box of cards

8/14/2019 Box of Cards

1/26

Box of Cards:Computer musics birth into real-time interactions

By Jenifer Jaseau

MUS 605 History of Electronic Music

Dr. Jeffrey StoletUniversity of Oregon School of Music and Dance

Pre-MIDI paper

Submitted Winter 2010


2/26

Box of Cards Jaseau 2

Introduction

The development of computer music began with the innovations of Max Mathews

at Bell Telephone Laboratories. His work focused on the controlled generation of sound

within a computer, leading to the development of computer-based systems we use today

in computer-generated interactive music systems. These developments started with the

Music-Nseries, an early computer sound synthesis software program. The first efforts at

digital synthesis were non-real-time, and after many trials from composers and

modifications from technicians, there became a need for real-time application. The

GROOVE system came out of the Music-N series in response for that need, becoming a

template for live computer music systems today. These two structures of computer

architecture became the basis for all computer systems we use today, including Max/MSP

and Kyma.

The Beginning

Max Mathews began working at Bell Telephone Laboratories in Murray Hill,

New Jersey after graduating from M.I.T with a doctoral degree in electrical engineering.

In 1957, Mathews worked in the acoustic research department at Bell Laboratories

developing computer equipment to study telephones. His task was to create a listening

test for telephones in order to judge the quality of a sound received. He constructed a

device that would convert an analog representation of sound, enabling the sound to be

received by a computer in binary or electrical form. He then needed an additional device

to get the sound back out of the computer and into an analog representation of sound that

a listener could hear. In this way, he developed a way to store and manipulate sound


3/26


4/26


digits were needed to obtain samples, which could be easily done by adding more

switches and resistors to the system.2

ADC

The ADC was the equivalent of a DAC with the exception that it needed a

feedback mechanism and a programmer, which was a small computer. The ADC had

three essential components: the comparer, the programmer and the DAC.

The ADC would convert sound pressure level variations into a series of discrete

numbers by using a current-to-voltage amplifier in a sequence of steps; first, the analog

voltage would be applied to the analog input terminal, then the programmer would set its

calculations to zero for each of the five branches. After five cycles of five decisions from

the comparer, the five branches would output a digital equivalent of the analog input.

The ADC was n times slower than the DAC (n representing the number of digits.) It had

limitations from n sequential decisions involved in converting a single number.

Then there was the question of storage, an essential component of creating

computer music. Accessibility to recall digital data within the limits of the system was

another primary concern. It was essential to the quantization process to maintain a

steady sampling rate, for if there were any variations in the sampling rate, it would then

equate to fluctuations similar to wow and flutter in the analog domain. The numbers of

samples were greater than the magnetic core memory of a computer so the computer

would store samples in bulk. Samples were stored and retrieved in sequence by a digital

magnetic tape that would record in bulk or groups called records, and these records

2Mathews, Technology of Computer Music, 26


5/26


would not store continuous data. There became a need for a small core memory or buffer

to be inserted between the tape and the converter in order to maintain a constant sampling

rate. The computer provided an external data connection to deliver samples to a

converter under the control of an external oscillator digital tape transport associated with

a computer. Core memory equaled a buffer, and the functions of control circuits were

accomplished with a program. This allowed the same computer that synthesized the

sounds to then communicate directly in sound with the external world, being us humans.

Once Mathews discovered the ability to use computers as a way to generate

sound, the next logical step was to write programs that could play music through the use

of a computer. He made several versions of sound synthesis software that gradually got

better and better, gradually using less computation power and gaining more of a user-

friendly environment for the composers use. This software program is commonly referred

to as the Music-Nseries, developed at Bell Laboratories. The development of Music-N

opened up many areas and opportunities for other programmers and composers to make

sounds that were new, and make environments for electronic compositions.

