This shows you the differences between two versions of the page.

Link to this comparison view

a_vocal_fold_model_for_disordered_voice_synthesis_to_develop_auditory_perception_in_speech_language_patologists [2014/04/04 13:36] (current)
bziolko created
Line 1: Line 1:
 +Mara Behlau(1,2), Glaucya Madazio(2), Marina Englert(1,2), Jorge Lucero(3), Jean Schoentgen(4), (Brazil & Belgium)
 +**A Vocal Fold Model for Disordered Voice Synthesis to Develop Auditory Perception in Speech Language Patologists**
 +(1) Universidade Federal de São Paulo – Unifesp, Brazil\\
 +(2)  Centro de Estudos da Voz – CEV, Brazil\\
 +(3) Universidade de Brasília – UNB, Brazil\\
 +(4) Université Libre de Bruxelles and National Fund for Scientific Research, Belgium\\
 +Introduction: Auditory analysis is the gold standard to evaluate dysphonia, training students is a challenge. Synthesizes stimuli may facilitate vocal identification, by manipulating vocal characteristics; however, these stimuli must sound natural. 
 +Objectives: Test a dynamic model of the vocal fold oscillation that simulates voice disorders with good naturalness and to introduce a training program with synthesized voices. 
 +Model: A surface wave propagates through the vocal fold tissues, glottal air pressure is approximated by using linearization and the glottal area is computed at the midpoint of the glottis. The vocal tract is represented with concatenated tubes. Three ways to simulated voice disorders: perturbations of vocal frequency (flutter, tremor and wow); additive noise (pulsatile noise and aspiration noise) and vocal fold tension asymmetry. 
 +Method: Six independent listeners submitted to two listening session where they should evaluate the naturalness, deviation, type and degree of two sets of vocal samples (Mr and St) of synthesizes voices with 20 files each. Results: The St vocal tract was classified as twice more artificial than Mr tract. The stimulus that varied on flutter were considered more natural and stimulus with high levels of pulsatile noise were classified as more artificial. The perceptual assessment showed a high naturalness of synthetic voice. The isolated manipulation is ideal for teaching, there is a greater intersubjects agreement especially on breathiness, the association of two or three parameters should be used on advanced stages of education and they bring more disagreement, also the deviations of symmetry must be more explored. Training proposal: 1.train breathiness alone, 2. train roughness alone, 3. combined variation of jitter and signal to noise ratio.  Possible Applications/Conclusion: Use of synthesizes stimulus for educational purpose (voice clinicians, SLP, Physicians), investigation purpose (understand the production and characteristics of abnormal vocal timbres) and for research purpose (offer anchor stimulus).
Copyright © XXII PVC Organizing Committee 2013. All Rights Reserved.