Analysis/Synthesis of the Andean Quena via Harmonic Band Wavelet Transform

Aldo Díaz; Rafael Mendes
DAFx-2015 - Trondheim
It is well known that one of the challenges in musical instruments analysis is to obtain relevant signal characteristics and information for sound description and classification. In this paper we study the Peruvian quena flute by means of the Harmonic Band Wavelet Transform (HBWT), a convenient representation for the sound content based on its 1/f fractal characteristics. In order to identify a relationship between fractal characteristics of musical sounds, we developed two sound transformations and establish a comparison between quena, a recorder and melodica wind instruments. The sound transformations implemented were noise filtering and pitch-shifting while the sound classification was focused on the γp fractal attribute. Our work led us to the conclusion that the HBWT quena representation favored the implementation of sound transformations and that the γp fractal feature had great potential in musical instruments recognition and classification applications. Keywords: sound fractal analysis; quena; 1/f noise; noise filtering; pitch-shifting.
Download