We propose an original technique for separating the spectrum of the noisy residual component from that of the harmonic, quasideterministic one, and to estimate the envelope of the residual, for the spectrum modeling of musical sounds. The algorithm for spectrum separation relies on nonlinear transformations of the amplitude spectrum of the sampled signal (obtained via FFT), which allow to eliminate the dominant partials without the need for precisely tuned notch filters. The envelope estimation is performed by calculating the energy of the signal in the frequency domain, over a sliding time window. Eventually the residual can be obtained by combining its spectrum and envelope, so that separate processing can be performed on the two.

Download

In this paper, an analysis of flute attacks processed by a model of the physiological auditory system is presented. In flute performance, the musician uses consonants (/d/, /g/, /k/, /t/ and /p/) in order to create particular effects as hard and soft attacks. These effects are very important in music interpretation. We found that the model discriminates the sounds very well and better than spectral analysis as worked out by standard methods. The model responses appeared very detailed and allowed attack classification by the application of very simple pattern recognition techniques.

Download