Abstract
Electro-acoustic music has been ignored by music theorists for years. One theory suggests that this deficiency is due to the lack of an objective visual representation of this type of music. This research is focused upon the audio analysis involved in the creation of a ‘‘pseudoscore’’ for electro-acoustic music. Computer analysis of music is a complicated task. The ‘‘holy grail’’ of automatic computer music analysis has most often been the translation of a continuous pressure variation into traditional Western musical notation. Since electro-acoustic music is rarely able to be transcribed into this type of notation, the automatic analysis of electro-acoustic music confounds many previous algorithms. A top-down approach is suggested in order to extract acoustic and musical information from recordings of electro-acoustics music. A top-down approach has already proven successful in extraction of musical tempo [E. Scheirer, M.I.T. Ph.D. thesis (2000)]. Since humans are able to translate a continuous pressure variation into useful acoustic information, when using a top-down approach to automatic music analysis, a cochlear model is used as a front end. Although this model apparently encodes some redundant information, cross- and auto-correlation techniques allow easier extraction of some acoustic information. This presentation will detail the current state of this research.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call