The idea of the Extended Sitar Performance system (ESP) was born out of my desire to create an instrument that could express more fully the range of my musical interests.
I have grown up in a time and place where there is a tremendous diversity of cultural influence. It is possible on a given day to hear music from every continent in the world, including Tuvan throat singing, Tibetan chanting, Chinese Opera and folk music, Polynesian dance music, African music from every part of Africa, Persian classical and popular music, Japanese shakuhachi, koto and taiko drumming, Celtic harp music, Jazz and Blues (interpreted by just as wide and various a group of performers from around the world), North & South Indian classical, dance and folk traditionsthe list could go on and on, and each music is often represented by well respected masters of their traditions.
As with many musicians who are exposed to such diversity of music culture I have responded with curiosity and interest. One of my main influences has been North Indian classical music, more specifically the teaching of Ustad Ali Akbar Khan, with whom I have studied for over twenty years. I have also received training and have experience in Blues, Jazz, western classical, various African and Brazilian traditions, Balinese Gamalan, as well as experimental electronic music. I have experienced many kinds of improvised music settings and have performed with a wide variety of musicians from all over the world, performing on instruments such as koto, shakuhachi, jong, pipa, sarod, sarangi, bansuri, harp, tabla, djembe, all the major western instruments, electronic and homemade instruments.
As an exponent of North Indian classical music I feel an obligation to pass on the knowledge of the tradition I have received. At the same time, in being true to who I am, I cannot help but create music that reflects the natural internalization of all my influences. I hope to do so in a manner that respects the deepest knowledge of music that I possess.
Creating the ESP has entailed the bringing together of many diverse forms of music technology - both digital & analog. It required an understanding of the sitar, MIDI, sound synthesis, digital sound processing, acoustics, basic electronics, computer programming, hardware fabrication and, of course, music.
For many years I have dreamed of accessing the sonic potential of sound synthesis, sampling, and real time sound manipulation, using sitar technique; with the ESP, I can realize this goal.
The ESP, as the name implies, is designed for live performance and can be understood as part of the growing musical endeavor known as "Interactive Music." To help define interactive music I will draw upon ideas from Robert Rowe's book: Interactive Music Systems: Machine Listening and Composing (1993).
"Interactive computer music systems are those whose behavior changes in response to musical input." (Rowe 1)
To add to the above statement I would include that the human performer's behavior also changes in response to the computer's musical output.
In interactive music the computer responds to musical input from a human performer and the human performer responds to the computer's output. For an improvisor this circle of interactivity can produce a very interesting musical challenge, making it clear that the burden of making such an interactive system "musical" lies with the human performer as well as with the proper hardware and software design of the system.
The design of an interactive music system can be broken down into three stages: sensing, processing, and response (Rowe 2-23).
The sensing part of an interactive music system turns human musical input into information usable by a computer. The main sensing apparatus of the ESP is a pitch-to-MIDI converter. The converter translates sound into MIDI1 information, specifically: the pitch of a sound, when a note starts and stops, and its initial amplitude.
The creative use of this somewhat limited amount of information lies at the heart of and is much of the art of interactive music.
The information provided by the sensing portion of the system can then be processed by a computer. There are many paradigms of how this information can be processed. One of the main paradigms used in the ESP is that of emulating the processes of a human musician, such as "listening", making musical "decisions", and "responding" in the form of musical output. A program based on this model could be thought of as a virtual "player" (Rowe ch. 4). This paradigm follows closely the same over-all structure of the interactive music system (sensing, processing, responding).
Other paradigms might involve emulating natural processes, such as the formation of crystals, or the changing flow of liquid as it encounters various surfaces and constrictions.
Both of these paradigms often involve the use of aleatoric processes: the use of random numbers and probability tables as part of the decision-making process. This adds the feeling that these processes are less directly controlled by human input as they are influenced by that input.
The processed information is then converted into audible musical sound via sound synthesis (software and hardware synthesizers), soundfile playback (samplers, computer files such as AIFF, CD players, etc.), MIDI-to-cv converters (which produce a control voltage used to control electronic and electro-mechanical devices), or can be used to control sound in other ways, with sound mixing and processing software or hardware.
