Convolutive Transfer Function Invariant SDR Training Criteria for Multi-Channel Reverberant Speech Separation

• Main Authors: Boeddeker, Christoph, Zhang, Wangyou, Nakatani, Tomohiro, Kinoshita, Keisuke, Ochiai, Tsubasa, Delcroix, Marc, Kamo, Naoyuki, Qian, Yanmin, Haeb-Umbach, Reinhold
• Format: Conference Proceeding
• Language: English
• Subjects: acoustic beamforming; Array signal processing; complex backpropagation; Measurement; Multi-channel source separation; Signal-to-Distortion Ratio; Source separation; Speech enhancement; Time-domain analysis; Training; Transfer functions
• ISSN: 2379-190X
• DOI: 10.1109/ICASSP39728.2021.9414661

Time-domain training criteria have proven to be very effective for the separation of single-channel non-reverberant speech mixtures. Likewise, mask-based beamforming has shown impressive performance in multi-channel reverberant speech enhancement and source separation. Here, we propose to combine neural network supported multi-channel source separation with a time-domain training objective function. For the objective we propose to use a convolutive transfer function invariant Signal-to-Distortion Ratio (CI-SDR) based loss. While this is a well-known evaluation metric (BSS Eval), it has not been used as a training objective before. To show the effectiveness, we demonstrate the performance on LibriSpeech based reverberant mixtures. On this task, the proposed system approaches the error rate obtained on single-source non-reverberant input, i.e., LibriSpeech test clean, with a difference of only 1.2 percentage points, thus outperforming a conventional permutation invariant training based system and alternative objectives like Scale Invariant Signal-to-Distortion Ratio by a large margin.