AES E-Library

Generalized MMSE STSA Estimator with ICA-Based Noise Estimation and Speech Prior Identification for Binaural Hearing Aids

In this paper, we propose a new method of sound localization preserved noise reduction method for binaural hearing-aid systems. In the previous works, the sound-localization-preserved MMSE STSA estimator with ICA-based noise estimation was proposed. This method can preserve sound localization by using common spectral gain for noise reduction at each channel of the left and right. However, in this conventional method, an approximation is used to estimate optimal common spectral gain. Therefore, it can be consider that the noise reduction performance of the conventional method can be more improved. Thus, we propose to replace the original MMSE STSA estimator by the generalized MMSE STSA estimator. From the objective and subjective evaluation, we confirmed that effectiveness of the proposed method.

Authors: Wakisaka, Ryo; Saruwatari, Hiroshi; Shikano, Kiyohiro; Takatani, Tomoya
Affiliations: Nara Institute of Science and Technology, Ikoma, Nara, Japan; Toyota Motor Corporation, Toyota, Aichi, Japan(See document for exact affiliation information.)
AES Conference:45th International Conference: Applications of Time-Frequency Processing in Audio (March 2012)
Paper Number:2-4 Import into BibTeX
Subject:Representation of Audio

Click to purchase paper or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!

This paper costs $20 for non-members, $5 for AES members and is free for E-Library subscribers.

Learn more about the AES E-Library

E-Library Location: (CD 45thPapers)   /conf/45/aes45-000028.pdf

Start a discussion about this paper!