UNESP - Universidade Estadual Paulista
Faculdade de Ciências - Departamento de
Computação
Bauru - São
Paulo
Brasil
jfmdi
ufpe
br
USP - Universidade de São Paulo
ICMSC - Instituto de Ciências Matemáticas de São
Carlos
Departamento de Computação
São Carlos - São Paulo
Brasil
edsmoreiicmsc
sc
usp
br
Signals can be sampled, stored and played back by digital computers. Furthermore, signals can be analized in such a way that compressed forms of their representation can be extracted, reducing the amount of memory needed to store them. The Fourier transform have been traditionally used in this way. This paper deals with the use of the Karhunen-Loève transform as an alternative method of representing a particular and dynamic signal, the musical tones, allowing bigger SAVINGS Of computing resources, as compared to the Fourier transform.