Harmonic Product Spectrum based approach for tonic extraction in Indian music

Tonic is one of the integral part of Indian music. It is base pitch for entire rag or melody. In this paper, we have discussed tonic extraction method for analysis of Indian music. The literature witnesses several computationally complex and slightly sluggish algorithms for tonic extraction. We have used Harmonic Product Spectrum (HPS) method to exploit the harmonic nature of music for pitch extraction. The pitch candidates proposed by HPS are optimized using two different methods: Highest peak and template matching. The state of the art and proposed algorithms have also been compared using Hindustani and Carnatic music corpuses. The objective analysis resulted in increased accuracy as compared to already existing algorithms for mono-pitch tonic extraction.