These concepts will be returned to and expanded in the discussion of specific software developed for the ESP.
As has been noted by Rowe (78), one of the attractions of interactive music systems is "their potential for improvisation." As an improviser I feel that one of the joys of improvisation is dealing with the unexpected. In designing the software for an interactive music system one of the determining factors of musical success is the balance between expected and unexpected behavior. This balance has a great deal to do with degree of ability and personal style.
This brings us to a very interesting question: is "interactive" music a style or a method? Rather than offer an answer here, I would propose this as an ongoing investigation, to be answered more fully by the music made with the ESP. It is an important question for me because I bring to the ESP many concepts that are rooted in a particular kind of music, namely North Indian Classical music.
The Extended Sitar Performance system is an attempt to extend the range of the sitar in ways that use traditional sitar techniques to control analog and digital electronic music processes. The ESP system (or ESP for short) is my name for an instrument that is a collection of connected computers, programs, synthesizers, pitch-to-MIDI converter(s), effects units, microphones, pick-ups, sitar(s), etc. Currently the system consists of:
Hardware (roughly in order of connection starting at the sitar)
1) Roland GK-2A Hex pick-up (attached to the bridge of the sitar with a custom made plate)
2) AKG C419 condenser microphone
3) Roland GI-10 (pitch-to-MIDI converter)
4) MPU-104 MIDI input selector, & Nexus, 3 in, 8 out, MIDI patcher
5) Opcode Professional Plus MIDI interface
6) Power Macintosh 8500 w/internal CD player
7) Macintosh SE
8) Synthesizers: Yamaha TX7, Korg M1
9) Ensonic Mirage Sampler
10) Yamaha SPX-90 effects unit
11) Niche audio control module
12) Mackie 1604 audio mixer
Software
All programs written for the ESP which are discussed in this paper have been written in Max, a graphical programming environment for musical applications (Puckette and Zicarelli ). They include RagaPatch, a virtual "player" which can "learn" raga-like structures and "improvise" in them; GuitarAccompanist, which loosely follows a human player's rhythm and pitch and provides chords, strummed or plucked, based on the human input; QUAD, moves stereo sound between four speakers based on human input, pre-recorded patterns or by drawing a "path" on the computer screen; TXMAX, takes human player input and, with the use of probability tables, re-maps the input to change synthesis parameters on a Yamaha TX7 synthesizer; and CD/AIFF.Player, controls CD & AIFF soundfile playback based on human musical input. Each of these programs are discussed in greater depth later in the paper.
Some work was previously done using the Hierarchical Music Specification Language (HMSL), developed and written by Phil Burk, Larry Polansky, and David Rosenboom at the Center for Contemporary Music (CCM) at Mills College and Mike Berry's GrainWave, also developed at the CCM. Future plans for the ESP will inevitably include use of both HMSL and GrainWave, however there will be no in-depth discussion of them in this paper.
For converting the sitar playing into MIDI data a pitch-to-MIDI converter is needed. After some investigation I decided to use the Roland GI-10 pitch-to-MIDI converter. It can accept two kinds of input, one from its made-for-guitar pick-up, the GK-2, and another from an external microphone (not included with the GI-10). The sitar, when played, is designed to resonate several strings at a time and can produce too much information for a pitch-to-MIDI converter to accurately determine pitch, so using a microphone as input to the GI-10 would create problems and would also make the system vulnerable to other external sounds, possibly transmitting unwanted information. With this in mind I adapted the GK-2 pick-up for use on the sitar.
Traditionally the sitar has three sets of playing strings, on two levels, one above the frets and one below (for a more detailed discussion see section on sitar notation below). The strings are made of steel, brass, and bronze. The GK-2 is essentially an electromagnetic pick-up and therefore is not affected by the vibrations of non-ferrous strings. Fortunately the main playing string of the sitar is made of steel. I was able to fabricate a pick-up holder for the GK-2 with which I could easily adjust the position of the pick-up so that it would only sense the main playing string.
The main playing string of the sitar is designed to be pulled in order to perform ornaments that can cover a range of a major sixth. This causes considerable deflection of the string and had to be taken into account when adjusting the position of the GK-2 pick-up.
