Utterance-level Permutation Invariant Training with Latency-controlled BLSTM for Single-channel Multi-talker Speech Separation

TL;DR

This paper introduces a latency-controlled BLSTM approach for single-channel multi-talker speech separation, balancing low latency with high performance, and compares training strategies to optimize results.

Contribution

It proposes using latency-controlled BLSTM during inference and compares chunk-level PIT with utterance-level PIT for improved speech separation.

Findings

uPIT outperforms cPIT with LC-BLSTM during inference.

Inter-chunk speaker tracing enhances uPIT-LC-BLSTM performance.

SDR gap between uPIT-BLSTM and uPIT-LC-BLSTM is within 0.7 dB.

Abstract

Utterance-level permutation invariant training (uPIT) has achieved promising progress on single-channel multi-talker speech separation task. Long short-term memory (LSTM) and bidirectional LSTM (BLSTM) are widely used as the separation networks of uPIT, i.e. uPIT-LSTM and uPIT-BLSTM. uPIT-LSTM has lower latency but worse performance, while uPIT-BLSTM has better performance but higher latency. In this paper, we propose using latency-controlled BLSTM (LC-BLSTM) during inference to fulfill low-latency and good-performance speech separation. To find a better training strategy for BLSTM-based separation network, chunk-level PIT (cPIT) and uPIT are compared. The experimental results show that uPIT outperforms cPIT when LC-BLSTM is used during inference. It is also found that the inter-chunk speaker tracing (ST) can further improve the separation performance of uPIT-LC-BLSTM. Evaluated on the WSJ0 two-talker mixed-speech separation task, the absolute gap of signal-to-distortion ratio (SDR) between uPIT-BLSTM and uPIT-LC-BLSTM is reduced to within 0.7 dB.