A Discrete Wavelet Transform based analysis of sounds of some musical instruments

In this paper a technique based on Discrete Wavelet Transform has been applied to characterize the sounds produced by some traditional Indian musical instruments, viz. sarangi, harmonium and tumbi. It is observed that when the sound signal is decomposed for different levels and the energy in the wavelet coefficients is calculated for each level. The energy retained in wavelet coefficients varies with the type of wavelet and its decomposition level. The maximum level of decomposition for the three sound signals is different and the signals are also reconstructed with the wavelet coefficients only up to the maximum level of decomposition. The hearing perception of the reconstructed signals is in agreement with the theoretical models.