-
摘要: 在回顾了基于语音客观质量评估和计算听觉场景分析的单声道混合语音分离方法的基础上, 针对该方法所采用的ITU-T P.563语音客观质量评估标准存在的使用限制以及计算量大的缺点, 提出了一种采用基于时域包络表示的语音客观质量评估算法来替代P.563算法的单声道混合语音分离方法. 该方法在几乎不降低原方法分离性能的前提下, 大大节约了算法运行所需的时间和资源消耗.Abstract: We firstly review the monaural speech separation method based on objective quality assessment of speech (OQAS) and computational auditory scene analysis (CASA). Considering of the defects of application limitations and time consuming of the employed ITU-T P.563 algorithm, we then propose an alternative method which combines CASA with the temporal envelope representations based OQAS algorithm. The proposed method greatly reduces the operation time and resource requirement, yet almost does not decrease the performance of separation.
-
点击查看大图
计量
- 文章访问数: 1927
- HTML全文浏览量: 55
- PDF下载量: 1364
- 被引次数: 0
下载:
