SAY WHAT YOU SEE? THE EFFECTS OF BEING ABLE TO SEE WHAT'S HAPPENING ON VOCABULARY DESIGN FOR SPEECH INPUT

摘要

This paper considers the way that vocabulary is used depending on the ability of users to see what they are doing. Blind, blindfolded and able-bodied participants took part in a user trial to complete a speech based task. A 'Wizard of Oz' paradigm was used to mimic the use of a speech recognition system that allowed user to speak freely and still be recognised without any training. The results indicated that the different user groups used speech in different ways. The paper will present these results and develops a number of recommendations for speech input.