It was immediately apparent that once we could get sound out of acomputer, we could write programs to play music on the computer. That

interested me a great deal. The computer was an unlimited instrument,and every sound that could be heard could be made this way. And the

other thing was that I liked music.3

Music-N

Music Iwas the first attempt to apply a computer to musical goals. Being the first

of the Music-N series, developed in 1957 on the IBM 704 computer4. The IBM 704 was

3Chadabe,Electric Sound, 108

4The very next model after IBMs first ever computer, the IBM 701


6/26


located at IBMs world headquarters in New York City. Mathews would run his music

program, get a digital magnetic tape, and then bring that tape back to Bell Labs to convert

it into sound using the DAC. Bell Labs were the only ones to have the right DAC

hooked to a digital tape transport that could play a computer tape. This first sound

generating computer program had only a single voice, one waveform, had the same rise

and decay characteristics, and only had the ability to change 3 expressive parameters;

pitch, loudness, and duration. The first ever piece composed digitally was a seventeen-

second piece by psychologist Newman Guttman entitledIn the Silver Scale. Mathews

said it was quite terrible, and the program was terrible, yet it was the first

5

. In later

Music-N programs, the sound synthesis software was written on an IBM 7094, unique

because it was the first computer to use transistors in addition to being very effective for

the work done in speech processing and visual signal processing. Some of the early

notable operation systems were developed for the IBM 7094 such as the Bellsys 1 and 2.6

Music IIfollowed shortly thereafter in 1958. Improvements such as an addition

of three voices, totaling four independent voices, had been made. Sixteen waveforms

were stored in memory thus introducing the wavetable oscillator, a new concept that

allowed a user to select from arbitrary waveforms such as sinusoidal, square and saw

tooth waves.

The wavetable oscillator works by circularly accessing the wavetable at multiples

of an increment and reading the wavetable content at that position. The wavetable is a

list, stored in memory, of the calculation for one cycle of the waveform. A computer

program achieved this by calculating the sample values of a waveform according to a

5Chadabe, 109

6Roads, Music Machine, 6


7/26


mathematical formula. After calculation, the samples are sent to the DAC. To generate a

periodic sound, the computer program read through the wavetable again and again,

sending the samples it read to the DAC in order to convert the waveform to sound.

Music IIIfollowed in 1960. Mathews wanted to create a set of universal building

blocks that gave a musician freedom and task to put blocks together into his or her

instrument. He thought if someone knew how to patch a Moog synthesizer, then they

could put together unit generators7

into a computer. This version introduced the concept

of modularity in addition of an array of orchestra of instruments to choose from. A score

was also implemented, listing notes in order according to starting times. Each note could

also be associated with a timbre in addition to the previous expressive parameters

mentioned before with Music I and II.

Word Got Out

Shortly after Music III, word got out there was a computer able to make music.

This discovery got many people excited, a few composers and a collection of folks from

Princeton University such as Hubert Howe, Godfrey Winham, and Jim Randall were

curious about the new development of computer based music and visited Mathews at Bell

Labs. But Bell Labs was not the only place that had a computer capable of making

music.

In 1961, John Pierce, director of Bell Labs and a crucial supporter of Mathews

work, heard of the pieceIlliac Suite for String Quartetand was interested in the work

7Software modules for signal processing. Languages based in the concept of unit

generators have been the foundation of most research in digital sound synthesis

algorithms to date. RoadsFoundations in Computer Music, 370


8/26


being done by the programmers and researchers there. Since Pierce did not know the

names of the researchers there, he addressed his letter to the Illiac computer at the

University of Illinois as Dear Computer, expressing his interest for the computer that

was composing music. The researcher at Illinois, Lejaren Hiller, then invited Pierce for a

visit; while there, Pierce met Harry Partch and graduate student, James Tenney.

Tenneys work so impressed Pierce that he hired Tenney as a resident composer to do

work at Bell Labs in computer music.

