Ankita Shukla Journal of Engineering Research and Application www.ijera.com ISSN : 2248-9622, Vol. 8, Issue 7 (Part -IV) July 2018, pp 33-37 www.ijera.com DOI: 10.9790/9622-0807043337 33 | Page Noise Reduction and Echo Cancellation Using Threshold Filters in Hands Free Communication Systems Ankita Shukla, Dr. Vineetasaxena Nigam DDI-PG, Department of Electronics and Communication, UTD-RGPV, Bhopal Professor, Department of Electronics and Communication, RGPV, Bhopal Corresponding author: Ankita Shukla ABSTRACT:-Background noise, far-end acoustic echo, and room reverberation dramatically degrade the performance of many hands-free speech communication systems, in practical environments. For example, for automatic speech recognition system, noises result in the mismatch between the training and testing conditions, further degrading the performance of recognition system in real-world conditions. The threshold filtering can be executed with the help of mask windows of different sizes. For this, first we have to detect the noisy pixel, if the detected pixel is contaminated it can be identified by the homogeneity level of the local region around that pixel. Suppose the signal is corrupted by noise, which means the pixel value, is both 0 and 255 then we must calculate the threshold value and compare it with the neighbouring upper and lower pixels and find out which of them is homogeneous with the threshold value and replace the pixel with that value. In addition to hands-free speech communication systems, the proposed threshold filter system in this thesis is also useful and preferable to many other applications. For example, for speech recognition systems, it is able to improve the recognition accuracy of the received speech signals in adverse environments. Keywords:-Echo Cancellation, Noise, Threshold Filter, Signal to Noise Ratio, Error, Commutation Time ------------------------------------------------------------------------------------------------------------------------------------- Date of Submission: 11-07-2018 Date of acceptance: 25-07-2018 ------------------------------------------------------------------------------------------------------------------------------------- I. INTRODUCTION In Hands-free communication systems, a microphone often picks up reverberation, background noise, and acoustic echoes together with a speaker’s voice which is desired speech signal. Reverberation is due to reflective acoustic environments and leads to degrade the auditory quality of speech signal. This can be remedied by using a DE reverberation technique. DE reverberation consists of recovering a desired speech signal from observed reverberant signals. Several DE reverberation approaches have been proposed [1 -4]. Generally, the research on methods of background noise reduction is being done by the two approaches. One of these is by making use of the single microphone speech enhancement techniques, and the other one is the multi-microphone techniques. Beamforming microphone arrays are very effective since suppress background noise by spatial-temporal filtering without distorting the desired speech signal. Thus the later techniques are preferred to single- microphone techniques. With full-duplex communication, echoes of the loudspeaker signals will join background noise to corrupt the desired speech signal. However, beamforming does not exploit the available loudspeaker signals as reference information for suppressing the acoustic echoes. This is accomplished by acoustic echo cancellation algorithms. In this research work, algorithms will be developed for techniques that allow for removing background noise and acoustic echo from the speech signal before further processing it. II. ACOUSTIC ECHO CANCELLATION In order to suppress echo, several conventional acoustic echo cancellation techniques can be applied. These techniques are based on adaptive filtering techniques. Adaptive filters are a powerful signal processing tool which can be used to model the unknown system and track possible system variations. A large set of adaptive filtering techniques has been developed during the last decades, differing in terms of performance (such as convergence speed, tracking, delay, complexity, and stability). In acoustic echo cancellation, the far- end echo path has to be modeled by the adaptive filter. The echo path is acoustic impulse response from the far-end signal emitted by loudspeaker to the microphone(s). Since this acoustic impulse response can be quite long and highly timevarying, the adaptive filter will require several hundreds or thousands of filter coefficients and high- performance (fast convergence rate), but low complexity adaptive filtering algorithms are desirable. Moreover, the delay introduced by the algorithm cannot be too large. For acoustic applications, cheap algorithms, such as the least RESEARCH ARTICLE OPEN ACCESS