Parag Chordia

You can now check out the raag emotion survey data for yourself. There is an interactive visualization tool here.

Just got into Philly for the ISMIR conference taking place at Drexel. At first I thought the city was bleak but they were banging the house music on Sunday night. Anyhow, I’m curious about Dmitri Tymoczko talk tomorrow. I wonder if he will address whether his theory of consonance has cognitive grounding and what insights or predictions can be made by the geometric interpretation.

Had a chance to visit Japan for the first time to attend the biennial Music perception and cognition conference in Sapporo. Unfortunately spent almost exactly as long on the plane as on the ground, but still was great. Saw the best water fountain at Moerenuma park (much better than Bellagio), it had a very meditative progression a touch of artifice applied to an elemental process. It was something like a geyser, waves and the tide. It reminded philosophically of the artist Andy Goldsworthy.

I presented work showing that raags do seem to elicit relatively consistent emotions whether or not the listener is enculturated. We also found the age-old minor key = sad, which still awaits a convincing theoretical explanation (which we are working on ) On the other hand, this was an explicit response survey using necessarily crude terms, such as “happy”, “sad”, “restless”, etc. Still, this is among the first studies to empirically show the association of raag and emotion which is taken for granted in Indian music theory. I would like to see if I can find a way to delve deeper using implicit measures. My personal feeling is that the true feelings elicited are complex and not really like emotions elicited by non-musical stimuli. When we project them onto basic axes such as arousal and valence we may get a consistent pattern, but probably we are losing most of the true information. There are more details in the abstract.

I presented my work (virtually) on realtime tabla recognition at ICMC 08. See below for some videos describing the work. I’m really excited about taking this forward so it can be used in realistic performance contexts with top-notch tabla players. You can read the publication here.

I had an interesting opportunity to make the argument recently that music can express wisdom. I was finalist for the Arete Initiative grant that sought new approaches to the study of wisdom. The crux of my claim was that music has a tremendous ability to modulate mood both subtly and dramatically. I think we all have had situations where music has dramatically altered our perspective on a situation, perhaps because it gave expression to something inexpressible, or made us ruminate, or simply because it put us in a better mood. in any case it was a harder sell than I thought it would be. Nevertheless an interesting opportunity to meet people doing work in areas like meditation and neuroscience, trolleyology (morality and neuroscience) and behavioral econ.

I presented a North Indian classical recital on Saturday night at the Sangati center with Ferhan Qureshi on tabla. The main piece was Raag Ahir Bhairav, which was followed by Raag Manjh Khamaj and a short Kafi.

Dangum is a piece for mrdangam, the main South Indian percussion instrument, and a machine listener. It was created by my students Jagadeeswaran Jayaprakash and Alex Rae. The basic format of the piece is call and response. The machine listener is able to listen to a mrdangam improvisation and determine the strokes played and their timing in realtime. This is an extension of Tabla Gyan a system that recognized tabla in realtime. A video of the performance is below. The Listening Machines website is here.

Slow Theka, a piece for processed sarod and a machine improvise that provides rhythmic accompaniment. The rhythmic model is based on the rhythmic and timbral feel of accompaniment in vlambit tintal, a slow cycle of sixteen beats used in North Indian classical music. The accompaniment is generated in realtime using basic rhythmic phrases that are combined according to probabilistic rules that control what strokes should be played when and the overall rhythmic tension at a given point in time. Here is a very brief excerpt from the concert. More details about the piece inside. The Listening Machines website is here.

I, along with my co-creator Alex Rae, spent last week at the Spark music festival in Minneapolis performing our piece for sarod and interactive electronics. You can listen to a version of the piece here and read a bit about it here. The piece followed the basic contour of Indian classical music, with a slow, meditative fist section followed by a rhythmic section. The interaction was based on following the melody and responding to directly to the phrase or more indirectly to more abstract qualities of the phrase. In this piece we are trying to take steps towards a machine listener that reacts to higher-level pitch gestures while at the same time exploiting mistakes that give an interesting chraracter to the computer’s responses.

A video, with synchronized accompanying slides, of a talk I gave at NIPS. There is a fairly easy
to understand introduction to raag as well as some examples that clearly demonstrate how even listeners
easily can hear violations of raag structure even when they are umfamiliar with Indian music.