arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1902.07033cs.SDeess.AS
收藏

用于语音分离的低延迟深度聚类

Low-Latency Deep Clustering For Speech Separation

Tampere University of Technology(坦佩雷理工大学)

Shanshan Wang, Gaurav Naithani, Tuomas Virtanen

摘要:

本文提出了一种低算法延迟适应的深度聚类方法,以非特定人语音分离。它包括三个部分:a)使用长短期记忆(LSTM)网络,而不是原始工作中使用的双向变体,b)使用低延迟操作所需的短合成窗口(此处为8 ms),以及c)在音频混合的开始使用缓冲器来估计与组成扬声器对应的聚类中心,然后利用该聚类中心来分离信号其余部分中的扬声器。缓冲持续时间将用作初始化阶段,在该初始化阶段之后,系统能够以8ms的算法延迟操作。我们评估我们提出的方法对两个扬声器的混合物从华尔街日报(WSJ0)语料库。我们观察到,与基线双向LSTM相比,LSTM的使用在源失真比(SDR)方面降低了约1dB的SDR。此外,与基线相比,使用8ms合成窗口而不是32ms使分离性能降低约2.1dB。最后,我们还报告了不同缓冲持续时间的分离性能,指出即使缓冲持续时间低至300 ms,也可以实现分离。

英文摘要:

This paper proposes a low algorithmic latency adaptation of the deep clustering approach to speaker-independent speech separation. It consists of three parts: a) the usage of long-short-term-memory (LSTM) networks instead of their bidirectional variant used in the original work, b) using a short synthesis window (here 8 ms) required for low-latency operation, and, c) using a buffer in the beginning of audio mixture to estimate cluster centres corresponding to constituent speakers which are then utilized to separate speakers within the rest of the signal. The buffer duration would serve as an initialization phase after which the system is capable of operating with 8 ms algorithmic latency. We evaluate our proposed approach on two-speaker mixtures from the Wall Street Journal (WSJ0) corpus. We observe that the use of LSTM yields around one dB lower SDR as compared to the baseline bidirectional LSTM in terms of source to distortion ratio (SDR). Moreover, using an 8 ms synthesis window instead of 32 ms degrades the separation performance by around 2.1 dB as compared to the baseline. Finally, we also report separation performance with different buffer durations noting that separation can be achieved even for buffer duration as low as 300ms.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