We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.SD

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Sound

Title: TaylorBeamixer: Learning Taylor-Inspired All-Neural Multi-Channel Speech Enhancement from Beam-Space Dictionary Perspective

Abstract: Despite the promising performance of existing frame-wise all-neural beamformers in the speech enhancement field, it remains unclear what the underlying mechanism exists. In this paper, we revisit the beamforming behavior from the beam-space dictionary perspective and formulate it into the learning and mixing of different beam-space components. Based on that, we propose an all-neural beamformer called TaylorBM to simulate Taylor's series expansion operation in which the 0th-order term serves as a spatial filter to conduct the beam mixing, and several high-order terms are tasked with residual noise cancellation for post-processing. The whole system is devised to work in an end-to-end manner. Experiments are conducted on the spatialized LibriSpeech corpus and results show that the proposed approach outperforms existing advanced baselines in terms of evaluation metrics.
Comments: In submission to ICASSP 2023, 5 pages
Subjects: Sound (cs.SD); Audio and Speech Processing (eess.AS)
Cite as: arXiv:2211.12024 [cs.SD]
  (or arXiv:2211.12024v3 [cs.SD] for this version)

Submission history

From: Andong Li [view email]
[v1] Tue, 22 Nov 2022 05:39:02 GMT (1716kb,D)
[v2] Tue, 29 Nov 2022 09:59:27 GMT (1716kb,D)
[v3] Wed, 30 Nov 2022 15:21:51 GMT (1716kb,D)

Link back to: arXiv, form interface, contact.