Speech Enhancement Techniques for Digital Hearing Aids
Tahun : 2019Pengarang : Komal R. Borisagar
Penerbit : © Springer Nature Switzerland AG
Ket : The book explains how noise can be extracted from the silent part of the speech signal. The various types of adaptive filter such as the least mean square (LMS), normalized LMS (NLMS), and the recursive least square (RLS) are described for noise reduction from the noisy speech signal along with voice activity detection (VAD). The different types of practical possible noises generated for the preparation of noisy speech signal and speech can be cleaned effectively. Prediction of the noise signal is the main task in the present work. The methods mentioned are compared based on various parameters such as the amount of noise reduction, the estimation of filter weights and the convergence rate of the filters, mean square error (MSE), and peak signal to noise ratio (PSNR). The presented methods provide significant results in noise reduction and can be observed within a time domain. Also, reconstructed speech can be observed with noise removal and proper intelligence.
Ketegori : DIGITAL TECHNIQUE
