Front Cavity Research Articles

Nonlinearities in articulatory/acoustic relations and the evolution of the human vocal tract

The evolution of the modern human vocal tract is characterized by a progressive increase in the degree of flexion of the base of the skull. The acoustic consequences of this basicranial flexion were analyzed using a transmission line analog model and calculating variations in vocal tract transfer functions as the trailing edge of a constriction is moved anteriorly. The behavior of an unflexed tract was modeled by trading back and front cavity lengths. That of the modern human vocal tract was modeled similarly, except that after the distance of the trailing edge of the constriction tube from the larynx increased past the superior end of the pharynx, the cross-sectional area of the back cavity was increased (instead of its length) as front-cavity length decreased. The modern human vocal tract had a larger area of stability for the second formant frequency for [i]-like constrictions than that produced by the unflexed tract. It appears that the human vocal tract has evolved in the direction of increased size of quantal locations for vowel production.

Some useful inexactnesses in speech.

The tension between ideas and facts is characterized as a duality between “something going on”and “something being wrong”. Example 1 is that formants can be estimated by a simple and seemingly absurd linear combination of cepstral coefficients. Example 2 is the apparently simplistic hypothesis that the second equivalent spectral peak F2' is the resonance of the front cavity of the vocal tract.

Acoustic estimations of the front cavity in apical stops

Sweep-tone measurements of front-cavity resonances in apical stops are reported. Volumes inferred from resonance frequency data are presented as a function of place of articulation. The results indicate that, for post-dental points of contact, a cavity located under the tongue tip is formed. It is shown that this sublingual cavity plays a very important role in shaping the spectral structure of apical stops.

The front‐cavity/F2′ hypothesis tested by data on tongue movements

The hypothesis that the second spectral peak F2′ resulting from perceptual linear prediction (PLP) corresponds to the resonance of the front cavity of the vocal tract is tested against data on tongue movements for non‐low‐back vowels. The front cavity is represented for this case as a uniform tube of length Lf. Scaling between movements of the tongue Δx (range of a few centimeters) and of the vocal‐tract constriction ΔLf (range of about 10 cm) is represented by an articulatory amplification factor α: ΔLf = αΔx. If x is a tongue‐surface coordinate parallel to the tract axis, then 1/F2′ = k + 4αx/c, where k is a constant and c is the speed of sound. When x is the x coordinate of a tongue‐dorsum pellet tracked by x‐ray microbeam, the correlations between 1/F2′ and x are 0.74 and 0.75 for two speakers' triple repetitions of /wiwɹjuəjiɹ/. The respective estimates for α are 5.5 and 5.6. The correlations rise to 0.91 and 0.92 if x is a weighted sum of the x and y coordinates of three tongue pellets. Two predictions are thus confirmed: (1) a linear relation between 1/F2′ and tongue movements, and (2) reasonable values for α. A more complex model of the front cavity may be needed to handle low back vowels. [The authors thank George Papçun of Los Alamos National Laboratory for access to his x‐ray data.]

Grain size gradient nickel alloys II: Fatigue properties

Grain size gradient nickel alloys were tested in a rotating-bending fatigue machine and found to display better fatigue behavior than non-gradient, homogeneous nickel specimens having the same average grain size. This demonstrates that the overall fatigue properties are improved through the concept of a grain size gradient. The following reasons are suggested for the improved fatigue behavior. For the case of a gradient alloy with fine grains at the exterior and coarse grains in the center, the combined effect of a fine-grained surface and the presence of Al 2O 3 particles (grain growth inhibitor) increases fatigue crack initiation resistance, while the coarse-grained central region increases fatigue crack propagation resistance. For an alloy having the reversed grain size gradient, the coarse-grained surface increases crack propagation resistance near the threshold region, and the central, Al 2O 3 - added region contributes to fatigue resistance by enhancing yield strength. In both cases, when the applied stress is small, the fatigue crack propagates intergranularly at the beginning of the fracture process, but changes to a transgranular mode with striations as the stress intensity factor ΔK increases. At higher stresses the fatigue crack growth occurred only in a transgranular mode with striations. Grain size was found to exert a stronger influence upon the overall fatigue behavior than do the dispersed Al 2O 3 particles, probably because of the two opposing effects of oxide particles on crack front deflection and cavity formation. The portion of intergranularly fractured surface area decreases with an increase in the applied stress, and fatigue life increases with an increase in the portion of intergranularly fractured surface area.

Acoustic characteristics of fricatives and fricative‐like models

Spectra of spoken fricatives and of the output of mechanical models designed to mimic fricatives were compared in order to test the applicability of such models to speech, and to explore the acoustic mechanisms of fricative production. Power spectra were computed from recordings of five speakers uttering the sustained fricatives [φ, f, θ, s, s, x]. Mechanical models were assembled using a variety of articulatory and aerodynamic data, guided by previous work with models [C. Shadle, J. Acoust. Soc. Am. Suppl. 1 75, S24 (1984)]. The models fell into three groups according to source location and type. To produce [s, s] an obstacle is required at the approximate location of the teeth. A very short front cavity is required for [φ, f, θ]; the jet of air must be located near the tube wall or other surface to mimic the role of the lips. For [x, c] the palate and possibly alveolar ridge act as a turbulence‐generating surface anterior to the constriction. Although perfect spectral matches are not attained in all cases, the mechanical models can be used to test the adequacy of source‐filter representations of the fricatives. [Work supported by NINCDS.]

