Geometry to Acoustics

Vocal Tract Geometry to Acoustics for /s/

A collection of images illustrating sagittal, axial, and spectrum views of someone vocalizing the letter S

Zhou, X. et al. (2013) “A Cine MRI-based study of sibilant fricatives production in post-glossectomy speakers”, ICASSP, May, 2013.

 

An MRI and a digital model demonstrating the midsagittal view of the vocal tract

Midsagittal view of a cine-MRI vocal tract and segmented structures including the tongue (pink) and airway (blue). Active tongue and jaw muscles are calculated by the inverse solver in Artisynth, from the vowel /i/ in “geese,” using FEM’s with two quality of resolution: Fehigh and FEreg. The table shows simulation result of the two models and formant frequencies taken from audio recordings in the lab and the MRI scanner.

 

Audio from Lab

Audio from MRI

Simulation FE-reg

Simulation FE-high

F1

268

267

262

256

F2

2272

2055

1905

1995