ISCA Archive Interspeech 2013
ISCA Archive Interspeech 2013

Musical noise analysis for Bayesian minimum mean-square error speech amplitude estimators based on higher-order statistics

Hiroshi Saruwatari, Suzumi Kanehara, Ryoichi Miyazaki, Kiyohiro Shikano, Kazunobu Kondo

In this study, we perform a theoretical analysis of the amount of musical noise generated in Bayesian minimum mean-square error speech amplitude estimators. In our previous study, a musical noise assessment based on kurtosis has been successfully applied to spectral subtraction. However, it is difficult to apply this approach to the methods with a decision-directed a priori SNR estimator because it corresponds to a nonlinear recursive process for noise power spectral sequences. Therefore, in this paper, we analyze musical noise generation by combining Breithaupt-MartinĀfs approximation and our higher-order-statistics analysis. We also compare the result of theoretical analysis and that of objective experimental evaluation to indicate the validity of the proposed closed-form analysis.


doi: 10.21437/Interspeech.2013-131

Cite as: Saruwatari, H., Kanehara, S., Miyazaki, R., Shikano, K., Kondo, K. (2013) Musical noise analysis for Bayesian minimum mean-square error speech amplitude estimators based on higher-order statistics. Proc. Interspeech 2013, 441-445, doi: 10.21437/Interspeech.2013-131

@inproceedings{saruwatari13_interspeech,
  author={Hiroshi Saruwatari and Suzumi Kanehara and Ryoichi Miyazaki and Kiyohiro Shikano and Kazunobu Kondo},
  title={{Musical noise analysis for Bayesian minimum mean-square error speech amplitude estimators based on higher-order statistics}},
  year=2013,
  booktitle={Proc. Interspeech 2013},
  pages={441--445},
  doi={10.21437/Interspeech.2013-131}
}