Towards Efficient Models for Real-Time Deep Noise Suppression

• Published in: ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) pp. 656 - 660
• Main Authors: Braun, Sebastian, Gamper, Hannes, Reddy, Chandan K.A., Tashev, Ivan
• Format: Conference Proceeding
• Language: English
• Published: IEEE 06.06.2021
• Subjects: Computational efficiency; Computational modeling; convolutional recurrent neural network; efficient neural networks; Noise reduction; Real-time systems; Recurrent neural networks; Reverberation; Speech enhancement
• ISSN: 2379-190X
• DOI: 10.1109/ICASSP39728.2021.9413580

With recent research advancements, deep learning models are be-coming attractive and powerful choices for speech enhancement in real-time applications. While state-of-the-art models can achieve outstanding results in terms of speech quality and background noise reduction, the main challenge is to obtain compact enough models, which are resource efficient during inference time. An important but often neglected aspect for data-driven methods is that results can be only convincing when tested on real-world data and evaluated with useful metrics. In this work, we investigate reasonably small recurrent and convolutional-recurrent network architectures for speech enhancement, trained on a large dataset considering also reverberation. We show interesting tradeoffs between computational complexity and the achievable speech quality, measured on real recordings using a highly accurate MOS estimator. It is shown that the achievable speech quality is a function of network complexity, and show which models have better tradeoffs.