With all the potential for sound control that the MIDI set-up of the sitar provides it is still lacking the ability to use the natural sound of the sitar. In order to insure that the sitar's sound would be part of the ESP system I have included the use of a condenser microphone, the AKG C419 (illustration #2). I chose the C419 because of its frequency response (20-20k Hz.) and because its design enables it to be clipped directly to the instrument, allowing me control of positioning which would remain the same even if the instrument is moved, unlike more conventional designs that require the use of a separate microphone stand.
The microphone can be used to feed the sound of the sitar into a sound system, into sound processing units and into a computer for processing or analysis.
I have also used a contact piezo microphone attached directly to the face of the sitar, which, even without the fidelity of the condenser microphone, has proved useful. In a recent concert a signal from the piezo microphone was sent to a computer to be processed by Mike Berry's GrainWave. The piezo allowed the use of nearly inaudible vibrations of the instrument to be amplified and processed. Currently I am investigating the use of piezo film which can be mounted under the feet of the sitar's bridges.
Early in the creation of the ESP system I investigated and experimented with making electromagnetic pick-ups for the sitar, it proved to be difficult, tedious and time consuming. Despite that, I plan to continue to investigate their possible use in the ESP.
The standard controller for using synthesizers is a keyboard, which is fine as a controller if you are a keyboard player. I have always dreamed of accessing the sonic potential of sound synthesis, sampling, and real time sound manipulation, using instrument technique that I have already acquired and developed. Sitar is my main instrument, to which I have devoted years of study. To be able to combine my musical interest in sound synthesis and manipulation with sitar technique seems a natural course.
As a performer in many musical situations I wanted to design an instrument that could enhance my musical contribution and also expand the possible playing situations in which I could use the sitar. The sitar, being a traditional instrument of North Indian classical music, brings with it techniques that are rooted in a particular style and concept of music making. I want to include these concepts and techniques into a larger and more personal definition of music.
The sitar can be tuned to virtually any system of tuning and is designed to be able to produce minute changes in intonation. One of the great benefits of using a pitch-to-MIDI converter is its ability to track the performed pitch which allows me to express this ability of the sitar with other forms of sound generation (synthesis, sampling).
The sitar is traditionally tuned so that all the scale pitches used in a raga are tuned in relation to one pitch (the tonic of the raga) or two pitches in combination (usually the tonic and the 5th). The process of tuning is learned from imitating one's teacher over a long period of time. This process includes the tuning of the strings as well as the adjustment of the frets, which are movable, and the required adjustment of finger position and pressure as the strings stretch and go out of tune in the normal course of playing. One eventually learns the proper tuning for each raga in one's repertoire.
There have been many theories of how the tuning in North Indian classical music is devised, including the classical theory of 22 microtones called srutis (Jairazbhoy 91). If this theory were correct a keyboard instrument could be tuned to perform all 22 microtones. Unfortunately these theories fall far short of the more complex reality.
To further complicate the issue, there is the use of andolan, a graceful bending between microtones that is required in the performance of certain ragas and the use of other ornaments such as meend, that require constant changes in string tension, to produce the pitch bending effect desired.
Because of the complexity of tuning and ornamentation in traditional sitar music it would be difficult or impossible to emulate it with conventional MIDI keyboard2 controllers.
The ability to include the sitar's tuning and intonational capabilities into the ESP allows me to experiment with expanding the use of sitar in traditional settings. I am often responsible for composing and performing music for Indian dance, including styles of dance from all over India (Kathak, Bharatnatyam, Odissi, Orissa, etc.). There is a great deal of experimentation going on in the contemporary traditional Indian dance world. Sometimes modern synthesizers are employed alongside traditional instruments. I think this kind of experimentation could benefit from the use of an instrument such as the ESP, an instrument that can express many aspects of the music tradition while allowing the use of greater sonic possibilities.
Much of my music, in collaboration with Loren Rush, Janis Mattox and Pauline Oliveros, has used systems of just intonation, including 5-limit and 7-limit tunings3 devised by Loren Rush and Alfred Owens (Rush, Mattox and Owens). The ESP is easily adaptable to such situations by virtue of its tuning capabilities.
![[opening page]](jpegs&gifs/opening.jpeg)