An approach to telephone communication between deaf and hearing persons

It may be possible for a deaf person and a hearing person to converse over the telephone using special equipment only at the deaf person's end of the connection. The deaf person drives a text‐to‐speech converter with an efficient keyboard system. The hearing person speaks sentences in a word‐at‐a‐time fashion. The words are handled at the deaf person's end by a large vocabulary isolated‐word recognition system that displays a sentence lattice (a sequence of sets of likely words for each word spoken). The deaf person then tries to find a sensible path through the sentence lattice. For such a system to be successful requires adequate performance in: (1) human text generation speed, (2) text‐to‐speech intelligibility, (3) human one‐word‐at‐a‐time speaking, (4) large vocabulary IWR, and (5) human disambiguation of sentence lattices. We report preliminary experimental results that bear on requirements (1), (2), (3), and (5); we outline an approach to (4), the implementation of which is simulated. Experimental studies of (1) and (5) are run on deaf subjects. [This material is based upon work supported by the National Science Foundation under Grant ECS‐8023527.]