Change search
ReferencesLink to record
Permanent link

Direct link
Musicians performance prosody
KTH, Superseded Departments, Speech, Music and Hearing.
2004 (English)In: Proceedings of Reading Symposium Music Language and Human Evolution, 2004Conference paper (Refereed)
Abstract [en]

Music and speech are specific to humans. In our time we have many opportunities to hear music that is interpreted and executed by machines rather than by living musicians. Such examples mostly sound quite pathologic, particularly for music from the classical western repertoire. This demonstrates the relevance of the performance to the musical listening experience. For many years a research group at the department of Speech Music Hearing, KTH has studied the reasons for the computer's shortcomings as a musician. Our method has mainly analysis-by-synthesis, i.e., we have the computer play music files on a synthesizer. A professional musician, the late violinist Lars Frydén, assessed the emerging performances and recommended how they could be improved. We implemented his recommendations as performance rules in the control program and then tested them on various music examples. After many years of such experiments we had a dozen or two performance rules. These performance rules significantly contribute to improving performance, and the reason for this is an interesting question. The rules seem to be of three types. One type, the grouping rules, serves the purpose of grouping, i.e., showing where the structural boundaries are in the composition and which tones belong together. Another type enhances the difference between musical categories such as note values or scale tones or intervals, e.g., by increasing the dissimilarities between them. A third type adds emphasis to unexpected tones and deemphasizes expected tones. It is thought-provoking that the principles of grouping, category enhancement and emphasis of the unexpected are not specific to music. They occur also in other types of communication, such as speech, architecture, and others. This suggests that they emerge from demands raised by the receiving system. For example it is tempting to speculate that emphasis by delayed arrival, common both in music and speech, and delaying the emphasised information somewhat, is appropriate because it allows the neural system to finish processing the old information before it starts with processing the emphasized and hence particularly important information. In any event it seems likely that music performance as well as speech is tailored to the human cognitive system and that a comparative study of these two examples of systematic interhuman communication by acoustic signals will contribute to the understanding of human perception and cognition.

Place, publisher, year, edition, pages
Keyword [en]
Synthesis, performance, emphasis, cognition, perception
National Category
Computer Science Language Technology (Computational Linguistics)
URN: urn:nbn:se:kth:diva-51841OAI: diva2:465134
Reading Symposium Music Language and Human Evolution. Univ of Reading, U.K. September 28th - October 1st 2004
tmh_import_11_12_14. QC 20111220Available from: 2011-12-14 Created: 2011-12-14 Last updated: 2011-12-20Bibliographically approved

Open Access in DiVA

No full text

Search in DiVA

By author/editor
Sundberg, Johan
By organisation
Speech, Music and Hearing
Computer ScienceLanguage Technology (Computational Linguistics)

Search outside of DiVA

GoogleGoogle Scholar
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 25 hits
ReferencesLink to record
Permanent link

Direct link