Speech Enhancement using Adaptive Filter with Bat Algorithm

• Main Authors: Thaitangam, Pheiga Gangmei, Laishram, Romesh, Devi, Khumukcham Usharani, Khwairakpam, Renuca, Singh, Morangthem Romesh, Oinam, Czenita
• Format: Conference Proceeding
• Language: English
• Subjects: Adaptive filters; component; Filtering algorithms; Finite impulse response filters; formatting; Noise measurement; Optimization; Signal to noise ratio; Speech enhancement; style; styling
• ISSN: 2473-943X
• DOI: 10.1109/ICCIC.2018.8782317

Communication using speech signal is an important aspect of a telecommunication system and the quality of a speech signal can be severely degraded by the presence of background noise. Therefore recovery of the speech signal in such a situation should be the prime focus of the designer and the problem should be dealt with appropriately. Speech enhancement is achieved by designing an adaptive filter for removal of noise from a noisy speech signal. In this paper, we proposed an adaptive filter whose filter coefficients are obtained using optimization techniques based on bat algorithm. The effectiveness of the algorithm is compared with the standard adaptive filter such as least mean square (LMS) and recursive least square (RLS). The results showed that the proposed technique has a better output signal to noise ratio (SNR) and stability. The output SNR of swarm-based algorithms are 2 to 4 dB better than the conventional algorithms for various input SNR. The computational time is compared with particle swarm optimization (PSO) based adaptive filter. The simulation result further showed that bat algorithm has less time complexity.