Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Source Modeling of Severely Pathological Voices Bruce R, Gerratt*, Jody Kreiman*, Norma Antonanzas-Barroso*, Brian Gabeiman*,
 

Summary: Source Modeling of Severely Pathological Voices
Bruce R, Gerratt*, Jody Kreiman*, Norma Antonanzas-Barroso*, Brian Gabeiman*,
and Abeer Alwan* *
"Division of Head and Neck Surge~ and *"Dept. of Electrical Engineering,
UCLA, Los Angeles, CA 90095 USA
Abstract: Our previous attempts at synthesizing severely pathological voices were hampered by perceptually-
significant errors in vowel quality. To determine whether these were caused by methodological limitations or by
source-tract interactions, normai and pathological voices were recorded using both a condenser microphone and a flow
mask system. Synthetic versions of each signal were produced and compared perceptually to the original signais.
Limitations of all-zero inverse filtering techniques as appiied to pathological voices wiii be discussed.
INTRODUCTION
Attempts at modeiing pathological phonation using anaiysis by synthesis depend on accurate modeling of
the pathological voice source and vocai tract resonances, The source finction is typically estimated using
inverse filtering, either of the glottal flow signal (e.g., 1) or of the acoustic signai as transduced with a condenser
microphone (2).
Because of the restricted frequency response of the flow mask system, our preliminary experiments
synthesizing pathological phonation used estimates of vocal tract characteristics derived from a microphone
signal, combined with estimates of vocai tract resonances derived from a flow mask signal (recorded from a
different segment of the same utterance). These studies suggested that the frequency response characteristics of
the flow mask sys~em may limit its applicability to severely pathological phonation, particularly when precise

  

Source: Alwan, Abeer - Electrical Engineering Department, University of California at Los Angeles

 

Collections: Computer Technologies and Information Sciences