James Tenney was the first composer to take an interest in computer music

research. He joined Mathews in 1961 as a composer in residence in psychoacoustics and

computer music where he remained until 1964, becoming one of the first composers to

work extensively in the area of digital synthesis. During his time there, Tenney created a

series of tape pieces in addition to writing software that utilized a computer to make

complex task oriented musical functions in way that had not yet been explored in music.

Mathews software made music in a traditional way such as making a printed score that

was encoded on punch cards. Tenney had a different approach; he worked on realizing

compositional intelligence by creating a software program able to make musical

decisions. His interest in developing a computer as an instrument and vehicle to make

musical decisions was influenced by other fields he found fascinating like

psychoacoustics, Gestalt psychology (in relation to auditory perception), information

theory, stochastic music, and the theories of John Cage. Tenney created several works

while at Bell Labs such asAnalog #1: Noise Study (1961) inspired by traffic noises inside

the Holland Tunnel,Dialogue (1963) exploring parameters that were controlled

according to probability distribution,Phases (for Edgard Varese) (1963) a piece said not


9/26


to have been made by a computer or a human, but by a hybrid of the two, andErgodos II

(for John Cage) (1964) created as a series of parameters and probabilities that change

over time.

Music IV and Music V

In 1963, Mathews was joined by Joan Miller to work on Music IV, a little more

convenient to use, although not more musically powerful than the previous version, but

computationally more sophisticated due to the technical advances in macro assembly

program that had been recently invented. The same computer, an IBM 7094, was also

available at the nearby Princeton University, providing the people there access to run

Music IV easily. They made improvements in user-friendliness and called the version

Music IV-B in 1962.

John Chowning, then a graduate student at Stanford University, was one of the

first people to work with Music IV. He had read Mathews The Digital Computer as a

Musical Instrument article in Science magazine, and in 1964, went to meet Mathews at

Bell Labs. Chowning came home with Music IV and a box of cards. Chowning and

friend, David Poole, began working on Music IV, for they had an IBM 7094 in their

facility. The Artificial Intelligence laboratory (AI lab) had a PDP-1 used to convert

samples into sound. Later in 1966, the AI lab bought a PDP-6 and Chowning and Poole

wrote Music 10, a version of Music IV for the PDP-6.

In 1965 there was a great increase in the making of different computers, and this

called for a non-machine specific language program. Music Vwas written in the high

level language of FORTRAN allowing the program to be utilized on a wide range of


10/26


computers. Previously the Music N series were written on a low level, machine specific

assembly language. This meant that the program could not run on any other computer

than the one it was written on. The desired goal of Music V was to be a universal

program using FORTRAN as its complier. FORTRAN had already been implemented at

Princeton turning the IV-B series into IV-BF in 1967.

In 1968, Mathews had a brigade of people involved in the development of Music

V including Miller in addition to some new faces such as F. Richard Moore, Jean-Claude

Risset, who replaced Tenneys position as composer in residence, and Vladimir

Ussachevsky.

The fundamental concepts of Music V were the unit generators and thestored

functions. Music V was a modular system of software defined unit generators. It had an

OSC acting as a waveform generator, an AD2 as a 2 input adder, and RAN as a random

number generator, all of which could be linked together to create a new instrument.

Instruments could than be widely varied in complexity and variety by means of

connecting to unit generators in a variety of ways. The stored functions became the

score; analogous to a traditional musical score except that this score specified all acoustic

properties of each instrument including information about the discrete sound called notes.

Each note contained information about an instrument definition, an algorithm for the

instruments unit generators, a starting time and the duration of the note.

The program was simple in the amount of coding required to put it on new and

different computers. It was efficient and ran rapidly. It was the means to spread

computer music into the world. The first main users of these music programs were

Guttman, Pierce and Mathews who were all fundamentally scientists. Mathews wanted


11/26


musicians to try the system and see if they could learn the computer language and express

themselves with it. Mathews and Pierce approached several composers, including

Copeland who denied their inquiry, hoping someone would be willing to embark on the

journey of composing with a computer. Some composers replied and Music V was sent

out to them with two boxes of punch cards and a note saying Good Luck!

