Hisayoshi Suzuki

MORPHOLOGICAL AND ACOUSTICAL ANALYSIS OF THE NASAL AND THE PARANASAL CAVITIES

Morphological measurements of the nasal and paranasal cavities were conducted to investigate their relevance to the acoustic properties of the human nasal tract. The magnetic resonance imaging (MRI) technique was used to measure the three-dimensional geometry of the vocal tract. The area function of the nasal tract was calculated for seven subjects based on data obtained during natural breathing. The entire vocal tract was measured for five subjects during sustained production of nasal consonants. A marked morphological difference was observed between our data and previously published data [A. S. House and K. N. Stevens, J. Speech Hear. Disord. 21, 218-232 (1956); G. Fant, Acoustic Theory of Speech Production (Mouton, The Hague, 1970), 2nd ed., p. 139] particularly in the middle portion of the nasal tract. Previous data derived from cadaver specimens showed a large cavity in the middle portion possibly due to an absent or dehydrated mucous membrane, while our data showed narrow passages due to thickly layered mucosa. It has been confirmed by an additional experiment that the wide cavity is reproducible by applying an adrenergic agent to the nasal mucosa. Transfer functions of the vocal tract and the nasal tract were calculated from measured data, and compared to spectra of real speech signals recorded subsequent to the MRI experiment. The results indicate that asymmetry between the two nasal passages can cause extra pole-zero pairs, and suggest that the paranasal cavities play an important role in shaping the spectral characteristics of human nasal sounds.

An experimental study of the open end correction coefficient for side branches within an acoustic tube

The open end correction coefficient (OECC) of acoustic tubes has been widely investigated under a free-field condition. This study examines OECCs in confined regions, such as side branches within the vocal tract. To do this, a number of mechanical acoustic models are used to examine the effects of the angle of the branch axis and the proximity of the walls of the main tract to the open end of the branch. The OECC is estimated by matching both the peaks and troughs (i.e., spectral maxima and minima) of the computed and measured transfer functions for each model. The results indicate that the OECC of a side branch depends on L/D, where L is the cross dimension of the main tract at the branching point, and D is the branch diameter. For side branches connected to the main tract through a narrow neck, the OECC of each end of the neck is determined using the ratio of the radius of the neck to that of the adjacent section. Two empirical equations for evaluating the OECC within a tract are derived from the present study. Finally, the range of appropriate OECC values for estimating an accurate vocal tract transfer function is discussed, based on the results presented here and morphologic measurements reported previously.

MRI measurements and acoustic investigation of the nasal and paranasal cavities

Morphological measurements of the nasal and paranasal cavities were performed to speculate on the acoustic properties of the human nasal tract. The magnetic resonance imaging (MRI) technique was used to measure the three‐dimensional structure of the vocal tract. The area function of the nasal tract was calculated from seven subjects at rest. The whole vocal tract was measured from five subjects during sustained pronunciation of nasal consonants. A marked morphological difference was observed between these data and the previous data particularly in the middle portion of the nasal tract. The previous data measured from cadaver specimens have a wide cavity in the middle portion possibly due to absent or dehydrated mucous membrane, while these data show narrow passages caused by thick mucosa. It was confirmed by an experiment that the above difference was reproducible by applying an adrenaline‐like agent onto the nose. The acoustic transfer function was calculated from these data, and the speech sound, record...

Analysis of acoustic properties of the nasal tract using 3-D FEM

In order to examine the acoustic effects of the complicated morphological construction of the nasal tract, we have analyzed acoustic models constructed according to measurements by magnetic resonance imaging (MRI). A prototype model was made to be as similar as possible with the actual nasal tract, which is asymmetrical in the left and right passages and has complicated cross-sectional shapes. Sound pressure, particle velocity and sound intensity in the model were calculated by the finite element method (FEM). Several modifications were applied to the shape of the prototype model in order to learn what acoustical effects are produced by the following modifications: (1) models having an elliptic shape, (2) models with and without a pair of maxillary sinuses, (3) a left-right symmetry model in which one passage is modified to be identical with the other passage, and (4) models having narrowed or blocked passages. The results show that the poles and zeros are produced and shifted by a mutual branching effect caused by the left-right asymmetry of the nasal passages and an additional side-branch effect of the sinus cavities. Those effects are also produced by complicated cross-sectional shapes of the nasal tract in the 3 kHz region. The reduced cross-section in the narrowed-passage models causes a shift of the pole and zero frequency and weakens the mutual side-branch effect of the left and right passages when either of them are excessively narrowed.

An experimental study of the open‐end correction coefficient for side branches within an acoustic tube and its application in speech production

The open‐end correction coefficient (OECC) for side branches within an acoustic tube was investigated experimentally, and combined with morphological data of the vocal tract to arrive at OECC estimates useful in speech production. A number of mechanical acoustic models were used to examine the effects of the angle of the branch axis, and the proximity of the main tube around the open end of the branch. The results indicate that the OECC (α1) depends on the cross‐dimension L of the main tract at the branching point and the branch diameter D (α1=0.414 +0.517⋅log10(L/D2), 0.158<L/D2<6.97). For side branches connected to the main tract by a narrow neck, the OECC of each end of the neck is determined by α2=0.82(1−ξ)/18ξ2−3.1ξ+1, where ξ= the ratio of radius of the neck to that of the adjacent section. Using morphological data of the vocal‐tract, the range of appropriate OECC values for estimating accurate vocal‐tract transfer function was evaluated. The OECC of the nasopharyngeal port was about 0.8 for the ora...

