Comparing User Performance on Parallel-Tone, Parallel-Speech, Serial-Tone and Serial-Speech Auditory Graphs
Résumé
Visualization techniques such as bar graphs and pie charts let sighted users quickly understand and explore numerical data. These techniques remain by and large inaccessible for visually impaired users. Even when these are made accessible, they remain slow and cumbersome, and not as useful as they might be to sighted users. Previous research has studied two methods of improving perception and speed of navigating auditory graphs - using non-speech audio (such as tones) instead of speech to communicate data and using two audio streams in parallel instead of in series. However, these studies were done in the early 2000s and speech synthesis techniques have improved considerably in recent times, as has the familiarity of visually impaired users with smartphones and speech systems. We systematically compare user performance on four modes that can be used for the generation of auditory graphs: parallel-tone, parallel-speech, serial-tone, and serial-speech. We conducted two within-subjects studies - one with 20 sighted users and the other with 20 visually impaired users. Each user group performed point estimation and point comparison tasks with each technique on two sizes of bar graphs. We assessed task time, errors and user preference. We found that while tone was faster than speech, speech was more accurate than tone. The parallel modality was faster than serial modality and visually impaired users were faster than their sighted counterparts. Further, users showed a strong personal preference towards the serial-speech technique. To the best of our knowledge, this is the first empirical study that systematically compares these four techniques.
Domaines
Informatique [cs]Origine | Fichiers produits par l'(les) auteur(s) |
---|
Loading...