Digital Music

Project Goal

Physical Modeling of Indian Classical Instruments. Identification of Indian Classical Musical Instruments.

An intelligent learning mechanism for musicians to learn hindustani music is proposed. A software workbench is being created to characterize, recognize and preserve Hindustani music. Hindustani classical music is structured around raga, taal, and thaat, music collection can be preserved in digital data base for learning using the workbench thus created.

Benefits

Characteristics of rich cultural heritage of gharanas of Hindustani music will be extracted and query and search functions will be used to look into the digital data base created and musicians can learn Hindustani music with this novel and low cost technique.

Methodology

An opcode is defined for each instrument. When used with proper parameters, the opcode will produce same sound as produced by the corresponding musical instrument.

Given an instrumental composition, some algorithm can be developed to identify the number and type of Indian Classical Instruments used in the composition.

The opcode module can be implemented in the language Csound. Csound provides signal generator and various signal processing tecniques. The opcode module, in a way, will extend the Csound language.

For analysis and feature extraction from a database of musical sounds, artificial neural network and Hidden Markov Model can be used. Timbre and cepstral coefficient are examples of critical features for use in instrument identification.

Project Investigator