International Journal of Electronics Engineering Research.
ISSN 0975-6450 Volume 9, Number 3 (2017) pp. 377-390
© Research India Publications
http://www.ripublication.com
A comparative Performance Analysis of LPC
and MFCC for Noise Estimation in Speech
Recognition Task
Syed Sibtain Khalid
1
Asst. professor, Department of Computer Science & Engineering,
Jamia Hamdard, New Delhi-62
Safdar Tanweer
2
Asst. professor, Department of Computer Science & Engineering,
Jamia Hamdard, New Delhi-62
Dr. Abdul Mobin
3
Adjunct Professor, Department Of Computer Science & Engineering,
Jamia Hamdard, New Delhi-62
Dr. Afshar Alam
4
Department Of Computer Science & Engineering,
Jamia Hamdard, New Delhi-62
Abstract
From the days of adam and eve Speech plays an important means in
communication system, its recognition is truly a challenging task. Generally
noise is an unwanted interface in any form of communication system. This
paper deals with the estimation of noise that are present at the front end of
original speech Signal, for estimation of noise we have calculated the
coefficient of LPC (Linear Predictive Coding) and MFCC (Mel Frequency
Cepstral Coefficient). To minimize the complexity of processing only three
classes of noise are considered, ten samples from each classes are taken for
computational work. Accurate deviation results are pin pointed for similar
kind of noises in both LPC and MFCC. Finally percentage deviation is
calculated. Resultant vector shows MFCC outperform LPC in terms of
percentage deviation for estimation of noise that will be the key thrust of this
paper. MATLAB 7.11 is used to carry out all the experimental work.
Keyword: Noise samples, LPC, MFCC, Deviation