Articles to come...
Articles to come...
Articles to come...
Music is ineffable communication that permeates the history of mankind. Kokelaar and Lavy (2002, p. 328) discuss the creative process of writing music, and suggest that to date no cohesive explanation for it has been proffered. This article explores at a high level some of the issues in musical composition, especially as it relates to computer-assisted musical composition. In fact, the emphasis of the research and development of this web site from this time forward will be computer-assisted musical composition. This will take the form of articles which propose and explore theories on that topic, and programs which demonstrate ideas and theories.
The question that is probably on most readers’ minds is this: why involve computers in musical composition, just let the gifted artists create music. There are several answers to that. The first is that computers save time. It is no wonder that almost all composers and arrangers now use the Finale or Sibelius programs to notate music. These software applications save time because they do repetitive functions such as draw key signatures and bar lines. Copy and paste can save hours for repetitive accompaniments. In addition, Guthmann (2013, p. 43) points out that composition software allows composers to hear playbacks without relying on performers. Guthmann also goes on to say that composition software makes revisions easy, and thus the final product more to the composer likes.
The next question that arises in a discussion such as this is whether a computer itself can be creative. This question has been debated extensively over time. Hofstadter (1991, p. 378) suggests that creativity can be approximated in a computer. This can easily be seen and heard with a program such as Band-in-a-Box, which is published by PG Music, Inc. With a single click of a button, an entire song can be written and played back. At first blush, this is pretty impressive. After “writing” several songs with Band-in-a-Box, though, one realizes that the creativity that each song demonstrates is not to the level of today’s popular songs. In fact, the songs may seem to get repetitive and boring. Regardless of your perceptions, Band-in-a-Box is an incredible demonstration of what can be done by a computer to write music.
There have been some notable attempts to codify musical composition into algorithmic systems. The earliest was a series of suggestions made by Hindemith (1939) where he presents rules for how many steps and skips in a single direction is recommended for melodies. If you follow these suggestions, you can create some Hindemith-like melodies. This web site will feature a program in early 2014 that demonstrates creating melodies by following Hindemith’s suggestions. The article that you are reading will be updated with a link to the software application and the related article. Another significant contribution was made by Schillinger (1941) in which he presented a comprehensive system of musical composition. Some notable composers such as Gershwin and Bernstein studied this approach, and assimilated the techniques into their compositions. In addition, the 12 tone composers brought stringent composition rules to a new level in their rigid system of 12 tone sequences. While this may seem to remove any creativity that composers have, if you listen to Le marteau sans maître by Boulez, you quickly realize that 12 tone music can be beautiful and provocative.
This web site does not advocate complete compositions written by a computer. In spite of the successes of Band-in-a-Box, it seems that a hybrid of human creativity along with the tools which a computer offers is ideal. Humans can provide the creative spark, and computers can do some of the drudgery that is associated with musical composition. In this way, human creativity will be advanced. In the final analysis, a marriage of human creativity with computer tools will lead the way to even better compositions. This is computer-assisted musical composition.
As computer-assisted musical composition is explored, we will select the best approach to solving the problem at hand. Sometimes it will be following a rule set. Sometimes it will be using neural networks. It is important, though, to make sure that no one single approach to AI becomes predominant.