The program required a high knowledge of acoustics and computer techniques,

and all compositions were disseminated as recordings on tape because the computers

were not able to be carried into concert halls being that they were far to large. What was

the attraction of using a mainframe computer to compose music? How did composers

use computers to composer? Chowning reflects, Generality was a great part of the

attraction, meaning that any sound-generating algorithm could be tried, tested, tweaked

and perfected in software.8

This was true for all software synthesis approaches.

Each composer used Music V in a different way. Tenney was interested in

continuing to develop timbres and the use random noises of various sorts. After Tenney

left, Risset was invited from France to work as a composer in residence, on his thesis,

analyzing the timbre of a trumpets sound. His work lead to new techniques for analysis

becoming analysis for synthesis, the most powerful tool for analyzing natural music

sounds.9

Music V had progressed to be a flexible sound synthesis program. The only rigid

aspect was the input process, which lead to the need for a graphical system of display for

interaction. Mathews desired to broaden and make easy the techniques for specifying

compositions. He felt that the Music V language was more important than the program

8Chadabe, 127

9The Music Machine, 8


12/26


13/26


was called cmusic10. By the end of the 1970s most of the work in computer music had

been started, worked on and was close to perfection, due to contributions from people at

Stanford, UCSD, IRCAM, and MIT.

In a later interview with Mathews done by Curtis Roads in 198011

, Mathews

spoke about the early reactions people had to computer music. Mathews said that people

were skeptic, had fear and a lack of comprehension for computer music. Composers

were the most interested, and performers were not interested at all in the new

developments. By 1980, many computer music centers had been established and

composers recognized that computer literacy was important to their compositional

training. The computer programs were used tospecify a composition in order to create it.

Box of Cards

Music V was a multi-pass program; input specifications were processed in stages

of calculations, meaning that you had to wait before ever hearing a sound. Music V

worked in several steps; first, sound was generated by the main computer which would

accumulate an inaudible table of numbers known as samples, and store those samples as

fluctuations of voltages on a magnetic computer tape. Then in a separate step, these

samples were converted to sound and stored on a normal audiotape. People from

Princeton University would drive to Bell Labs to convert their paper cards onto a

magnetic audiotape. It was primarily Moores job to convert these samples to sound.

This process would take about two weeks. It was common to be surprised at the end

result of the tape, for the sound did not sound as they had expected. Batches were sent in

10An acoustic compiler program

11Music Machine, Roads 8


14/26


30-second segments because the composers had to be sure they worked. Then the

process would start all over again, using the computer program to punch audio

information onto the paper cards, and then send the cards to Bell Labs to be converted in

hope that this time, the sound would be representative of what the composer had tried to

specify. Barry Vercoe who was working at Princeton in 1968 and would continually drive

to Bell Labs to get bits of tape converted.

It was the only working converter and it was a long trek one had to go to BellLabs to convert the sound and drive through this dreadful traffic, and you could

only play it when you got back to Princeton and think, My God, thats not whatI wanted at all12

This angst of driving, and waiting to hear the result of the box of cards really

increased the desire for a real-time performance capability. Although the alterations made

great strides when compared to the Music I, it still lacked some intelligence. The

program was not intuitive in any way, and was not able to operate in real-time. In 1967,

Mathews and Moore began developing the GROOVE system, obtaining real-time editing,

control and the performance of musical scores by 1970. Commercial hybrid synthesizers

in the 1970s incorporated microcomputers and organ-type keyboards that had a

keyboard scanner program to record information as it was played and relayed that

information to the synthesizer13

. Although many attempts were made, such as the Illiac

computer in 1967 at the University of Illinois, or the PIPER 1 from the University of

Toronto in 1965, it was again Max Mathews that pioneered the first fully developed

hybrid system in the USA.

