The estimation of noise is very crucial part of the speech enhancement; hence the algorithm is propounded to compute the spectrum of noise for the exceedingly variable noisy environment. Mostly, the updating of the noise is done by taking the average of noisy signal's spectrum with the help of smoothing factor that depends upon a time as well frequency and also manipulated based on speech presence probability. Instead of using recursive averaging in this paper exponential function is used to update the spectrum of noisy speech. The enhanced clean speech spectrum is obtained by the noisy signal spectrum and gain function. Further, the gain function is established by using the spectrum of noise power which is measured on the behalf of speech presence probability (SPP) which is the determined by the Exponential function. Objective evaluations like Perceptual Evaluation of Speech Quality (PESQ) and Segmental SNR are utilized to measure the accomplishment of the proposed algorithm. Simulation outcomes reveal the proposed algorithm is much preferable than the existing algorithms.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Noise estimation for real-time speech enhancement


    Contributors:


    Publication date :

    2018-03-01


    Size :

    2188773 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Real-Time Estimation of Accident Likelihood for Safety Enhancement

    Oh, J.-S. / Oh, C. / Ritchie, S. G. et al. | British Library Online Contents | 2005



    A Short-Time Spectrum Estimation Algorithm of Speech Enhancement under the Distortion Control

    Liu, X. / Ban, C. / Feng, X. | British Library Online Contents | 2011


    Real-time image enhancement

    Wong, V. S. | NTRS | 1981


    Suppression of Wideband Noise by Application of Speech Enhancement System

    Liu, X. / Zhou, Y. / Zheng, M. et al | British Library Online Contents | 1998