Evaluation of Audio Based Searching for Indian Traditional Music

The ever-increasing amount of digital audio media in recent years, led to the demand of content based retrieval of multimedia. On the other India classical Music is unaffected from this innovation. Hence, this paper puts forward a comparative study on content based identification and retrieval for Indian Classical Music. The paper concentrates on comparative implementation of various features like Mell Frequency Cesptral Coefficients (MFCC), Bark Scale based features etc for dependable, fast and robust recognition of Indian classical music which can be operate on the contemporary computing platforms. Besides the comparison, a novel customized feature has also been proposed to minimize the drawbacks in the space and time complexity of already presented features. The algorithms have been tested on Indian Classical Music, which demonstrate certain characteristics which make the recognition harder in contrast to Western music.