hillenbrand: phonation1 phonation note: audio demos made with fsyn: original pitch, monotone, and...
TRANSCRIPT
Hillenbrand: Phonation 1
Phonation
Note: Audio demos made with fsyn: original pitch, monotone, and inverted pitch. FDR demo original pitch and monotone only.
Note: Audio demos made with fsyn: original pitch, monotone, and inverted pitch. FDR demo original pitch and monotone only.
Hillenbrand: Phonation 2
Information Conveyed by the Source
For voiced speech, the spectrum of the laryngeal buzz constitutes the source part of source-filter theory. A great deal of the burden of phonetic coding is carried by the filter (e.g., /b/-/d/-/g/; /s/-/S/; /å/-/i/-/u/-/ú/-/ü/, etc.) But, a good deal of speech information is conveyed by the source. For example:
1. Intonation (melodic) contour: Pattern of f0 over time conveys information about the grammatical structure of the utterance (e.g., phrase boundaries and sentence type), as well as affective information.
2. Rhythmic pattern: Pattern of stressed and unstressed syllables can convey lexical information (OBject vs. obJECT) and emphatic stress (e.g., given vs. new).
3. Loudness: Controlled mostly at the source (but filter has some effect on loudness as well).
Hillenbrand: Phonation 3
4. Voice quality:
• Clear or “modal” phonation• Whisper• Breathiness• Roughness• Hoarseness• Diplophonia• “Pressed” voice• Glottal fry• Falsetto• You name it: Other hard-to-classify variations in vocal quality
5. Some segmental phonetic information:
Example: Timing of voicing onset relative to articulatory release is a major cue to the voice-voiceless distinction (more later)
Hillenbrand: Phonation 10
CRICOTHYROID JOINT
Note that when the cricoid moves up (i.e., closing the gap between the cricoid and the thyroid), the arytenoids are rotated away from the thyroid angle. We will see that this has a lot to do with the control of fundamental frequency.
Hillenbrand: Phonation 11
MOTIONS OF THE ARYTENOIDS
Gliding motion of arytenoids brings VFs toward midline.
Hillenbrand: Phonation 12
ROCKING MOTION OF ARYTENOIDS
Rocking forward (toward thyroid angle) brings VFs forward (obviously) and (less obviously) toward midline. Rocking backward (away from thyroid angle) brings VFs backward (obviously) and (less obviously) away from midline.
Thyroid Notch This Way->
Hillenbrand: Phonation 15
Five Layers of VFs:1. Epithelium (very thin, very flexible)2. Superficial layer of LP (thin, gelatinous, very flexible)3. Intermediate layer of LP (rubbery, less flexible)4. Deep layer of LP (like thick thread)5. Vocalis muscle
“Cover-Body” Organization of VFs:Cover= Epithelium + Superficial LPTransition= Intermediate + Deep Layers of LPBody = Vocalis muscle
It is the cover which is most heavily involved in VF vibration – both the side-to-side motion that we all know about, but also the up and down motion that you may be less familiar with.
Hillenbrand: Phonation 16
How the layers of the VFs are organized (from the
Kent Speech Sciences text)
Hillenbrand: Phonation 18
Sequence of events in phonation, beginning with:
Steady lung pressureSteady flow thru VFsAbducted VFs (i.e., away from midline)
1. A steady (DC) muscular force is applied to adduct the folds; i.e., to bring the VFs toward midline.
A Vo(volume velocity; i.e., air flow)V (particle velocity)
2. Bernoulli force increases: The Bernoulli Principle states that an increase in particle velocity is accompanied by an aerodynamic force that is exerted at right angles to the angle of flow.
FBFB
Angle of Flow
Hillenbrand: Phonation 19
The other way to think about FB is to think of it as a drop in pressure or sucking force inside the glottal aperture. Either way, the result is a force that bring the VFs toward midline.
3. Muscular force and Bernoulli force combine to bring the VFs to midline, where they meet.
A Zero (Glottal Area)
VoZero (Volume Velocity; i.e., airflow)
V Zero (Particle Velocity)
FM = Steady (Muscular force)
FB = Zero (Bernoulli force)
Psg = Very rapid and dramatic increase (Subglottal pressure)
4. When folds meet at midline, there are two opposing forces acting:
the muscular force acts to keep the VFs approximated
Psg acts to blow the VFs apart
At some point, Psg will reach a high enough value to win the contest, and:
Hillenbrand: Phonation 20
5. VFs are blown apart, moving away from midline
A (glottal area)Vo(volume velocity; i.e., air flow)
V (particle velocity)
6. The mvt of the VFs away from midline is opposed by:
The DC muscular force, which is still in effectThe elasticity of the VF tissue
The VFs will move toward midline again, and the process is repeated, from step 1.
Hillenbrand: Phonation 21
Vibratory Motion of the Vocal Folds . Note the “Vertical Phase Difference”; i.e., the VFs open bottom edge 1st, followed by top edge; close bottom edge 1st, followed by top edge.
Hillenbrand: Phonation 22
Note that when the VFs separate, they do not just move side-to-side. The folds – especially the top edge – are also displaced upward quite a bit. This is not surprising given the upward direction of the aerodynamic force that causes them to separate in the first place.
Hillenbrand: Phonation 23
THE TWO-MASS MODEL OF PHONATION
Note that the two masses of the vocal folds are represented by a spring and mass system. What factors will control the vibrating frequency of this system?
