David Cope: Virtual Mozart - Experiments in Musical Intelligence
CD CRC 2452, 1999, Centaur Records, 136 St. Joseph Street, Baton Rouge, LA 70802; telephone (225) 336-4877; fax (225) 336-9678; World Wide Web http://www.centaurrecords.com/
Reviewed by Nico Schüler
Already since the early 1980s, Professor David Cope, who teaches at the University of California - Santa Cruz, started to develop his LISP-computer system "Experiments in Musical Intelligence" (EMI), which combines analysis and composition processes. (His system was extensively described in his two books Computers and Musical Style [Madison, Wisconsin: A-R Editions, 1991] and Experiments in Musical Intelligence [Wisconsin: A-R Editions, 1996], as well as in numerous articles.) His goal has been to write music in a specific style. Mr. Cope's analyses are based on hierarchical analysis, drawing on Schenkerian analysis and on Chomsky's generative grammar of natural languages. Cope's EMI as well as his "Simple Analytic Recombinancy Algorithm" (SARA) can analyze each component of a composition for its hierarchical musical function, match patterns for "signals" of a certain composer's style, and reassemble the parts sensitively, using techniques drawing on natural language processing. Part of the analysis process involves a pattern searching algorithm that, in contrast to pattern-searching algorithms by other authors, seek patterns without any preconceived notion of their content. That means, the analyst does not need to know which patterns are supposed to be matched. Mr. Cope himself wrote: "EMI employs a limited set of variables called controllers, which affix musical parameters to vague outlines within which patterns are accepted as viably recognizable." (Cope in his monograph Experiments in Musical Intelligence, p. 36.) All analytical results, stored in object system files, can be recombined in the eventual composition process. To ensure a logical order of the recombination process, Mr. Cope programmed "augmented transition networks." (These augmented transition networks have already been successfully used since the 1970s for automatic language generation.) Many compositions generated on the basis of analytical results are proof of Cope's success. Some of these compositions, in the style of Palestrina, Bach, Mozart, Chopin, Brahms, Joplin, Stravinsky, Gershwin, Prokofiev, etc., were published on CDs at Centaur Records (Bach by Design, CRC 2184; Classical Music Composed by Computer, CRC 2329).
His newest CD, Virtual Mozart, contains a Symphony and a Concerto for piano and orchestra in the style of Wolfgang Amadeus Mozart, both "composed" in 1995. Sixteen Mozart concerti were used as models for the "virtual" piano concerto, whereby Mozart's symphonies nos. 6 through 31 were used as a database for the symphony. The musical forms were given with the three-movement form for the concerto and the four-movement form for the symphony. The motivic material, in both compositions, is - mostly - strikingly Mozartean, similarly the harmonic progressions as well as modulatory schemata. Only occasionally, the experienced listener might be taken aback at passages that seem not quite to fit in these otherwise astonishingly "original" Mozart compositions. Interestingly, as Mr. Cope mentioned in the CD cover, the symphony sounds indeed, at times, like later Mozart symphonies, even though symphonies beyond no. 31 were purposely avoided in the database. This phenomenon is certainly a question of "musical intelligence": Is it an example of pre-determined style characteristics of the late Mozart found in his earlier symphonies?
The excellent performance certainly added to the "originality" of the compositions: Only historically authentic instruments were used. The orchestra, directed by the UC Santa Cruz professor Nicole A. Paiement, consisted of two flutes, two oboes, two clarinets, two bassoons, timpani, each two violin I and II, two violas, two celli, and one bass. Even the piano (in the concerto), played by UC Santa Cruz professor Linda Burman-Hall, was a historical instrument from ca. 1800. However, the limitations in the availability of historical instruments were responsible for a re-orchestration: the incorporation of the brass parts into those for the woodwinds.
One CD - two more examples of Mr. Cope's pioneering work in computer-assisted simulation of specific musical styles. And while Cope's books on Experiments in Musical Intelligence have been praised in numerous reviews, his CDs surely demonstrate the success of his computer system. And this is one more reason to look forward to Mr. Cope's forthcoming book, The Algorithmic Composer (at A-R Editions), that will complete his trilogy of books on computer-assisted composition.
* * *
This review was originally published as: Nico Schüler: "David Cope: Virtual Mozart. Experiments in Musical Intelligence," Computer Music Journal 24/4 (Winter 2000): 80-81.
* * *
last revised: 3-20-2003