In the paper an attempt to match musical instruments in terms of timbral features of their
sound is made for a collection of 53 high quality concert violins. The starting point of the
analysis is calculation of a set of features based on harmonic analysis, e.g. odd-to-even harmonics
amplitudes ratio, intensity of the first and higher harmonics etc., and a set of linguistic
descriptors of violin timbre related to these features. The semantically disjoint categories of
timbre characteristics are considered. The result of the analysis is the allocation of instruments
to those semantic categories with the expectation of the discovery of their similarities in individual
timbre dimensions. Although evident matching did not occur, certain possibilities
of uncovering expert preferences have been noticed. The outcome of the research provides
supportive cue for the design of a method of inferring preference models from the objective
characteristics of musical sounds.
[1]
Bozena Kostek.
Soft Computing in Acoustics: Applications of Neural Networks, Fuzzy Logic and Rough Sets to Musical Acoustics
,
1999
.
[2]
Ewa Lukasik,et al.
Inducing Jury's Preferences in Terms of Acoustic Features of Violin Sounds
,
2004,
ICAISC.
[3]
Bozena Kostek.
Perception-Based Data Processing in Acoustics: Applications to Music Information Retrieval and Psychophysiology
,
2005,
Studies in Computational Intelligence.
[4]
Witold Pedrycz,et al.
Supervised Semantic Scene Classification Based on Low-Level Clustering and Relevance Feedback
,
2004,
EWIMT.