TubeTalker Demonstration

 

“TubeTalker” is a moniker for the area function model described in:

 

Story, B.H., (2005).  A parametric model of the vocal tract area function for vowel and consonant simulation, J. Acoust. Soc. Am., 117(5), 3231--3254.

 

story_jasa2005.pdf

 

The following are demonstrations of some of the speech sounds produced with TubeTalker coupled to a digital waveguide model of the vocal tract and a glottal area function type model for the voice source.  

 

Slide Show from XV Pacific Voice Conference
at PIXAR Studios 03.11.05

 

story_pvsf2005_pixar.pps

 

 

“Oh yeah”

 

oh_yeah_base.wav

ohyeah1_low.wav

ohyeah1_med.wav

ohyeah1_high.wav

23 cm long vocal tract

11 cm long vocal tract

 

 

“oCae” series (VCV)

 

"oae" (no consonant)

obae.wav

odae.wav

ogae.wav

odrae.wav

omae.wav

onae.wav

ongae.wav

 

 

Audio and video demos from ICVPB 2008 conference
in Tampere, Finland 08.09.08

 

These demos are based on the same “TubeTalker” model for changing the vocal tract shape. The voice source is a kinetic model of the 3D glottis based on Titze (2006), The myoelastic aerodynamic theory of phonation, NCVS, pp. 197-214.

 

Animation of kinematic model of the 3D glottis:

 

1.    Adductory maneuvers without vibration

2.    Adductory maneuvers with vibration

 

 

Adductory Glides:

 

Long "hahaha"

Short "hahaha"

 

 

Building the sentence “He had a rabbit and a boat.”

 

Audio

1.    Vowel sequence + pitch changes

2.    Vowel sequence + pitch changes + adductory changes

3.    Consonant articulation + #2 (complete sentence)

4.    Faster version of #3

5.    Even faster...

6.    Complete sentence but with breathy voice.

7.    Complete sentence but with laryngeal tremor.

8.    Complete sentence with lengthened vocal tract and wide epilarynx

9.    Complete sentence with lengthened vocal tract and constricted epilarynx

10. First attempt at singing the sentence

11. Second attempt at singing the sentence.

12. Another attempt at singing - but one consonant is changed. Which one?

 

Animations (note: so that vocal tract shape changes can be easily observed, these run about 2.5 times slower than real time)

 

1.    Vowel sequence for the sentence - tubular representation and formant tracks

2.    Complete sentence: vowel sequence + consonants (tubular representation and formant tracks)

3.    Complete sentence: tubular representation with 3D view

4.    Complete sentence: tubular representation in vocal tract profile view