Hillenbrand: Phonation 24
ANOTHER VIEW OF THE TWO-MASS MODEL
Note that VFs open bottom edge followed by top edge, and close bottom edge followed by top edge.
View from above->
Light gray = top edge
Dark = bottom edge
Hillenbrand: Phonation 25
Control of F0 in the Two-Mass Model
Fundamental Frequency Can be Increased by:
1. Increasing Stiffness: This is done by increasing the longitudinal tension of the VFs, exactly like stretching a rubber band. The stiffness increase results in an increase in natural vibrating frequency.
2. Decreasing the Effective Mass of the VFs: When the VFs are stretched, a smaller portion of the folds vibrates. This is equivalent to decreasing the mass of the VFs. The decrease in mass results in an increase in natural vibrating frequency.
MORAL: Longitudinal Tension F0 Longitudinal Tension F0
Hillenbrand: Phonation 26
INTRINSIC LARYNGEAL MUSCLES AND THE CONTROL OF F0
Four paired muscles (i.e., one on left, one on right), one unpaired muscle.
Paired:
1. Lateral Cricoarytenoid (LCA) Adductor (Closer)
2. Posterior Cricoarytenoid (PCA) Abductor (Opener)
3. Cricothyroid (CT) Longitudinal tension increaser/decreaser
4. Thyroarytenoid (TA) [Internal (vocalis) / External]Function depends on behavior of other muscles
Unpaired:
Interarytenoid (IA) [Transverse & Oblique]Adductor
Hillenbrand: Phonation 27
LATERAL CRICOIDARYTENOID (LCA)
This muscle pulls downward and forward on the arytenoids. Contraction has the effect of rocking the arytenoids forward. Given the “toe in” angle of the arytenoids, this forward rocking motion adducts (closes) the VFs (and may increase medial compression; i.e., squeezing force).
The LCA may also reduce the longitudinal tension on the VFs. (Note: Only the right LCA is shown in this picture.)
Hillenbrand: Phonation 28
Posterior Cricoarytenoid (PCA)
This muscle pulls back on the arytenoids. This has the effect of rocking the arytenoids backward. Given the “toe in” angle of the arytenoids, this backward rocking motion abducts (opens) the VFs (and may decrease medial compression; i.e., squeezing force).
The PCA may also increase the longitudinal tension on the VFs.
View from the Back
Hillenbrand: Phonation 29
Cricothyroid Muscle (CT)
This muscle pulls the cricoid up, reducing the distance between the cricoid and the thyroid. Most Important: This mvt rotates the cricoid lamina back and away from the thyroid notch. This pulls the arytenoids away from the thyroid notch, increasing the tension on the VFs.
Hillenbrand: Phonation 30
Main Point: The CT increases the longitudinal tension of the VFs, decreasing effective mass, and increasing F0.
Hillenbrand: Phonation 31
Thyroarytenoid Muscle (TA)
• Note internal and external parts of TA.• Internal TA also called vocalis muscle.
Hillenbrand: Phonation 32
Interarytenoids (IA)
• Note transverse (side-to-side) and oblique parts of IA.
• Contraction of transverse IA produces gliding motion of arytenoids; result is adduction and medial compression (squeezing).
• Contraction of oblique IA may cause apices of arytenoids to approximate.
Hillenbrand: Phonation 34
Relationship Between Glottal Area and Glottal Volume Velocity (Air Flow)
• When area is large, flow is high. No big surprise: When a faucet is full on, flow is high.
• The flow waveform is steeper than the area waveform. (Flow is proportional to Area3; e.g., if area is doubled, flow increases by a factor of 23= 8.)
Hillenbrand: Phonation 35
Time Organization/Frequency Organization
This figures shows just two extremes:
•A nearly impulse-like waveshape (high time organization – events are “compressed” in time) with lots of energy spread to the upper harmonics (low frequency organization).
•A nearly sinusoidal waveshape (low time organization – events are spread evenly over time) with nearly all of the energy at the fundamental frequency (high frequency organization).
Hillenbrand: Phonation 36
Time Organization/Frequency Organization
Note that more impulsive-looking waveforms produce more energy spread into the higher frequencies. The more smooth and sinusoidal-looking waveforms have a greater amount of their energy concentrated at the 1st harmonic (f0), and less energy in higher frequency harmonics.
Hillenbrand: Phonation 37
Effects of Gradual vs. Abrupt Glottal Closure
The glottal waveforms above differ only in the abruptness of glottal closure. Notice that the glottal waveform with more gradual closure is fairly weak in higher frequency harmonics. Conversely, the glottal waveform showing more abrupt closure shows stronger upper harmonics. Rapid glottal closure is accomplished mainly by the lightest and most flexible portion of the VFs – the VF cover (i.e., epithelium & superficial layer of the LP).
More abrupt closure, more energy spread to harmonics above f0.
Gradual closure, energy concentrated strongly at f0.
Hillenbrand: Phonation 38
MORAL: Time organization and frequency organization are inversely related.
•When time organization is high (like an impulse), frequency organization is low (energy is spread or “splatters” into the higher frequencies).
•Conversely, when time organization is low (like a sinusoid), frequency organization is high (energy is concentrated near a single frequency).
SO:
•Transient-looking waveforms have a lot of energy spread into the higher frequencies.
•Sinusoidal-looking waveforms have most of their energy near the fundamental.