Examination of the open end correction coefficient for side branches within an acoustic tube

The open-end correction of an acoustic tube has been widely investigated in a free-field condition (cf. Hall, 1987). However, no experimental data give accurate open-end corrections for side branches within an acoustic tube, although some proposals have been made [see Fant (1960)]. To incorporate side branches such as the paranasal sinuses and piriform fossa in an acoustic model of the vocal tract, an end correction coefficient is definitely needed. Thus this study estimated the end correction coefficient of side branches experimentally and computationally. A number of acoustic tubes were used to examine effects of (1) the angle from the branch to the main tube, (2) the distance between the open end and the opposite tube wall, and (3) the geometric shape around the open end of the branch. The results indicate that the end correction coefficient depends on the dimension in front of the open end, and is independent of the angle of the branch to the main tube. The open-end correction coefficient (Ce) of a branch can be described by the diameter (D) of its open end and the distance (L) from the end to the opposite tube wall as shown: Ce = 0.414+0.517 × log (L/D2), for 0.158 < (L/D2) < 6.24. Application of this empirical formula to actual vocal geometries will be discussed.

Speech production by a vocal cords‐vocal tract‐vocal tract wall vibration model

This research was aimed at finding the effects of finitehess in the mechanical impedance of the vocal organs on speech sounds. A speech synthesis program has been developed considering the vocal tract wall vibration that causes sound leakage from the vocal cavity to the nasal cavity even when the velum is closed. The radiation of sound from the outer surface of the vocal tract is not always negligibly small, and its effect is enhanced by the nasal cavity. The synthesis program has the parameters of the physical properties and pressure of ambient gases, as well as the mechanical impedance of a yielding vocal tract wall. The wall vibration is treated as a small perturbation in the area function of the vocal tract. The following items will be discussed in the paper: comparisons between vowels with and without vocal tract wall vibration, sounds from the mouth opening, nostrils, and vocal tract wall, and the distribution of the strength of the wall vibration along the vocal tract. The effect of the pressure of...

Speech Production Model Involving the Subglottal Structure and Oral-Nasal Coupling due to Wall Vibration

Publisher Summary The chapter describes a speech production model considering factors such as the impedance of the velum, the inflation of the vocal tract volume, and the effects of the subglottal system. Based on the measurements of acoustic waveforms and mechanical vibrations at several points of the vocal organ, a speech production model is proposed having oral nasal coupling through the velum even when it closes. It is assumed that the velum is composed of two vibrating plates connected to each other by a spring and a mechanical resistance. A syllable/bi/ is synthesized and examined, as an example, by the model, considering the velum leakage as well as a small inflation of the vocal tract volume by the inner air pressure just before the mouth opening. Concerning the speech production model, which has both a subglottal structure and a supraglottal one, it is shown that the waveform of glottal flow is changed, and synthesized vowels have zeros in their spectra because of an interaction between the sub- and supraglottal structures through the small opening area due to incomplete vocal cords closure.

Helium speech database and some new aspect of diver’s speech

Speech is distorted in pressurized helium‐air. Many aspects, except for upward shifts of formant and pitch in the helium speech, however, have not been known yet. A helium speech database system including more than 5 000 words and messages of divers was developed. The vibration of the cheek, the neck, and the nose were also recorded for some utterances. They were taken at the simulation dives up to the depth of 300 m at JAMSTEC since 1982. The helium speech and the normal speech were compared in terms of (1) duration of phonemes, and (2) pitch frequencies and formant trajectories in two different hearing/speaking conditions: the first condition is a case when a diver spoke as he was listening to his own unscrambled speech, and in the other condition the diver was not hearing his unscrambled speech. As a result, the shorter duration of word‐initial consonants, higher pitch by psychological stress, and rough movement of formant trajectories were observed in addition to the well‐known formant and pitch frequ...

MRI measurement and 3‐D FEM analysis of the nasal and paranasal cavities

Morphological measurement of the nasal and paranasal cavities were performed by means of the magnetic resonance imaging (MRI) technique. Three‐dimensional acoustic tube models were constructed using the MRI data. Using the finite element method (FEM), the Helmholtz equation was solved for sinusoidal pressure wave input at the velopharyngeal port to obtain sound pressure, particle velocity, and sound intensity in the nasal cavity. In the FEM modeling, the nasal tract containing the paranasal sinuses is divided into 3470 meshes and 4625 nodes. The cross‐sectional shape of the tract was approximated by an elliptical shape whose area and circumference were matched to the observed data. The frequency transfer function as a ratio of sound pressure at the velopharyngeal port to that at the nostrils, and input impedance at the velopharyngeal port were calculated and compared with those of the classical electric circuit model of the nasal tract. A FEM model incorporating shape asymmetry between the left and right ...