12Chadabe, 113

13Computer Music Tutorial, Roads, 614


15/26


GROOVE

The GROOVE system stands for Generated Real-Time Operations on Voltage

Controlled Equipment. It was a hybrid system meaning that is was a digital computer

used to control an analog synthesizer. F. Richard Moore who worked with the

programming aspect of the GROOVE system assisted Mathews in developing this

system. This program was linked to a real-time generation system that allowed a

composer to be in direct contact with the processes of digital synthesis. The great

advantage of the GROOVE system was that is had real-time editing, control and a

performance of musical scores. It had many input devices, such as a joystick and a knob,

for conducting a score that was prepared beforehand and had input into the computer.

The software language was intended to be easily understood using standard

concepts such as time, notes, duration, and velocity as building blocks. These building

blocks could then be put together to create a score, which was a list of defined and

undefined parameters that would change over time either from instructions from the

score, or from data coming from the many sensors.

The GROOVE system used a mini-computer connected to an analog sound

synthesis system. The computer in use was a Honeywell DDP-224 minicomputer and

attached to this computer was a large auxiliary disk drive, a digital tape drive, an

interface for the analog device that incorporated twelve 8-bit and two 12-bit DACs, and

sixteen relays for switching functions. Information was updated every one-hundredth of

a second. There was also an additional set of converters that provided the horizontal and

vertical coordinates for a cathode ray display unit that displayed visual representations of

the instructions for control provided by the composer.


16/26


It contained a graphical monitoring system that looked at a score as a recording of

control functions for an analog synthesizer. Functions of time specified the way

frequencies and other parameters that one controls in an analog synthesizer. The

parameters would change over time to make music. This graphical monitoring system

had software that generated a linear time-scale on the horizontal axis. The time span was

ten seconds and acted as a timing block or page. The screen would automatically clear

itself at the end of the page to display the next page. Ten different functions could be

displayed without crowding. The use of the monitoring system linked with the real-time

generation system allowed for the intimate connection of digital synthesis between the

composer and the computer.

Commands were entered into the computer by means of a conventional typewriter

terminal using a mnemonic code that was translated by the computer into sequential

controls values for the analog device interfaces. Since air pressure and air molecules vary

over time to give a listener the perception of sound, so must a computer-generated sound

change over time. It is not just a commandthat changes a sound, but agesture from a

human that occurs over a time span that allows music to breathe as if coming from the

natural setting of sound occurring through fluctuations of sound pressure.

There were support input devices that provided varying device functions during

the performance of a computer score. The computer score contained a list of control data

that changed over time, and allowed specified parameters to be modified in real-time.

These support devices consisted of a 24-note keyboard (now MIDI triggers), four rotary

knobs, and a 3D joystick. The voltage output from the knobs and joystick were


17/26


multiplexed14

to a single ADC so as to be recognized by the computer. The keyboard

was connected directly to a 24-bit binary register allowing each key to control the state of

a bit. The output of control data was regulated with a variable-rate clock pulse generator.

The computer acted as a control device in hybrid systems and not as a direct source of

audio signals. Adjustments to the clock rate would vary the rate of change in the events,

not vary the events themselves. Therefore a performance could be halted at any given

time.

The analog section of the system contained twelve voltage-controlled oscillators

(VCO), seven voltage-controlled amplifiers (VCA), two voltage-controlled filters (VCF),

and a variety of signal processing functions selected from an array of seventy-two

different circuits mounted on plug in cards. Fifty fine-resolution potentiometers provided

the manual specification of the basic controls associated with the circuits. All

interconnections for the whole system were routed via a five hundred element central

patch field.

The synthesizer had devices that were built from components that were laying

around or they would build one. It had a patching system of patch boards that plug into a

holder allowing users to change connections rapidly. It also had a display system that

showed a subset of 14 lines of control information as a cursor went across the monitor.

This made for continuous control of parameters as opposed to event based control, so

slightly varying vibrato became something varying in time.

The digital sequencers made it possible to use a computer to control an analog

14A transmission system that carries two or more individual channels over a single

communication path.


