Boosting Objective Scores of a Speech Enhancement Model by MetricGAN Post-processing
This work addresses speech enhancement for applications like noise suppression, but it is incremental as it builds on existing Transformer and MetricGAN methods.
The study tackled speech enhancement by applying a modified Transformer architecture and fine-tuning it with MetricGAN to improve perceptual evaluation of speech quality (PESQ) scores, achieving results that significantly outperformed the Deep Noise Suppression challenge baseline in both subjective and objective evaluations.
The Transformer architecture has demonstrated a superior ability compared to recurrent neural networks in many different natural language processing applications. Therefore, our study applies a modified Transformer in a speech enhancement task. Specifically, positional encoding in the Transformer may not be necessary for speech enhancement, and hence, it is replaced by convolutional layers. To further improve the perceptual evaluation of the speech quality (PESQ) scores of enhanced speech, the L_1 pre-trained Transformer is fine-tuned using a MetricGAN framework. The proposed MetricGAN can be treated as a general post-processing module to further boost the objective scores of interest. The experiments were conducted using the data sets provided by the organizer of the Deep Noise Suppression (DNS) challenge. Experimental results demonstrated that the proposed system outperformed the challenge baseline, in both subjective and objective evaluations, with a large margin.