Speech interruption effects in multi-modal tasks

There is a tendency for technology driven development to occur when there are rapid improvements in the power or sophistication of software and hardware. Technology driven development (TDD) can appear to offer an advantage over the current design solutions. However, the advantages of new over old technology may be illusory because the difficulties with current solutions are well understood whereas those of the new technology are largely unexplored. This seems to be particularly true when the technology is embedded in a complex multi-task multi-modal interface and inadequate assessment of technology impact is often worst with regard to decrements in cognitive performance. All too often assessments put great emphasis on preference and not performance as a discriminatory index in choosing between alternatives. Software and hardware for speech input and speech output have improved exponentially in recent years. New algorithms and methods for speech analysis have become possible with cheaper high performance computing. The sophistication of output models have helped to ensure that crude synthetic voices are now a matter of choice and are not the default. The paper discusses the effects of speech interruption in multi-modal tasks