18/26


synthesizer by storing its performance information into memory in the form of functions

of time for each synthesis parameter. The functions could be edited to change the

performance.

The GROOVE system served as an instrument, although initially intended for

research applications. It recorded time functions in sampled form at 100-200 Hertz (Hz),

storing this information on a disk. The program played back the functions and combined

them with other functions of time generated by a performer playing on thesensors of the

instrument. These combined functions were used to control the analog synthesizer. The

program had editing facilities to change stored functions; change could me made to one

sample of one function without affecting anything else. One could then get a printout of

the functions, allowing musicians to edit their improvisation. The functions were

displayed on a scope that could move to a particular sample in the function. One could

then hear a sustained sound, flip the editing switch on, change the value of the function

and hear what it was doing to the sound.15

The data being edited was the stored time

function on disk instead of the live data coming into the computer from the various

sensors. One could finally edit with real-time feedback.

People of GROOVE

A few individuals chose to work with the GROOVE system such as Laurie

Spiegel and Emmanuel Ghent each working in different ways. Spiegel begun working in

1973 and was interested in the real-time interaction process and compositional logic. She

made several pieces such asAppalacian Grove (1974),Drums (1975) and The Expanding

15Music Machine, 8


19/26


Universe (1975) among many others. Drums was composed by using a bank of resonant

(high-Q) filters, built by Mathews, which oscillate if pulsed. She sent the amplified sound

of literally turning a digital bit on and off through the filters16

. In other compositions she

used analog and digital input and output, hardware modules, library-resident code

modules that had been written by previous users, and gizmos built by Max Mathews.17 In

1974 she wrote software to synchronously compose music and animated video and called

it VAMPIRE.18

As she puts it I had begun to conceive of music not specifically as a

sonic art, but as the art of composing abstract pattern of change within time. (Spiegel,

191) This program allowed for control of sound based on the GROOVE system, with an

addition on a drawing program that did not yet exist. The VAMPIRE died in 1979 when

its CORE was dismantled, the digital equivalent of having a stake driven through its

heart. (191)

In a personal correspondence with Spiegel, she recounts the difficulties of being

one of the first people to work with computers as a way to generate and control sound.

Back then it was incredible amounts of work to get the simplest things to happenmusically and it really generally sounded pretty awful compared to today's refined

highly-controllable sounds. There was a lot of excitement but a lot of frustrationtoo. People gave the few of us who were into computers for doing music a lot of

grief and flack because computers were still viewed by the general public(including just about everyone in the arts) as "dehumanizing" - cold clinical, to be

feared, in general inheriting the characteristics of those who tended to owncomputers. Access was extremely limited. And at Bell Labs we had to pretend not

to be doing music and were always afraid of getting caught because it was notreally permitted under the "regulated monopoly" status that BTL had up until they

broke the company up. But you would have love the sense of discovering soundsand ways of doing music that were altogether new, even if it did sometimes take a

whole 6 months work to get something like a computer-controlled reverb to work,including not knowing if it ever would or not.

16correspondence with Spiegel

17Chadabe, 158

18Video And Music Program for Interactive Real-time Exploration


20/26


The most notable person working with the GROOVE system was Ghent. He was

already interested in rhythmic and tempo relationships and interacting with algorithms

while composing. Ghent was looking for a device that could represent any rhythmic

and/or tempo relationships.

I remember many occasions working with Laurie all night at this and thinkingwhat the computer was doing was incredible. It would produce lines that were

musically so interesting, but who ever would have thought of writing a musicalline like that? We had the sense that here we had hired the computer as a musical

assistant and it was producing something that we never would have dreamed of.19

His most notable piece wasPhosphones (1971) being the most interesting because

the Mimi Gerrard Dance Company used the piece for a performance, in addition to

Gerrard and her husband James Seawright developing a lighting system for synchronized

theatrical lighting that was based off the technology of the GROOVE system.

