Abstract: Auditory Scene Analysis（ASA） is the process in which the auditory system segregates a scene into streams corresponding to different sources. A system for auditory segmentation is proposed via analyzing onsets and offsets of auditory events. The proposed system detects onsets and offsets, generates segments by matching corresponding onset and offset fronts, and resynthesizes these segments to auditory stream for a listening test.

Key words: auditory segmentation, event detection, multi-scale analysis, onset, offset, Computational Auditory Scene Analysis（CASA）

摘要： 听觉场景分析（Auditory Scene Analysis，ASA）系统能将一个场景分解为与不同声源对应的语音流。分割是ASA的主要步骤，借助分割可将一个听觉场景分解成多个片断。实现基于上升缘和下降缘分析的语音分割系统需检测上升缘与下降缘，通过匹配对应的上升缘与下降缘的波前来生成语音片断，将这些片断重构成语音流。

关键词: 语音分割, 事件检测, 多尺度分析, 上升缘, 下降缘, 计算听觉场景分析