BACKGROUND AND RESIDUAL NOISE REDUCTION FOR SPEECH ENHANCEMENT: SPECTRAL SUBTRACTION AND MEDIAN FILTER APPROACH

Disha Desai
2015 unpublished
Analysis of speech signal is used to illustrate parts of speech that comprise momentary the speech component and develop an algorithm to extract those components. In noisy environment in which channel is bit noisy, speaker and listener may experience several annoying noises which causes significant degradation in efficiency of communication system which makes communication bit difficult. It may happen that speaker has to speak much louder and listener has to give more concentration in listening
more » ... process to maintain fruitful communication. Spectral subtraction method is normally used for background noise removal but there will be still some residual noise is present in this enhanced speech that needs to be suppressed for efficient communication system. Median filter can be used for second stage enhancement which removes residual noise from already enhanced speech in first stage.
fatcat:hamthsxlhjg5hjbclkval3avbi