Deep Time Delay Neural Network for Speech Enhancement with Full Data Learning

TL;DR

This paper introduces a deep time delay neural network (TDNN) for speech enhancement that captures long-range temporal contexts efficiently, uses full data learning to maximize training data utilization, and achieves better performance with lower inference time compared to RNNs.

Contribution

The paper proposes a novel TDNN architecture with full data learning for speech enhancement, improving performance and reducing inference time compared to traditional RNNs and BLSTMs.

Findings

Outperforms DNN in speech enhancement tasks.

Achieves comparable or better results than BLSTM.

Reduces inference time significantly compared to RNN-based models.

Abstract

Recurrent neural networks (RNNs) have shown significant improvements in recent years for speech enhancement. However, the model complexity and inference time cost of RNNs are much higher than deep feed-forward neural networks (DNNs). Therefore, these limit the applications of speech enhancement. This paper proposes a deep time delay neural network (TDNN) for speech enhancement with full data learning. The TDNN has excellent potential for capturing long range temporal contexts, which utilizes a modular and incremental design. Besides, the TDNN preserves the feed-forward structure so that its inference cost is comparable to standard DNN. To make full use of the training data, we propose a full data learning method for speech enhancement. More specifically, we not only use the noisy-to-clean (input-to-target) to train the enhanced model, but also the clean-to-clean and noise-to-silence data. Therefore, all of the training data can be used to train the enhanced model. Our experiments are conducted on TIMIT dataset. Experimental results show that our proposed method could achieve a better performance than DNN and comparable even better performance than BLSTM. Meanwhile, compared with the BLSTM, the proposed method drastically reduce the inference time.