Computer music is the use of computing technology in music composition, helping human composers create new music or allowing computers to create music on their own, such as with programs that use rules to make music. It includes the study and use of new and existing computer software technologies and basic aspects of music, such as sound synthesis, digital signal processing, sound design, sonic diffusion, acoustics, electrical engineering, and psychoacoustics. The field of computer music has its beginnings in the early days of electronic music and the first experiments and innovations with electronic instruments at the start of the 20th century.
History
Much of the work on computer music has focused on the connection between music and mathematics, a connection that was noticed long ago by the Ancient Greeks, who talked about the "harmony of the spheres."
In 1950, the first computer-generated musical melodies were created using a computer called the CSIR Mark 1, later renamed CSIRAC, in Australia. Some newspapers in America and England claimed that computers might have played music earlier, but research has shown these claims to be false because there is no evidence to support them. People may have guessed that computers could play music because they made noises, but there is no proof that they actually did.
The first computer to play music was the CSIR Mark 1 (CSIRAC), designed and built by Trevor Pearcey and Maston Beard in the late 1940s. A mathematician named Geoff Hill programmed the CSIR Mark 1 to play popular songs in the early 1950s. In 1950, the CSIR Mark 1 was used to play music for the first time, marking the first known use of a digital computer for that purpose. The music was not recorded, but it has been reconstructed accurately. In 1951, the CSIR Mark 1 publicly played the "Colonel Bogey March," a version of which only exists in the reconstructed recording. However, the CSIR Mark 1 only played standard musical pieces and was not used to create new ideas in music composition, unlike the work of Max Mathews, who helped develop modern computer-music practices.
In late 1951, the first computer music performance in England was the British National Anthem, programmed by Christopher Strachey on the Ferranti Mark 1. Later that year, short parts of three songs were recorded by the BBC: the National Anthem, "Baa, Baa, Black Sheep," and "In the Mood." This recording is considered the earliest known recording of a computer playing music, as the CSIRAC music was never recorded. The recording is available online at the University of Manchester website. In 2016, researchers at the University of Canterbury, Christchurch cleaned up and restored this recording, and the improved version can be heard on SoundCloud.
Two major developments in the 1950s were the start of using computers to create digital sound and the creation of programs that compose music using rules, not just playing pre-written music. Lejaren Hiller and Leonard Isaacson, two musical chemists, worked on experiments with algorithmic composition from 1956 to 1959. Their work led to the 1957 performance of the Illiac Suite for String Quartet. Max Mathews, working at Bell Laboratories, created the influential MUSIC I program and its later versions, which helped spread the use of computers in music through an article in Science in 1963. The first professional composer to use digital synthesis was James Tenney, who created music using Max Mathews' MUSIC III system at Bell Labs, beginning with Analog #1 (Noise Study) in 1961. After Tenney left Bell Labs in 1964, he was replaced by Jean-Claude Risset, who studied how to create realistic musical sounds and composed Computer Suite from Little Boy in 1968.
Early computer-music programs often did not run in real time, but the first experiments on CSIRAC and the Ferranti Mark 1 did operate in real time. From the late 1950s, programs took hours or days to run on expensive computers to create only a few minutes of music. One solution was to use a "hybrid system" that combined digital control with analog synthesizers. Examples of this were Max Mathews' GROOVE system (1969) and MUSYS by Peter Zinovieff (1969).
Until now, some research has focused on studying the qualities and structure of sound. Examples include work by Hiller and Isaacson in Urbana, Illinois, USA; Iannis Xenakis in Paris, France; and Pietro Grossi in Florence, Italy.
In May 1967, the first experiments in computer music in Italy were conducted by the S2FM studio in Florence in partnership with General Electric Information Systems Italy. Pietro Grossi used the Olivetti-General Electric GE 115 computer as a performer. Three programs were prepared for these experiments, written by Ferruccio Zulian and used by Grossi to play works by Bach, Paganini, and Webern and to explore new sound structures.
John Chowning's research on FM synthesis from the 1960s to the 1970s made digital sound creation more efficient. This led to the development of the Yamaha DX7 digital synthesizer, released in 1983.
In Japan, computer music experiments began in 1962 when Keio University professor Sekine and Toshiba engineer Hayashi used the TOSBAC computer. This resulted in a piece called TOSBAC Suite, inspired by the Illiac Suite. Later Japanese works include a piece by Kenjiro Ezaki at Osaka Expo '70 and Panoramic Sonore (1974) by music critic Akimichi Takeda. Ezaki also wrote an article titled "Contemporary Music and Computers" in 1970. Since then, Japanese research in computer music has mostly focused on commercial uses in popular music, though some serious musicians used large systems like the Fairlight in the 1970s.
In the late 1970s, systems like the Roland MC-8 Microcomposer, which used a microprocessor to control an analog synthesizer, became available. The release of the Yamaha DX7, along with cheaper digital chips and microcomputers, allowed real-time computer music generation. In the 1980s, Japanese personal computers like the NEC PC-88 included FM synthesis sound chips and supported audio programming languages like Music Macro Language (MML) and MIDI interfaces, which were often used to create music for video games. By the early 1990s, microprocessor-based computers had advanced enough to allow real-time computer music creation using general-purpose programs and algorithms.
