To Analysis the Effects of Compressive Sensing on Music Signal with variation in Basis & Sensing Matrix

Abstract

Compressive Sensing (CS) is a new approach for compression and reconstruction of compressed signals using very minute observations. These minute observations are also called the number of measurement. The basic benefits of CS are that the number of measurements which are required for proper reconstruction of the compressed signal is very less than the conventional method. If we go through the literature then, we get that for proper reconstruction of signal a theory is given by Shannon. This theory states that the sampling frequency must be higher than twice the highest frequency component in that signal. So the limitation of the conventional method is that it requires so much storage to store and a large bandwidth to transmit the data. Both the things are so much scarce now days, as we know that if we have to required high resolution of signal then the storage which required to store this is also so much high. As there are various parameters in the theory of CS. But the two parameters are so much important than the others. These two parameters are basis and sensing matrices. Various types of other properties like RIP property and IID property also shows a big role in CS theory. By changing the sensing and measurement matrix the SNR value can also be enhanced. In this paper Gaussian matrix is taken as a sensing matrix & DST, DCT considered as the Basis matrices. The combination of basis and sensing matrix enhances the quality & level of compression. As the quality of compression enhanced it enhances the Signal to Noise ratio too. We cannot check the quality by using only one signal so comparison is made using Single Tone, Multi Tone and Vocal Song. 1 1 minimization technique is used for reconstruction of compressed signal