Effects of Audio Compression in Automatic Detection of Voice Pathologies

Abstract

] T. Hildebrand and P. Rüegsegger, "A new method for the modelindependent assessment of thickness in three-dimensional images," J. Microsc., vol. 185, pp. 67-75, Jan. 1997 Effects of Audio Compression in Automatic Detection of Voice PathologiesNicolás Sáenz-Lechón * , Víctor Osma-Ruiz, Juan I. Godino-Llorente, Manuel Blanco-Velasco, Fernando Cruz-Roldán, and Julián D. Arias-LondoñoAbstract-This paper investigates the performance of an automatic system for voice pathology detection when the voice samples have been compressed in MP3 format and different binary rates (160, 96, 64, 48, 24, and 8 kb/s). The detectors employ cepstral and noise measurements, along with their derivatives, to characterize the voice signals. The classification is performed using Gaussian mixtures models and support vector machines. The results between the different proposed detectors are compared by means of detector error tradeoff (DET) and receiver operating characteristic (ROC) curves, concluding that there are no significant differences in the performance of the detector when the binary rates of the compressed data are above 64 kb/s. This has useful applications in telemedicine, reducing the storage space of voice recordings or transmitting them over narrow-band communications channels.Index Terms-Gaussian mixture models, MP3 compression, support vector machines, voice pathology detection.

Common sounds in bedrooms (CSIBE) corpora for sound event recognition of domestic robots

“…Two popular lossy compression algorithms were examined, the Ogg Vorbis transcoding was implemented with libvorbis 3 , the MP3 encoding with libmp3lame 4 and the MP3 decoding with LAME's mpglib version. The effects of lossy encoding has been studied in the literature for speech (Besacier et al, 2001;Pollak and Behunek, 2011;Sáenz-Lechón et al, 2008) and music (Uemura et al, 2014;Urbano et al, 2014) classifications, but this paper provides the first analysis for sound event recognition. The cited papers from the literature examined only the MP3 encoding while both Ogg and MP3 codecs are reviewed here.…”

Análisis de la influencia de las técnicas de compresión de voz en la detección de anomalías vocales

“…En [16] se reporta una precisión del 89,7% usando la metodología EMD (Empirical Mode Decomposition) para clasificación, asumiendo una relación señal-ruido del canal de 30 dB. En [17] se analizan los efectos de la compresión de señales de voz en formato mp3 en la detección de patologías, usando como métodos de clasificación GMM (Gaussian Mixture Models) y SVM (Support Vector Machines); se obtiene como resultado precisiones del 87,05% usando SVM y 85,37% usando GMM, para señales comprimidas a 8 kpbs. En [18] se propone un método para mejorar la precisión de un sistema que transmite las señales de voz haciendo uso de redes inalámbricas.…”

Fusing voice and query data for non-invasive detection of laryngeal disorders

“…Voice is a convenient non-invasive modality of information about subject's larynx and audio analysis is being applied increasingly (Godino-Llorente et al, 2001;Dibazar et al, 2002;Hadjitodorov & Mitev, 2002;Godino-Llorente & Vilda, 2004;Alonso et al, 2005;Moran et al, 2006;Sáenz-Lechón et al, 2006;Shama et al, 2007;Sáenz-Lechón et al, 2008;Gelzinis et al, 2008;Salhi et al, 2008;Hariharan et al, 2009;Henriquez et al, 2009;Marinus et al, 2009;Markaki & Stylianou, 2009Martínez et al, 2012;Saudi et al, 2012;Vaiciukynas et al, 2012Vaiciukynas et al, , 2014b. Even telephone-based voice samples (Moran et al, 2006) and compressed recordings (Sáenz-Lechón et al, 2008) may lend itself to voice pathology detection.…”

“…Even telephone-based voice samples (Moran et al, 2006) and compressed recordings (Sáenz-Lechón et al, 2008) may lend itself to voice pathology detection. Voice data-based detection of mildly pathological larynx by Vaiciukynas et al (2014a,b) also gave promising results.…”