Models of Throughput Rates for Dictation and Voice Spelling for Handheld Devices

摘要

Since the emergence of the personal digital assistant (PDA), developers have attempted to create input methods that allow users to enter accurate data at speeds that approach those achieved with the personal computer. Common text entry methods (handwriting and soft keyboard) allow for rates that are unacceptably slow for many purposes. The objective of this paper is to consider the possible benefits of speech-to-text input mechanisms (dictation and voice spelling) for handheld devices. By modeling throughput based on varying rates of speech, correction speeds, and system recognition accuracies, we can compare expected speech throughput rates to current throughput rates for PDAs.