctly predicted the better performance of multiplicative vs. additive text.

Unfortunately, the index predicted that the “wave” pattern (Figure 1b) would be more detrimental to reading than the “culture” background (Figure 1a) because the wave background had a larger RMS contrast (0.27 and 0.15, respectively). Additive low contrast text on the culture pattern was the most difficult condition to read

Figure 1. (below) The task was to find a target word (triangle, circle, or square) in the text and click on the corresponding shape below. The backgrounds are (left to right) the culture pattern, the wave pattern, and a plain background.

Figure 1a

Figure 1b

Figure 1c

Methods

In the above experiment the task was to find a target word in text paragraphs. For this experiment, the target words were cut from the original text stimuli, so that the backgrounds exactly matched those in the original conditions. 28 participants completed the word task, and 14 of those did the letter task for the additive transparency condition.

Word task: Individual target words were placed on a plain background. As in the original experiment, there were three shapes at the bottom of the screen, and the participant’s task was to click on the corresponding shape as quickly as possible.

Letter task: Each of the target words was chopped into single letters and presented on a plain background.

Results

For each participant, median reaction times for each condition were calculated for each of the experiments. Accuracy was also recorded. For the word task, accuracy was close to perfect.

Figure 2. The left bar set shows the log search times in seconds for the three backgrounds in the previous experiment. The center bar set shows the log recognition times for the target words alone. The right bar set shows the log of the number of errors in the letter task.

    Results

For each participant, median reaction times for each condition were calculated for each of the experiments. Accuracy was also recorded. For the word task, accuracy was close to perfect.

As Figure 2 shows, the recognition times in the word experiment varied with background in the same way as the search times in the paragraph experiment. The average letter identification accuracy is the same for the two textured backgrounds, so it does not provide an explanation for the poorer performance on the culture background.

The number of times that all the letters of a word were incorrectly identified is a letter task measure that correlates with the word and paragraph task performance:

 
     Culture Wave Plain 
       25     14    0

Conclusions

Readability of text on complex backgrounds can not be predicted by the average RMS contrast when the background is not homogeneous at the scale of the letters. In our tasks, words were identifiable if any of the letters were identifiable.

Reference

Scharff, L. F. V., & Ahumada, A. J. Jr. (2002 JOV). Predicting the readability of transparent text, Journal of Vision.

Acknowledgements

Funding for this work was provided by the Airspace Operations Systems (AOS) Project of NASA's Airspace Systems Program. It was partially supported by NASA Ames Research Center cooperative agreement NCC 2-1095 with the San Jose State University Foundation. We are grateful for the assistance of Robin Rustad and Lori Shird.