This lighting system was called CORTLI20

and was funded by a grant from the

NEA. The piece required complex and rapid stroboscopic changes in lighting. The

realization of this polyphony of music and lighting in relation to the dance is extremely

dramatic. As the placement, intensity, and color of the lighting are precisely programmed,

and change with great rapidity in relation to the music, a subtlety of interaction results

which had not been possible before.21

The sound source forPhosphones was almost entirely comprised to a special

group of resonator circuits dubbedResons designed by Mathews for use in his electronic

violin. When tuned and adjusted to ring when dampened, the produce the array of

percussive sounds that characterize the piece.

19Chadabe, 162

20Acronym for Composing and Outputting Real Time Lighting Information

21Computer Music JournalVol. 32 no. 4 Winter 2008


21/26


In a personal correspondence with Moore in regards to his presence at the first

showing ofPhosphones he wrote: Nothing quite like what had ever happened before.

Actually, nothing quite like it has ever happened since. Gerrard reflects in the

difference between the original performance and one that was hep 30 years later, The

initial reaction ranged from amazement to hostility. Thirty years later the piece was

almost universally praised.

In 1977, Ghent begun a series of twenty-nine studies calledProgram Musicbased

on algorithmic modes of interaction. They were studies for what was to come22

This

series never got a chance to grow for Ghent received a Bell Labs notification that the

DDP-224 computer was to be removed from service thus ending the life of the GROOVE

system. Mathews did not recreate the program for other computers because by that time,

he thought people should use a digital sequencer and synthesizer.

Computer Architecture

In the bookFoundations of Computer Music, Roads questions what architecture is

ideal for digital audio synthesis. For this there is no single answer, yet there are four

principle design strategies for influencing architecture.

Modularized systems question how communication occurs between individual

modules. Kyma is a great example for how to get around the issue of communication

between modules by creating a group of modules that are connected by lines on a

graphical display.

22Chadabe, 163


22/26


The collection of suitable primitive elements, such as memory, shift registers, and

arithmetic-logic units questions how they are controlled through microprogramming.

Max/MSP allows a user to define their own parameters for how devices can change a

music element over time by programming their own patch. This leads to a variety of

success and errors in microprogramming depending on the skill of the user and the

purpose of the patch.

The pipeline approach is a great advantage because of the synthesis power

achieved with a small amount of hardware, yet it is difficult due to determining where in

the datum lays within the pipeline in addition to it being difficult to re-patch an

instrument on the fly. The way around these difficulties is to flush the pipeline after

every sample, which is a lot of work for a processor to do and therefore not a great

approach.

Lastly, the general-purpose computer architectures, be they micro or mini. Direct

digital synthesis was performed with a 16-bit or a 8-bit microprocessor until the

appearance of the 32-bit microprocessor with built in floating point instructions. This

was a great advantage because the amount of specificity between 0. and 1. is almost

limitless thus allowing more finite specification of musical parameters that can slowly

and accurately modulate over time in small and precise ways. This is useful for

spacialization between many speakers, or the control of an LFO over an oscillator,

amplifier or generator.

In many languages, these activities can be substituted for variables, also known as

functions in the language of FORTRAN and BASIC. Roads recommends this capability


23/26


because it seems to mirror something about the way we think and therefore make a

language more natural (513).

A system that grew out of GROOVE was PLAY, the first software sequencer,

built in 1977 at the Electronic Music Studio at State University of New York at Albany

by Joel Chadabe and Roger Meyers. Chadabe and Myers based the PLAY system on a

process model of music with functions and timers similar to the modules derived from

analog synthesis. PLAY was developed from the knowledge of GROOVE, the

Conductor program, and the Buchla Series 500 electric music box and used the PDP-11

computer. It maximized composers flexibility in conceptualizing the temporal process.

It used a small portable computer that controlled an external synthesizer in real-time with

the capability of interaction. It had two stages, the first is design, and the second is

operation.

