A high-performance multiprocessor system for speech processing applications

A universal multiprocessor system has been developed for a variety of real-time speech processing applications, including isolated-word and connected-word recognition and medium-bit-rate speech analysis/synthesis. The throughput capability of the system for speech recognition is about 1000 actual word templates in real time. The main advantage of the system is that the same hardware can be used in different areas of man-machine communication; only different software modules have to be downloaded, depending on the application. Therefore, this system can serve as a basic component in future telecommunication services, where voice input/output is required for dialogue control.<
>