An Analysis on Source-filter Model Based Artificial Bandwidth Extension System

To minimize transmission bandwidth in speech based communication systems such as telephone systems, narrowband representation of the signal is generally used. This band-limited signal can able to represent the vowel sounds without much distortion, but it is not suitable for representing consonant sounds, mostly fricative consonants ( th /, sh/, / Isl /, xl/ ch, /, etc). Artificial Band Extension (ABE) techniques are used to generate a wideband signal from the narrowband signal. Since most of the high frequency components and the fricative consonants were absent in the narrowband representation of the sound, it is a challenging task to create those missing components in the wideband equivalent signal. In this work, implemented a source-filter model based ABE system .The spectral envelope extension is carried out using classified codebook approach method and evaluated its performance with suitable metrics.