DeepFilterNet2: Towards Real-Time Speech Enhancement on Embedded Devices for Full-Band Audio

Schröter H, Escalante-B AN, Rosenkranz T, Maier A (2022)


Publication Type: Conference contribution, Conference Contribution

Publication year: 2022

Book Volume: 17

Conference Proceedings Title: International Workshop on Acoustic Signal Enhancement (IWAENC 2022)

Event location: Bamberg DE

URI: https://github.com/Rikorose/DeepFilterNet

DOI: 10.1109/iwaenc53105.2022.9914782

Open Access Link: https://arxiv.org/abs/2205.05474

Abstract

Deep learning-based speech enhancement has seen huge improvements and recently also expanded to full band audio (48 kHz). However, many approaches have a rather high computational complexity and require big temporal buffers for real time usage e.g. due to temporal convolutions or attention. Both make those approaches not feasible on embedded devices. This work further extends DeepFilterNet, which exploits harmonic structure of speech allowing for efficient speech enhancement (SE). Several optimizations in the training procedure, data augmentation, and network structure result in state-of-the-art SE performance while reducing the real-time factor to 0.04 on a notebook Core-i5 CPU. This makes the algorithm applicable to run on embedded devices in real-time. The DeepFilterNet framework can be obtained under an open source license.

Authors with CRIS profile

Related research project(s)

Involved external institutions

How to cite

APA:

Schröter, H., Escalante-B, A.N., Rosenkranz, T., & Maier, A. (2022). DeepFilterNet2: Towards Real-Time Speech Enhancement on Embedded Devices for Full-Band Audio. In International Workshop on Acoustic Signal Enhancement (IWAENC 2022). Bamberg, DE.

MLA:

Schröter, Hendrik, et al. "DeepFilterNet2: Towards Real-Time Speech Enhancement on Embedded Devices for Full-Band Audio." Proceedings of the International Workshop on Acoustic Signal Enhancement (IWAENC 2022), Bamberg 2022.

BibTeX: Download