论文部分内容阅读
提出一种新的复杂环境噪声下无参考源语音质量客观评价方法.该方法基于准干净语音构造和有参考源评价模型,实现接近于有参考源客观评价的性能.首先,采用改进的最小值控制递归平均算法和多带谱减法获得带噪语音的准干净语音;然后,将该准干净语音作为改进的主观语音质量评估(PESQ)算法的参考语音,计算参考语音与带噪语音之间的失真误差,获得带噪语音的客观评价分值.实验结果表明:该算法客观评分相关度达到0.927,与有参考源语音质量客观评价标准的相关度0.931相比,实现了99%的接近,与无参考源的客观评价标准相比,提高了7.4%.
A new method for objective evaluation of speech quality without reference source under complex environment noise is proposed.The method is based on quasi-clean speech construction and reference source evaluation model to achieve the performance close to the objective evaluation of a reference source.First, using the improved minimum Then the quasi-clean speech is used as the reference speech of the improved subjective speech quality assessment (PESQ) algorithm to calculate the difference between the reference speech and the noisy speech The objective evaluation score of this algorithm is 0.927, which is 99% compared with the 0.931 correlation coefficient of the objective evaluation standard of reference speech quality Compared with the objective evaluation standard without reference source, it increased by 7.4%.