基于TMS320C6678的合成语音检测系统的设计与实现

摘要

针对合成语音检测系统在大规模电信网应用中的实时性需求，在分析合成语音检测原理和多核DSP任务并行的基础上，提出了一种基于TMS320C6678的合成语音检测算法并行实现方法，该方法实现了任务级并行流水和核间高效通信。实验结果表明，该方法是可行、有效的，并且基于TMS320C6678的合成语音检测系统的实时处理能力有很大提升。%Aiming at the real-time requirement of synthetic speech detection in large-scale telecommunication network application, based on the analysis of synthetic speech detection and parallel task in multicore DSP,this paper designs a method of synthetic speech detection algorithm parallel implementation based on TMS320C6678,the method implements task-level parallel pipeline and efficient inter-core capability. Experimental results show that the method is feasible and effective, and the real-time processing capability of synthetic speech detection system based on TMS320C6678 has improved so much.