Adaptive Speech Enhancement Technique using Time Variable LMS Algorithm
JyoshnaGirika1, Md. Zia Ur Rahman2
1JyoshnaGirika, Department of Electronics and Communication Engineering, KoneruLakshmaiah Education Foundation, Vaddeswaram, Guntur, (Andhra Pradesh), India.
2Md Zia Ur Rahman, Department of Electronics and Communication Engineering, KoneruLakshmaiah Education Foundation, Vaddeswaram, Guntur, (Andhra Pradesh), India.
Manuscript received on 02 June 2019 | Revised Manuscript received on 10 June 2019 | Manuscript published on 30 June 2019 | PP: 2713-2718 | Volume-8 Issue-8, June 2019 | Retrieval Number: H7208068819/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Elimination of noise from the speech signal is the majortask in applications like communications, hearing aids, speech therapy etc. This facilitates to provide good resolution speech signal to the user. The speech signals are mainly affected due to the various natural as well as manmade noises. As the nature of these noises random in its nature fixed coefficient filtering techniques are not suitable for clutter elimination task. Hence, in this work an adaptive algorithm has driven noise canceller for speech enhancement applications which has an innate ability to change its weight coefficients depending on the statistical nature of the unwanted component in the original speech signal. In our experiments in order achieve better convergence as well as filtering capability we propose Time Variable Least Mean Square (TVLMS) algorithm rather than constant step parameter. The computational complexity of the speech enhancement process is also a key aspect due to the excessive length of the speech signals in practical scenario. Hence, to lower the computational complexity of the speech enhancement process we propose Sign Regressor TVLSM (SRTVLMS), which is a hybrid realization of familiar sign regressor algorithm and the proposed TVLMS. Using these two techniques noise cancellation models are developed and tested on real speech signals with unwanted noise contaminations. The experimental outputsconfirm that the SRTVLMS based signal enhancement unit performs better than its counterpart with respect to convergence rate, computational complexity and signal to noise ratio increment.
Keyword: Adaptive Noise Cancellation, Convergence Rate, Computational Complexity, Speech Signal, Signal Enhancement, Variable Step Size.
Scope of the Article: Artificial Intelligent Methods, Models, Techniques