1. Acoustic Phonetics 3/9/00

2. Acoustic Theory of Speech Production Modeling the vocal tract –Modeling= the construction of some replica of the actual physical object under study –Tube to simulate the vocal tract –Model gradually changes to what we know the vocal tract shape to be

3. Pipe Models A. A uniform straight tube (closed at one end & open at the other) *This tube is closed at one end to represent the closed vibrating vocal folds, The open end represents the lips separated for the production of vowel sounds. l1l1 A. ClosedOpen

4. Pipe Models B. A uniform straight tube of a greater length than A. closed at one end and open at the other end. * First modification of this model was to change the length. l2l2 B.

5. Pipe Models C. A straight tube of varying cross-sectional diameter. * Next modification is to shape the pipe so that the diameter varies along the length. Now the tube flares at its open end (lips). C.

6. Pipe Models D. A bent tube of varying cross sectional diameter. * The final modification is a bent tube to represent the right angle of the vocal tract D.

7. The Simple Pipe Model: Modeling One Vowel A. Closed at one and and open at the other: 17.5 cm long –Exhibits resonance under the right conditions: Frequency of sound; length of tube Odd-quarter wavelength –Wavelength of a sound = Distance traveled by a soundwave during one period of vibration Wavelength depends on the frequency of vibration & speed of sound  = s/f (Where is the wavelength in meters, s is the speed of sound in meters/sec, and f is in frequency in hertz

8. The Simple Pipe Model: Modeling One Vowel In this example the tube length is adjusted to the wavelength of a tone (Similar to only a mid-central vowel in speech) –Other resonance's occur when the tube changes shape –The energy of human speech is concentrated at relatively low frequencies and higher areas can be neglected Concentration on first 3 resonance's of the vocal tract or formants. –Formants are identified by formant number (F1, F2, F3) Typical Male (production of a mid central vowel): F1= 500 Hz, F2= 1500 Hz, F3= 2500 Hz

9. Adjusting the Length of the Tube: Modeling Vowels of Men, Women, & Children B. First modifications of the simple tube is length. –Length of the tube determines the frequencies of its resonance's –Longer tube has lower resonance frequencies –Shorter tube has higher resonance frequencies Short tube = Child (8-10 cm) Moderately long= Women (15 cm) Long= Men (17.5 cm)

10. Calculated Formant Frequencies (F1, F2, F3) F1F2F3 Neonate Woman Man 109432825470 583 17492915 50015002500 * Neutral, or central, vowel in which the tongue and other articulators are positioned so that the vocal tract has a uniform cross-sectional area from the glottis to lips

11. Adjusting the Cross-Sectional Area of the Tube: Making More Vowels C. Length and cross sectional area variation –/ i/ /u/, / a / and /æ / (Most frequently occurring in the English language) –Modifications to the pipe for this model is by varying the tube’s cross-sectional area so it resembles the vocal tract configurations of the specific vowel –/ i/- Back cavity enlarged relative to the front cavity /u/- There is a constricted midway along the tube and at its mouth / a /- The front cavity is enlarged relative to the back cavity /æ /- The front cavity is relatively enlarged compared to the back cavity

12. Vowels: Vocal Tract Configurations

13. Acoustic Vowel Quadrangle 1. Defines the acoustic space of vowel production 2. Interpret articulatory dimensions of the quadrangle can be interpreted acoustically 1. The high vowels / i/ & /u/ have low F1, but low vowels / a / & /æ / have high F1. *Relative value of F1 is associated with tongue height 2. The back vowels /u/ & / a / have a low F2, but the front vowels / i/ & /æ / have a high F2. *Relative value of F2 is associated with A-P position of the tongue i æ a æ a u u i Women Men F1(kHz) F2 (kHz)

14. Example Using the vowel quadrangle, describe the difference in formant structure for the following word pair: –seat- suit

15. The Voice Source: Resonators Harmonics

16. The Source-Filter Theory Resonator & Source Energy –Filter = Vocal tract Energy passed in a frequency selective manner Production of different vowels changes the filter shape –Source = Harmonic spectrum of voicing Fundamental frequency & its harmonics Source spectrum is the acoustic energy activating the formants Source-filter theory of vowel production: –The energy form the source (vibrating vocal folds) is modified by the resonance characteristics of the filter (vocal tract)

17. Source-Filter Theory of Speech Production

18. Source-Filter Theory Radiation Characteristic- Exerts final shaping on the acoustic signal –Takes into account the way the vocal tract opens into the atmosphere (baffle effect) –+6 dB increase at every octave Source energy has a -12 dB/octave and the radiation characteristic is +6 dB the result is -6 dB/octave fall-off in energy –Result= Radiated pressure spectrum

19. Extending the Theory: Consonants Fricatives: –Voiceless fricative /s/ –Energy source: Not vocal fold vibration, but turbulent noise generated in the vocal tract –Cavity lying in front of the noise source is the primary resonator Length of front cavity determines lowest frequency resonance –Short length: Labiodental (/f/, /v/) –Longer length: Alveolar (/s/, /z/)

20. Model: Filtering of Frication Noise Point of Noise Generation Noise shaped by resonance of the front cavity & sometimes back.

21. Different Filtering: Vowels vs. Consonants Filtering for consonants involves both formants & antiformants –Anitformants= the opposite of formants in that they result in a loss of sound energy rather than enhancement Antiformants arise when: –Vocal tract is radically constricted (stops, fricatives) –Vocal tract is bifurcated or split into 2 passages (nasal)