Accounting for Tonal Qualia in Hindustani Music: A Statistical Learning Approach
When listening to music, different tones evoke different phenomenal experiences. In both Western and non-Western music, listeners report that tones variously evoke a sense of expectation, anticipation, surprise, instability, inappropriateness, poignancy, strength, energy, repose, etc. In different contexts the same pitch may evoke dramatically different qualia. What accounts for the different qualia experiences? That is, what contextual properties contribute to the distinctive feelings evoked by a tone? This research proposes to address this question from the perspective of statistical learning. In particular, can statistical learning account for some of the phenomenal experiences evoked by listeners of Hindustani music. (Parag Chordia, David Huron)
Automatic Melody Transcription in Indian Classical Music
Indian Classical Music presents an interesting case for melodic transcription due to the complexity and fluidity of the melody and because no symbolic scores exist. Because of this, automatic transcription tools are particularly relevant. The current research attempts to separate the melodic line from a polyphonic mixture containing tabla (pitched percussive accompaniment) and tanpura (timbrally rich drone). The isolated line is then pitch tracked to create a continuous pitch vs. time representation which can be used to derive more abstract symbolic representations to study a variety of fundamental questions: What types of micro-tonal inflections are used in Hindustani music? Can melodies be identified from the distribution of notes used in passages? What expressive effects are used to highlight melodic expressions? (Parag Chordia)
Automatic Transcription of Tabla
Tabla is the most important percussion instrument in North India; its distinctive timbre is ubiquitous in classical, folk, and popular music. Tabla music is a sophisticated improvisation-based system that focuses on timbre and rhythm, with a complex fingering technique that allows performers to crisply juxtapose strokes of differing timbres. This research attempts to teach a machine to perceive the timbral and rhythmic structure of tabla music. Aside from furthering research in automatic transcription, the immediate motivations for this research are to create representations of tabla performances that can be used for analysis, and that will allow the musical patterns of tabla music to form the basis for new creative works. (Parag Chordia)
Listening Machines
Listening Machines is a concert series featuring pieces by the faculty and students from Georgia Tech's Music Technology group. The concert series explores concepts of machines listening and improvisation and musical human-machine interaction. (Gil Weinberg, Jason Freeman, Parag Chordia, Frank Clark, Chris Moore, Scott Driscoll, Travis Thatcher, Mark Godfrey)