COMPUTER MOTHERBOARD ARCHITECTURE WITH INTEGRATED DSP FOR CONTINUOUS AND COMMAND AND CONTROL SPEECH PROCESSING

摘要

PURPOSE: A computer motherboard architecture is provided to change architecture which integrates digital signal processing hardware into the board as the direct audio input interface to handle command and control as well as continuous speech processing. CONSTITUTION: Speech enters the system via microphone input(13), which converts the speech to an electrical signal. The data pathway then leads this signal to the CODEC(14) or a chip, which performs an analog to digital(A/D) conversion of the signal. This digital signal then travels to the DSP(15) for processing. At this point the system enters a decision state(16) as to whether the DSP is operating in command and control or continuous mode. If it is in command and control mode, the actual command is extracted and then sent to the CPU for processing(19). If however, the DSP(15) is operating in continuous speech mode, the DSP must convert it to phonemes at step(18) and then send these to the CPU(19) for processing in conjunction with a software-based speech engine.