Second International Workshop on Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA 2001)

Florence, Italy
September 13-15, 2001

Noise Reduction for Vocal Pathologies

Lorenzo Matassini (1), Claudia Manfredi (2)

(1) Max-Planck-Institut für Physik komplexer Systeme, Dresden, Germany
(2) Dept. of Electronic Engineering, Univ. of Florence, Firenze, Italy

We propose a noise reduction scheme in order to correct vocal disorders. The filter is based on concepts and tools originated from nonlinear time series analysis and deterministic chaos theory field. This work can provide a better understanding of voice dysfunctions, with a clear advantage for subjects a&# 11;ected by vocal pathologies. With similars tools, indeed, physicians can perform better surgical interventions and researchers can build devices to improve voice quality also without a surgigal treatment. We describe the filter and the idea of geometric signal classification in a feature space in order to visualize the effect of the noise reduction step. We focus our attention on patients a&# 11;ected by T1A glottis cancer and describe which kind of corrections are necessary to improve the quality of the voice, i.e. to move the feature vector from the sick cluster to the healthy zone.

Index Terms. Chaos Theory, Nonlinear Noise Reduction, Feature Space, Vocal Disorders

Full Paper

Bibliographic reference.  Matassini, Lorenzo / Manfredi, Claudia (2001): "Noise reduction for vocal pathologies", In MAVEBA-2001, 156-162.