Advances
Improvements in computer power and software for handling digital media have changed the way music is created and performed using computers. Modern small computers are strong enough to produce complex sounds using many different methods. Computer music tools are now common in music creation, so much so that we rarely notice them: digital synthesizers, mixers, and effects are now standard in making and recording music, replacing older analog methods.
Research
There is a lot of activity in the field of computer music as researchers work to create new ways to make, write, and perform music using computers. Around the world, many groups and schools study and research computer and electronic music. Examples include CCRMA (Center for Computer Research in Music and Acoustics at Stanford, USA), ICMA (International Computer Music Association), C4DM (Centre for Digital Music), IRCAM, GRAME, SEAMUS (Society for Electroacoustic Music in the United States), CEC (Canadian Electroacoustic Community), and many universities and colleges globally.
Later, composers like Gottfried Michael Koenig and Iannis Xenakis used computers to create sounds and scores for their music. Koenig made computer programs that used math to help write music, based on his own method of composing. This is different from Xenakis’ work, as he used math to explore musical ideas. Koenig’s programs turned math calculations into codes that could be written as musical notes. These codes could be copied by hand and played by musicians. His programs, Project 1 and Project 2, are examples of this. Later, Koenig used similar ideas to let computers create sounds directly. A program called SSP does this. All of these programs were created by Koenig at the Institute of Sonology in Utrecht in the 1970s. In the 2000s, Andranik Tangian made a computer program to help plan the timing of rhythms in musical pieces called rhythmic canons and fugues. These were later turned into harmonic compositions, Eine kleine Mathmusik I and Eine kleine Mathmusik II, performed by a computer.
Computers have also been used to copy the style of famous composers like Mozart. One example is David Cope, who created programs that analyze other composers’ works to make new music in a similar style. His most famous program is called Emily Howell.
A research project called Melomics, from the University of Málaga in Spain, created a computer system named Iamus that writes complex, multi-instrument music for editing and performance. In 2012, Iamus made a full album with the same name, which New Scientist called "the first major work composed by a computer and performed by a full orchestra." The group also made a tool for others to use its technology and shares its music online.
Computer-aided algorithmic composition (CAAC, pronounced "sea-ack") is the use of computer programs that follow rules or steps to create music automatically. This term combines two older labels that were too vague. "Computer-aided composition" does not clearly describe the use of rules to create music, as programs that write music with notation or sequencing software could also be called computer-aided. "Algorithmic composition" is too broad because it does not mention the use of a computer. The term "computer-aided" is used in the same way as "computer-aided design."
Machine improvisation
Machine improvisation uses computer programs to create new music based on existing musical pieces. This is often done by combining musical phrases from existing music, whether performed live or recorded before. To create music that sounds like a specific style, these programs use machine learning and pattern recognition to study examples of existing music. The patterns found are then used to create new variations that match the style of the original music, a process called stylistic re-injection. This is different from other methods that use computer programs to create music without analyzing existing examples.
Style modeling involves creating a computer model that captures important musical features from data. Statistical methods are used to identify repeated patterns or structures, which are then combined to make new music. Style mixing can be achieved by analyzing a collection of music examples from different styles. Machine improvisation builds on earlier work in statistical modeling, starting with Hiller and Isaacson's Illiac Suite for String Quartet (1957) and Xenakis' use of Markov chains and stochastic processes. Modern methods include techniques like lossless data compression, prediction suffix trees, and string searching. Style mixing can be done by combining models from different musical sources, first demonstrated by S. Dubnov in the piece NTrope Suite using the Jensen-Shannon joint source model. Later, the factor oracle algorithm (a type of computer model built in linear time and space) was used in music by Assayag and Dubnov, forming the basis for systems that use stylistic re-injection.
The first use of statistical style modeling was the LZify method in the Open Music software. This was followed by the Continuator system, which allowed real-time machine improvisation by interpreting the LZ incremental parsing method through Markov models. This system was developed by François Pachet at Sony CSL Paris in 2002. A version of the Factor Oracle machine improvisation is also available in the Matlab Computer Audition toolbox, as well as in the NTCC software.
OMax is a software environment created by IRCAM. It uses OpenMusic and Max, and is based on research into stylistic modeling by Gerard Assayag and Shlomo Dubnov, as well as research on computer-assisted improvisation by G. Assayag, M. Chemillier, and G. Bloch (known as the OMax Brothers) in the Ircam Music Representations group. One challenge in using factor oracle models for audio signals is converting continuous data into a discrete format. This issue was solved with the Variable Markov Oracle (VMO), a Python-based tool that uses an information rate criteria to find the most effective representation.
Using artificial intelligence to create new melodies, copy existing music, and copy artists' voices has become a recent development that some say is affecting the music industry.
Live coding
Live coding, also called interactive programming, on-the-fly programming, or just-in-time programming, is the practice of writing computer programs during a live performance. Some people believe it offers a more structured way to create music compared to laptop musicians, who use computers to perform. Live coders sometimes think that laptop musicians may not have the same level of excitement and energy as traditional live musicians.