In the design phase, the composer designs a specific compositional process using

any modules. The design phase uses 3 steps; first, a composer specifies data generators

such as a list of numbers, or a random number generator, or gets numbers in real-time

from some type of device, then determines how the data generators will affect the

attributes of sound such as pitch, rhythm, envelope shape and loudness. Second, the data

generators are organized as modules that are interconnected. Third, the composer sets the

rate of system clock and timing for each individual module. The typical operating

frequency for the system clock ranged from 20 Hz to 2 kHz.

In the operation stage, the composers process plays back and the composer

interacts with the playback according to the design. The composer will control, in real

time, the variables designed to be controlled or tuned, allowing real-time adjustment.


24/26


Significant changes could be made in the modules interconnections without any

discontinuity in playback.

The three fundamental concepts derived from the PLAY system that we still use

today are: function, such as specifying a data list; patch, defining modules; and playback;

setting a clock and letting it play.

GROOVE into Now

The importance of the GROOVE system lies in the fact that it was the first truly

interactive system comprised of many devices such as modules, sensors and a graphical

monitor. The modules allowed a user to define their own sound digitally from an analog

synthesizer. One could control musical parameters over time by moving a sensor, or

adjusting the sensors position over time. Nowadays sensors include infrared sensors,

game controllers, graphical tablets, accelerometers, a computer mouse, and basically

anything that sends out data. The use of sensors and modules that could change over

time made composing music with a computer less like using a machine by becoming

responsive to human gestures, but also lead to difficulties. These difficulties were the

problems faced by early composers and technicians working with these systems.

The systems built in Max/MSP are very similar to the system developed by the

Music-N series with the addition of GROOVE developments of real-time performance

capabilities.

GROOVE founded the system of design and playback we use today. One could

design a score and then play it back while modulating some specific parameters of the

analog synthesizer. The difference between then and now is then the storage medium


25/26


was paper cards and now the storage medium is a combination of a buffer, the random-

access-memory (RAM), and the computers hard disk.

GROOVE is in disguise today, the use of sensors to change a score over time, and

the ability to go back and manipulate what parameters was changed during an

improvisation. Kyma is an excellent example of the GROOVE in disguise. Developed

by Carla Scaletti and Kurt Hebel at the University of Illinois, Kyma is made up of many

objects that reflect the GROOVEs modular system. Instead of using hardware modules,

Kyma uses software objects. These objects can be placed together and connected using a

virtual patch cord, connecting all the module objects in various ways. An object can be

played back by itself, or in combination with other objects. This is the first step of

designing a real-time sound synthesis piece. The second step after defining all the

objects is to place these objects on a timeline. The timeline will then play back the

parameters of control for each of the modules. A user can choose to set some of the

objects used or to modulate the objects over time either with some kind of sensor, or by

using an envelop as a function of time that will change defined parameters by following

the value of the line set by the composer. This is fundamentally the same structure that

the GROOVE system used, a sequence of events that becomes the predetermined score,

and an array of devices to change the content within a specified parameter thus becoming

real-time editing, control and performance of musical scores.


26/26


Conclusion

Most computer music systems today are comprised from the basic functions and

developments of the Music-N series and the GROOVE system. Mathews and his

colleagues formed the common foundation for what a composer can design and monitor.

The basic concept of input and output is still the fundamental seed for the designing of a

computer music system, yet the path between input and output can form many different

journeys, which composers can freely design and manipulate. With systems such as

Max/MSP and Kyma, computer music today presents a composer with a world of

limitless possibilities for sound reproduction, synthesis, and the ability to control it in

real-time. Mathews, still alive today, must be reveling in the astonishment that he began

forging this path over 50 years ago.

It opened opportunities that had been unthinkableit enabled me to try all kinds of

ideas, listen to them in real time, modify them in real time, and thereby get a chance to

experiment in ways that would be prohibitive using standard methods like paper and

pencil and human musicians. -Emmanuel Ghent23

23Chadabe, 163

box of cards

Documents