Title :
Number theoretical means of resolving a mixture of several harmonic sounds
Author_Institution :
Signal Process. Lab., Tampere Univ. of Technol., Tampere, Finland
Abstract :
In this paper, a number theoretical method is developed for the purpose of analyzing the spectre of a mixture of harmonic sounds. The method is based on the properties of prime numbers and on non-linear filtering. It is shown that a number theoretical approach is of vital importance in order to detect and observe harmonic sounds in musical polyphonies. The method is verified by applying it to the automatic transcription of piano music.
Keywords :
acoustic signal processing; harmonic analysis; music; nonlinear filters; number theory; signal detection; signal resolution; automatic transcription; harmonic sounds; musical polyphonies; nonlinear filtering; number theoretical method; piano music; Harmonic analysis; Hidden Markov models; Mathematical model; Probability; Robustness; Signal processing algorithms; Signal resolution;
Conference_Titel :
Signal Processing Conference (EUSIPCO 1998), 9th European
Conference_Location :
Rhodes
Print_ISBN :
978-960-7620-06-4