摘要
运用TMS320C5416实现了说话人自动识别系统,系统利用一种新的语音信号r阶的倒谱线性回归系数等参数构成识别的特征矢量集,运用模糊矢量量化技术实现了与文本有关的说话人识别。实验结果表明系统具有识别精度高、识别速度快等特点,是一种有效的说话人自动识别的硬件实现方法。
In this paper, TMS320C541 was used to realize auto speaker recognition system. The system utilizes a new eigenvector aggregate of phonetic signal and takes advantage of fuzzy technique to realize speaker recognition based on text related. The experiment result reveals to be with high recognition rate, real-time, and with little memory consumption, so this system is an effective hardware realization.
出处
《电声技术》
北大核心
2004年第8期35-37,共3页
Audio Engineering
