Transform based speech enhancement for auditory signals based on iterative weiner filtering

Abstract

One of the important features of Speech processing is speech enhancement. In a noisy environment, speech enhancement plays a vital role. Many research works are being done in speech enhancement methods in recent years but still, it can't be attained. It mainly depends on Speech intelligibility which can improve the speech quality. In this research work, signal representation is considered and the various transforms are applied and compared. The analysis is done with the help of two parameters and the results are compared. Here the enhancement process is focused on using Advanced DCT (ADCT) and Discrete fractional Cosine transform. The ADCT has the advantage of energy compaction and flexible window switching. Iterative Wiener Filtering is used for filtering the coefficients. Pitch Synchronous Analysis (PSA) is combined for finding the exact pitch period.