Release Burst and Formant Transition as Acoustic Cues for English Stop Consonants Perception by Native Speakers of Indian Languages

Acoustic cues, burst and formant transition, are investigated for English stop consonant perception by native speakers of Indian languages. Synthesized initial stop CV syllables are used. Burst is found to be usually of low intensity and nearly invariant for labials and apicals. It is not at all important in perception of labials, but carries a significant perceptual weight for velars, and is a sufficient cue for /di/. Formant transition though perceptually very important is not a sufficient cue for velars /gV/, /di/ & /de/. These results offer further evidence to perceptual importance of front cavity resonance, serving as a simplifying reference to articulation. Hence pointing at the functional equivalence of the two cues. Both providing a spectrally continuous change from the release into the following vowel. In general we suggest that burst carries perceptual weight only when its spectrum is near to the front cavity resonance of the vowel. And formant transition though very important, is a sufficient c...

Stop consonant place perception with single-formant stimuli: Evidence for the role of the front-cavity resonance

The third formant and the second formant were found on average to cue the place of articulation of intervocalic stop consonants equally well when the stop consonants occurred before the vowel/i/. This result and others provide some support for the notion that the fundamental resonance of the front cavity plays an important role in the perception of the phonetic dimension of place of articulation.

Choral speech and phonetic invariance

In rhythmic choral speech, all speakers produce the same phonetic message at the same rate. A recording of such speech can be viewed as an algebraic summation in which speaker dependent variability is essentially reduced to the level of background noise. Intelligibility tests and acoustic analyses will be done on recordings of a large group of speakers producing short phrases containing stressed vowels in CVC syllables. The results will be considered in light of two competing theories of perceptual normalization; Kuhn [J. Acoust. Soc. Am. 58, 428–433 (1975)] which suggests that front cavity information is invariant across speakers; and the standard formant ratio hypothesis. Preliminary data from a mixed chorus of 30 members indicates that intelligibility is maintained for phrase length utterances, and further that several consonants and most vowels are readily perceptible, even in excised syllables.

Stop bursts and transitions as functionally equivalent, context-sensitive cues

The experiment reported here was primarily designed to test the functional invariance of stop bursts as perceptual cues. Bursts from three sets of nine CV syllables for each stop-consonant place of articulation were transposed within each set and played for listeners. Although some measure of perceptual invariance was found within each of the labial and apical sets, and within two subsets for the velar bursts, the perceptual weight carried by the bursts attained significance in few of the syllables tested. Further, the perceptual weight of the bursts and formant transitions of the syllables tended to be reciprocally related. These two cues appear to be functionally equivalent, context-dependent cues, each contributing to the rapid spectral changes that follow consonantal release. The results indicate the importance of the role of front cavity resonance in signaling place of articulation.

On the front cavity resonance and its possible role in speech perception.

Spectrographic data are presented which suggest that it may be possible to estimate the frequency of the fundamental resonance of the cavity behind the mouth opening, the ’’front cavity resonance,’’ from information in the speech signal. It is shown that place of articulation information in the steady states, transitions, and bursts of F2 (or sometimes F3) can be reinterpreted to be information from the front cavity resonance. Furthermore, a number of synthesis results that have appeared anomalous when described in terms of numbered formants seem to find a coherent explanation in terms of the front cavity resonance. Implications for theories of speech perception include the possibility that an estimate of front cavity resonance frequency may serve for continuous articulatory reference. Subject Classification: 70.20, 70.30.

A study of fine structural changes in the cartilage-to-bone transition within the developing chick vertebra.

AbstractForty‐eight chick embryos were killed at 9–16 days of incubation age. Tissue was obtained from the fourth or fifth cervical vertebra, immersed in Karnovsky's fluid, post‐fixed in osmium tetroxide, dehydrated in ethanol, stained “en bloc” with uranyl acetate in ethanol, and embedded in Epon 812. Vertebrae were oriented for cross‐sectional microtomy in cephalic to caudal sequence. Thin sections were stained with uranyl and lead salt solutions saturated with tribasic calcium triphosphate to prevent decalcification.Chondrocytes within the cartilaginous vertebral body occur in various stages of degeneration without orderly arrangement. Both reversible and irreversible stages are found at the cartilaginous resorption front. Electron‐lucent osteoid and mineralization appear in the intercellular matrix at about 12.5 days. Rapidly invading blood vessels form a highly variable resorption front and irregular marrow cavity. Capillaries with accompanying cells border on the front, but else‐where open capillaries allow blood elements to be in direct contact with cartilage. Chondroclasts are associated with small areas of calcified cartilage. At about 14 days trabeculae are formed at the resorption front by osteoblasts which deposit bone osteoid on uncalcified cartilaginous matrix. The matrix is eroded away. A free trabecula of bone without a core of calcified cartilaginous matrix remains.Basic differences between developmental growth processes in the epiphyseal plate and vertebral body may stem from the large amount of uncalcified cartilaginous matrix at the latter's resorption front.

Control Rule of Vocal-Tract Configuration

Objectives of the research is to find out a simple but an effective rule for the control of vocal-tract configuration in the synthesis of speech. One of the following two alternatives is often selected as the control rule: (1) independent linear interpolation, (2) mathematical descriptions with several parameters. The former has an advantage of simplicity but also has suffered from over simplified deformation. The latter is hardly applicable to consonant. The proposed processes are as follows: (1) Define three key points for each target configuration; they are maximum area points of front and back cavity V1, V2, and maximal constriction point P between these two. (2) Trace the trajectory of these key points between two target configurations in a certain time fashion. (3) Calculate the temporal change of each section by the interpolation from starting target value to ending target value with considering a restriction due to the pass of V1, V2, and P point. The proposed rule has two advantages: (1) applicable to any target configurations including consonants, (2) the restrictions avoid the results from over-simplified deformations. Some of the result will be shown with aids of our vocal tract pattern display.