English

Neural Network Alternatives to Convolutive Audio Models for Source Separation

Sound 2017-09-26 v1 Audio and Speech Processing

Abstract

Convolutive Non-Negative Matrix Factorization model factorizes a given audio spectrogram using frequency templates with a temporal dimension. In this paper, we present a convolutional auto-encoder model that acts as a neural network alternative to convolutive NMF. Using the modeling flexibility granted by neural networks, we also explore the idea of using a Recurrent Neural Network in the encoder. Experimental results on speech mixtures from TIMIT dataset indicate that the convolutive architecture provides a significant improvement in separation performance in terms of BSSeval metrics.

Keywords

Cite

@article{arxiv.1709.07908,
  title  = {Neural Network Alternatives to Convolutive Audio Models for Source Separation},
  author = {Shrikant Venkataramani and Y. Cem Subakan and Paris Smaragdis},
  journal= {arXiv preprint arXiv:1709.07908},
  year   = {2017}
}

Comments

Published in MLSP 2017

R2 v1 2026-06-22T21:52:18.912Z