Speech enhancement is an essential aspect of the field of speech processing research. In most cases the performance of back-end speech technology (such as speech recognition) depends on the quality of speech enhancement output. One typical multi-channel speech enhancement method is microphone array beamforming. However the beamforming performance decreases when it works in a low Signal-to-Noise Ratio (SNR) environment. We propose a speech enhancement method called Paraboloid with Microphone Array (PMA) to improve the microphone array performance. The PMA is a combination enhancement method. It combines a beamforming speech enhancement modular and an acoustic enhancement modular achieved by a paraboloid. This method can be described as follows: (1) The target signal is enhanced by the microphone array and the acoustic focusing method, which we achieved by attaching a circular microphone array to a paraboloid. These two methods enhance the signal from different perspectives, thus making the enhanced signals complementary. (2) We employ the Independent Component Analysis (ICA) method to combine the output from the two abovementioned methods, which achieved the speech signal's secondary enhancement. In this study we analyze the speech enhancement property of the parabolic quantitatively. Computer simulation shows that the proposed method performs well in the low Signal-to-Noise Ratio (SNR) environment. The real-world experimental results are similar to the computer simulation results. Besides, the subjective experiments also verify the feasibility of the proposed method.

Authors: Geng, Yanzhang; Zhang, Tao; Yaw, Mensah Samuel; Wang, Heng
Affiliation: School of Electrical Information Engineering and the Texas Instruments DSP Collaboration Lab, Tianjin University, Tianjin, China
JAES Volume 70 Issue 1/2 pp. 5-23; January 2022
Publication Date: January 23, 2022

Click to purchase paper as a non-member or you can login as an AES member to see more options.

No AES members have commented on this paper yet.

Subscribe to this discussion

To be notified of new comments on this paper you can subscribe to this RSS feed. Forum users should login to see additional options.

Start a discussion!

If you would like to start a discussion about this paper and are an AES member then you can login here:
Username:
Password:

If you are not yet an AES member and have something important to say about this paper then we urge you to join the AES today and make your voice heard. You can join online today by clicking here.