This paper deals with the problem of automatic recognition of music. Segments of digitized music are processed by means of a Continuous Wavelet Transform, properly chosen so as to match the spectral characteristics of the signal. In order to achieve a good time-scale representation of the signal components a novel wavelet has been designed suited to the musical signal features. particular care has been devoted towards an efficient implementation, which operates in the frequency domain, and includes proper segmentation and aliasing reduction techniques to make the analysis of long signals feasible. The method achieves very good performance in terms of both time and frequency selectivity, and can yield the estimate and the localization in time of both the fundamental frequency and the main harmonics of each tone. The analysis is used as a preprocessing step for a recognition algorithm, which we show to be almost independent on the instrument reproducing the sounds. Simulations are provided to demonstrate the effectiveness of the proposed method.

Instrument-independent analysis of music by means of the continuous wavelet transform / Olmo, Gabriella; Dovis, Fabio; Paolo, Benotto; Claudio, Calosso; Pierluigi, Passaro. - STAMPA. - (1999), pp. 716-726. ((Intervento presentato al convegno Wavelet Applications in Signal and Image Processing VII tenutosi a Denver, CO nel July, 1999 [10.1117/12.366827].

Instrument-independent analysis of music by means of the continuous wavelet transform

OLMO, Gabriella;DOVIS, Fabio;
1999

Abstract

This paper deals with the problem of automatic recognition of music. Segments of digitized music are processed by means of a Continuous Wavelet Transform, properly chosen so as to match the spectral characteristics of the signal. In order to achieve a good time-scale representation of the signal components a novel wavelet has been designed suited to the musical signal features. particular care has been devoted towards an efficient implementation, which operates in the frequency domain, and includes proper segmentation and aliasing reduction techniques to make the analysis of long signals feasible. The method achieves very good performance in terms of both time and frequency selectivity, and can yield the estimate and the localization in time of both the fundamental frequency and the main harmonics of each tone. The analysis is used as a preprocessing step for a recognition algorithm, which we show to be almost independent on the instrument reproducing the sounds. Simulations are provided to demonstrate the effectiveness of the proposed method.
9780819432995
File in questo prodotto:
File Dimensione Formato  
Instrument independent - SPIE 99.pdf

non disponibili

Tipologia: 2. Post-print / Author's Accepted Manuscript
Licenza: Non Pubblico - Accesso privato/ristretto
Dimensione 1.5 MB
Formato Adobe PDF
1.5 MB Adobe PDF   Visualizza/Apri   Richiedi una copia
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11583/2499929
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo