Vocal tract resonances and the sound of the Australian didjeridu (yidaki) I. experiment

The didjeridu, or yidaki, is a simple tube about 1.5 m long, played with the lips, as in a tuba, but mostly producing just a tonal, rhythmic drone sound. The acoustic impedance spectra of performers' vocal tracts were measured while they played and compared with the radiated sound spectra. When the tongue is close to the hard palate, the vocal tract impedance has several maxima in the range 1-3 kHz. These maxima, if sufficiently large, produce minima in the spectral envelope of the sound because the corresponding frequency components of acoustic current in the flow entering the instrument are small. In the ranges between the impedance maxima, the lower impedance of the tract allows relatively large acoustic current components that correspond to strong formants in the radiated sound. Broad, weak formants can also be observed when groups of even or odd harmonics coincide with bore resonances. Schlieren photographs of the jet entering the instrument and high speed video images of the player's lips show that the lips are closed for about half of each cycle, thus generating high levels of upper harmonics of the lip frequency. Examples of the spectra of "circular breathing" and combined playing and vocalization are shown.     

Do trumpet players tune resonances of the vocal tract?

The acoustic impedance spectrum was measured in the mouths of seven trumpeters while they played normal notes and while they practiced "bending" the pitch below or above the normal value. The peaks in vocal tract impedance usually had magnitudes rather smaller than those of the bore of the trumpet. Over the range measured, none of the trumpeters showed systematic tuning of the resonances of the vocal tract. However, all players commented that the presence of the impedance head in the mouth prevented them from playing the very highest notes of which they were normally capable. It is therefore possible that these players might use either resonance tuning or perhaps very high impedance magnitudes for some notes beyond the measured range. The observed lack of tuning contrasts with measurements for the saxophone which, like the trumpet, has weak resonances in the third and fourth octaves. Saxophonists are only able to play the highest range by tuning resonances of the vocal tract, so that the series impedance has a very strong peak at a frequency near that of the desired note. This difference is explained by the greater control that the trumpet player has over the natural frequency of the vibrating valve.     

Some insight into the acoustics of the didjeridu

The Australian didjeridu is a unique and interesting instrument. Despite the fact that the bore shape is almost random in nature and varies considerably across different instruments, the didjeridu timbre is readily recognizable. This is also true despite the fact that the player can manipulate the timbre more than in most wind instruments, by changing the shape of his vocal tract. In this study we examine the didjeridu spectrum in detail, in order to determine the characteristics that are similar across different instruments, those that are constant for a given instrument, and those that are readily influenced by the player. To this end we recorded and analyzed the sounds of eight instruments of different quality, all of them played across a range of timbres. Examining the resultant spectra, along with the resonance frequencies of these instruments, leads to a number of interesting conclusions. One of these is that the random nature of the instrument bore is actually conducive to creating its typical timbre. We also give a preliminary explanation of the differences between good and poor instruments.     

A methodological and preliminary study on the acoustic effect of a trumpet player's vocal tract

A methodological study is presented to examine the acoustic role of the vocal tract in playing the trumpet. Preliminary results obtained for one professional player are also shown to demonstrate the effectiveness of the method. Images of the vocal tract with a resolution of 0.5 mm (2 mm in thickness) were recorded with magnetic resonance imaging to observe the tongue posture and estimate the vocal-tract area function during actual performance. The input impedance was then calculated for the player's air column including both the supra- and subglottal tracts using an acoustic tube model including the effect of wall losses. Finally, a time-domain blowing simulation by Adachi and Sato [J. Acoust. Soc. Am. 99, 1200-1209 (1996)] was performed with a model of the lips. In this simulation, the oscillating frequency of the lips was slightly affected by using different shapes of the vocal tract measured for the player. In particular, when the natural frequency of the lips was gradually increased, the transition to the higher mode occurred at different frequencies for different vocal-tract shapes. Furthermore, simulation results showed that the minimum blowing pressure required to attain the lip oscillation can be reduced by adjusting the vocal-tract shape properly.     

A simplified model of mouth radiation impedance closed by mask cavity

Acoustic radiation impedance of the mouth is an important parameter when the vocal tract is modelled by the equivalent electrical circuit. If the vocal tract is closed by a cavity, as when the speaker wears some kind of mask, total impedance acoustically loading the vocal tract becomes serial connection of the mouth radiation impedance and the mask impedance. In that case the mouth radiation impedance has to be changed compared to free field conditions. This paper introduces a simplified approach to the modelling of that change by an appropriate reduction coefficient. The analysis based on an experiment preformed by measurement in the vocal tract physical model accompanied with analytical estimation has shown that the value of such reduction coefficient is 0.5. The results reveal that for a vocal tract closed with mask cavity the change in mouth radiation impedance introduced in an equivalent electrical circuit can be approximated by the value for free field radiation decreased by about 50%.     

Towards the Automatic Study of the Vocal Tract From Magnetic Resonance Images

Over the last few decades, researchers have been investigating the mechanisms involved in speech production. Image analysis can be a valuable aid in the understanding of the morphology of the vocal tract. The application of magnetic resonance imaging to study these mechanisms has been proven to be reliable and safe. We have applied deformable models in magnetic resonance images to conduct an automatic study of the vocal tract; mainly, to evaluate the shape of the vocal tract in the articulation of some European Portuguese sounds, and then to successfully automatically segment the vocal tract's shape in new images. Thus, a point distribution model has been built from a set of magnetic resonance images acquired during artificially sustained articulations of 21 sounds, which successfully extracts the main characteristics of the movements of the vocal tract. The combination of that statistical shape model with the gray levels of its points is subsequently used to build active shape models and active appearance models. Those models have then been used to segment the modeled vocal tract into new images in a successful and automatic manner. The computational models have thus been revealed to be useful for the specific area of speech simulation and rehabilitation, namely to simulate and recognize the compensatory movements of the articulators during speech production.     

Some roles of the vocal tract in clarinet breath attacks: natural sounds analysis and model-based synthesis

A simplified physical model mainly devoted to the reproduction of some transients of clarinet-like instruments is presented. From time-frequency analyses of natural clarinet sounds, it is shown that the vocal tract can play a significant role in some attacks as well as in the permanent regime. The model proposed consists in supplying a pressure source at the entrance of a cylindrical bore attached to the mouthpiece, allowing one to reach various vocal tract configurations. For real-time synthesis purposes, a digital scheme solving the physical problem is proposed. It is shown that this synthesis model is able to reproduce some of the complex features observed during the attacks of the natural sounds analyzed, as well as known effects of the vocal tract in permanent regime.     

A theoretical study of f0-f1 interaction with application to resonant speaking and singing voice

An interactive source-filter system, consisting of a three-mass body-cover model of the vocal folds and a wave reflection model of the vocal tract, was used to test the dependence of vocal fold vibration on the vocal tract. The degree of interaction is governed by the epilarynx tube, which raises the vocal tract impedance to match the impedance of the glottis. The key component of the impedance is inertive reactance. Whenever there is inertive reactance, the vocal tract assists the vocal folds in vibration. The amplitude of vibration and the glottal flow can more than double, and the oral radiated power can increase up to 10 dB. As F0 approaches F1, the first formant frequency, the interactive source-filter system loses its advantage (because inertive reactance changes to compliant reactance) and the noninteractive system produces greater vocal output. Thus, from a voice training and control standpoint, there may be reasons to operate the system in either interactive and noninteractive modes. The harmonics 2F0 and 3F0 can also benefit from being positioned slightly below F1.     

Discrimination of speaker size from syllable phrases

The length of the vocal tract is correlated with speaker size and, so, speech sounds have information about the size of the speaker in a form that is interpretable by the listener. A wide range of different vocal tract lengths exist in the population and humans are able to distinguish speaker size from the speech. Smith et al. [J. Acoust. Soc. Am. 117, 305-318 (2005)] presented vowel sounds to listeners and showed that the ability to discriminate speaker size extends beyond the normal range of speaker sizes which suggests that information about the size and shape of the vocal tract is segregated automatically at an early stage in the processing. This paper reports an extension of the size discrimination research using a much larger set of speech sounds, namely, 180 consonant-vowel and vowel-consonant syllables. Despite the pronounced increase in stimulus variability, there was actually an improvement in discrimination performance over that supported by vowel sounds alone. Performance with vowel-consonant syllables was slightly better than with consonant-vowel syllables. These results support the hypothesis that information about the length of the vocal tract is segregated at an early stage in auditory processing.     

How do clarinet players adjust the resonances of their vocal tracts for different playing effects?

In a simple model, the reed of the clarinet is mechanically loaded by the series combination of the acoustical impedances of the instrument itself and of the player's airway. Here we measure the complex impedance spectrum of players' airways using an impedance head adapted to fit inside a clarinet mouthpiece. A direct current shunt with high acoustical resistance allows players to blow normally, so the players can simulate the tract condition under playing conditions. The reproducibility of the results suggest that the players' "muscle memory" is reliable for this task. Most players use a single, highly stable vocal tract configuration over most of the playing range, except for the altissimo register. However, this "normal" configuration varies substantially among musicians. All musicians change the configuration, often drastically for "special effects" such as glissandi and slurs: the tongue is lowered and the impedance magnitude reduced when the player intends to lower the pitch or to slur downwards, and vice versa.     

Vocal tract resonances and the sound of the Australian didjeridu (yidaki). III. Determinants of playing quality

Traditional didjeridus have a broad range of bore geometries with many details not immediately apparent to a player, and are therefore suitable for examining the relationship between perceived quality and physical properties. Seven experienced players assessed the overall playing quality of 38 didjeridus that spanned a wide range of quality, pitch, and geometry, as well as 11 plastic cylindrical pipes. The ranking of these instruments was correlated with detailed measurements of their acoustic input impedance spectra. Most significantly, the ranked quality of a didjeridu was found to be negatively correlated with the magnitude of its acoustic input impedance, particularly in the frequency range from 1 to 2 kHz. This is in accord with the fact that maxima in the impedance of the player's vocal tract can inhibit acoustic flow, and consequently sound production, once the magnitude of these impedance maxima becomes comparable with or greater than those of the instrument. This produces the varying spectral peaks or formants in the sound envelope that characterize this instrument. Thus an instrument with low impedance and relatively weak impedance maxima in this frequency range would allow players greater control of the formants in the output sound and thus lead to a higher perceived playing quality.     

Perception of acoustic scale and size in musical instrument sounds

There is size information in natural sounds. For example, as humans grow in height, their vocal tracts increase in length, producing a predictable decrease in the formant frequencies of speech sounds. Recent studies have shown that listeners can make fine discriminations about which of two speakers has the longer vocal tract, supporting the view that the auditory system discriminates changes on the acoustic-scale dimension. Listeners can also recognize vowels scaled well beyond the range of vocal tracts normally experienced, indicating that perception is robust to changes in acoustic scale. This paper reports two perceptual experiments designed to extend research on acoustic scale and size perception to the domain of musical sounds: The first study shows that listeners can discriminate the scale of musical instrument sounds reliably, although not quite as well as for voices. The second experiment shows that listeners can recognize the family of an instrument sound which has been modified in pitch and scale beyond the range of normal experience. We conclude that processing of acoustic scale in music perception is very similar to processing of acoustic scale in speech perception.     

Mammalian laryngseal air sacs add variability to the vocal tract impedance: physical and computational modeling

Cavities branching off the main vocal tract are ubiquitous in nonhumans. Mammalian air sacs exist in human relatives, including all four great apes, but only a substantially reduced version exists in humans. The present paper focuses on acoustical functions of the air sacs. The hypotheses are investigated on whether the air sacs affect amplitude of utterances and/or position of formants. A multilayer synthetic model of the vocal folds coupled with a vocal tract model was utilized. As an air sac model, four configurations were considered: open and closed uniform tube-like side branches, a rigid cavity, and an inflatable cavity. Results suggest that some air sac configurations can enhance the sound level. Furthermore, an air sac model introduces one or more additional resonance frequencies, shifting formants of the main vocal tract to some extent but not as strongly as previously suggested. In addition, dynamic range of vocalization can be extended by the air sacs. A new finding is also an increased variability of the vocal tract impedance, leading to strong nonlinear source-filter interaction effects. The experiments demonstrated that air-sac-like structures can destabilize the sound source. The results were validated by a transmission line computational model.     

Acoustic analysis of trill sounds

In this paper, the acoustic-phonetic characteristics of steady apical trills--trill sounds produced by the periodic vibration of the apex of the tongue--are studied. Signal processing methods, namely, zero-frequency filtering and zero-time liftering of speech signals, are used to analyze the excitation source and the resonance characteristics of the vocal tract system, respectively. Although it is natural to expect the effect of trilling on the resonances of the vocal tract system, it is interesting to note that trilling influences the glottal source of excitation as well. The excitation characteristics derived using zero-frequency filtering of speech signals are glottal epochs, strength of impulses at the glottal epochs, and instantaneous fundamental frequency of the glottal vibration. Analysis based on zero-time liftering of speech signals is used to study the dynamic resonance characteristics of vocal tract system during the production of trill sounds. Qualitative analysis of trill sounds in different vowel contexts, and the acoustic cues that may help spotting trills in continuous speech are discussed.     

Cyclicity of laryngeal cavity resonance due to vocal fold vibration

Acoustic effects of the time-varying glottal area due to vocal fold vibration on the laryngeal cavity resonance were investigated based on vocal tract area functions and acoustic analysis. The laryngeal cavity consists of the vestibular and ventricular parts of the larynx, and gives rise to a regional acoustic resonance within the vocal tract, with this resonance imparting an extra formant to the vocal tract resonance pattern. Vocal tract transfer functions of the five Japanese vowels uttered by three male subjects were calculated under open- and closed-glottis conditions. The results revealed that the resonance appears at the frequency region from 3.0 to 3.7 kHz when the glottis is closed and disappears when it is open. Real spectra estimated from open- and closed-glottis periods of vowel sounds also showed the on-off pattern of the resonance within a pitch period. Furthermore, a time-domain acoustic analysis of vowels indicated that the resonance component could be observed as a pitch-synchronized rise-and-fall pattern of the bandpass amplitude. The cyclic nature of the resonance can be explained as the laryngeal cavity acting as a closed tube that generates the resonance during a closed-glottis period, but damps the resonance off during an open-glottis period.     

Mathematical model of acoustic speech production with mobile walls of the vocal tract

A mathematical speech production model is considered that describes acoustic oscillation propagation in a vocal tract with mobile walls. The wave field function satisfies the Helmholtz equation with boundary conditions of the third kind (impedance type). The impedance mode corresponds to a threeparameter pendulum oscillation model. The experimental research demonstrates the nonlinear character of how the mobility of the vocal tract walls influence the spectral envelope of a speech signal.     

The processing and perception of size information in speech sounds

There is information in speech sounds about the length of the vocal tract; specifically, as a child grows, the resonators in the vocal tract grow and the formant frequencies of the vowels decrease. It has been hypothesized that the auditory system applies a scale transform to all sounds to segregate size information from resonator shape information, and thereby enhance both size perception and speech recognition [Irino and Patterson, Speech Commun. 36, 181-203 (2002)]. This paper describes size discrimination experiments and vowel recognition experiments designed to provide evidence for an auditory scaling mechanism. Vowels were scaled to represent people with vocal tracts much longer and shorter than normal, and with pitches much higher and lower than normal. The results of the discrimination experiments show that listeners can make fine judgments about the relative size of speakers, and they can do so for vowels scaled well beyond the normal range. Similarly, the recognition experiments show good performance for vowels in the normal range, and for vowels scaled well beyond the normal range of experience. Together, the experiments support the hypothesis that the auditory system automatically normalizes for the size information in communication sounds.     

A Novel Source-Filter Stochastic Model for Voice Production

The novel stochastic model to produce voiced sounds proposed in this paper uses the source-filter Fant theory to generate voice signals and, consequently, it does not consider the coupling between the vocal tract and the vocal folds. Two novelties are proposed in the paper. The first one is the new model obtained from the unification of two other deterministic one mass-spring-damper models obtained from the literature and the second one is to build a stochastic model which can generate and control the level of jitter resulting even in hoarse voice signals or with pathological characteristics but using a simpler model than those ones discussed in the literature. An inverse stochastic problem is then solved for two cases, considering a normal voice and other obtained from a case of paralysis on the vocal folds. The parameters of the model are identified in the two cases allowing the validation of the model.     

Acoustic impedance of an artificially lengthened and constricted vocal tract

Voice training techniques often make use of exercises involving partial occlusion of the vocal tract, typically at the anterior part of the oral cavity or at the lips. In this study two techniques are investigated: a bilabial fricative and a small diameter hard-walled tube placed between the lips. Because the input acoustic impedance of the vocal tract is known to affect both the shaping of the glottal flow pulse and the vibrational pattern of the vocal folds, a study of the input impedance is an essential step in understanding the benefits of these two techniques. The input acoustic impedance of the vocal tract was investigated theoretically for cases of a vowel, bilabial occlusion (fully closed lips), a bilabial fricative, and artificially lengthening the tract with small diameter tubes. The results indicate that the tubes increase the input impedance in the range of the fundamental frequency of phonation by lowering the first formant frequency to nearly that of the bilabial occlusion (the lower bound on the first formant) while still allowing a continuous airflow. The bilabial fricative also has the effect of lowering the first formant frequency and increasing the low-frequency impedance, but not as effectively as the extension tubes.     

Two-dimensional model of vocal fold vibration for sound synthesis of voice and soprano singing

Voiced sounds were simulated with a computer model of the vocal fold composed of a single mass vibrating both parallel and perpendicular to the airflow. Similarities with the two-mass model are found in the amplitudes of the glottal area and the glottal volume flow velocity, the variation in the volume flow waveform with the vocal tract shape, and the dependence of the oscillation amplitude upon the average opening area of the glottis, among other similar features. A few dissimilarities are also found in the more symmetric glottal and volume flow waveforms in the rising and falling phases. The major improvement of the present model over the two-mass model is that it yields a smooth transition between oscillations with an inductive load and a capacitive load of the vocal tract with no sudden jumps in the vibration frequency. Self-excitation is possible both below and above the first formant frequency of the vocal tract. By taking advantage of the wider continuous frequency range, the two-dimensional model can successfully be applied to the sound synthesis of a high-pitched soprano singing, where the fundamental frequency sometimes exceeds the first formant frequency.