DocumentCode :
1987233
Title :
Performance analysis of enhanced noisy compressed speech signal corrupted by Gaussian and real world noise using recursive filter
Author :
Suman, M. ; Khan, Habibulla ; Latha, M. Madhavi ; Kumari, D. Aruna
Author_Institution :
K.L. Univ., Guntur, India
fYear :
2015
fDate :
2-3 Jan. 2015
Firstpage :
340
Lastpage :
348
Abstract :
Speech Enhancement refers to the improvement in the intelligibility and or the quality of the degraded speech signal using signal processing techniques. Till recent days speech enhancement is a very difficult problem because the noise content in the speech signals varies its nature and characteristics with time and application to application. Using speech enhancement techniques the quality and intelligibility of a speech signal can´t be preserved simultaneously. So generally a trade off is maintained between these two. In speech communication there are number of applications where speech enhancement is required for Example: VoIP, hands free communication, hearing aids, answering machines, speech recognition, teleconferencing systems, car and mobile phones. In this work the main focus is on the development of speech enhancement algorithm that maintains a proper tradeoff between quality and intelligibility in the speech signal. This can be made possible using the time and spectral information in the speech signal. This work also focus on the problem of enhancing the compressed version of the speech signal, to improve the intelligibility of the speech signal. The performance measures like Signal to Noise Ratio (SNR), Mean opinion Score (MOS), Pitch and Formants used to find the performance of a speech enhancement algorithm which varies from application to application.
Keywords :
Gaussian noise; compressed sensing; recursive filters; speech enhancement; SNR; VoIP; answering machines; degraded speech signal; hands free communication; hearing aids; mobile phones; noisy compressed speech signal enhancement; recursive filter; signal processing techniques; signal to noise ratio; speech recognition; teleconferencing systems; Current measurement; Equations; Noise; Noise measurement; Speech; Speech enhancement; Line Spectral Frequencies (LSF); Linear predictive Coding; Multi stage vector quantization;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Signal Processing And Communication Engineering Systems (SPACES), 2015 International Conference on
Conference_Location :
Guntur
Type :
conf
DOI :
10.1109/SPACES.2015.7058280
Filename :
7058280
Link To Document :
بازگشت