Acoustic Variables Research Articles

Sound Visualization Demonstrates Velopharyngeal Coupling and Complex Spectral Variability in Asian Elephants.

Simple SummaryWhether a vocalization is emitted though the mouth or nose impacts on its acoustic shape through specific vocal tract resonances. In human language, nasalized compared to oral vowels can change the meaning of words. African elephants mainly use low-frequency rumbles to communicate and can utter them either through the mouth or trunk. In this study, we used an acoustic camera to visualize the sound emission of rumbles in Asian elephants, which have been studied considerably less than African elephants. We recorded nine captive female Asian elephants and analyzed the acoustic structure of 203 calls. We found that most rumbles (64%) were emitted through the trunk, 21% through the mouth, and surprisingly, 13% simultaneously through the mouth and trunk. Some of the rumbles were combined with orally emitted roars. The nasal rumbles concentrated most spectral energy in lower frequencies, whereas the oral and mixed rumbles showed a broader spectral energy distribution and concentration in higher frequencies, and were louder. The roars were the loudest, broadest and highest in frequency. This study is the first to demonstrate coupled oral and nasal emission in a non–human animal, thereby setting an important framework to study the functions of this acoustic variability in elephant communication and the evolution of vocal flexibility across species.Sound production mechanisms set the parameter space available for transmitting biologically relevant information in vocal signals. Low–frequency rumbles play a crucial role in coordinating social interactions in elephants’ complex fission–fusion societies. By emitting rumbles through either the oral or the three-times longer nasal vocal tract, African elephants alter their spectral shape significantly. In this study, we used an acoustic camera to visualize the sound emission of rumbles in Asian elephants, which have received far less research attention than African elephants. We recorded nine adult captive females and analyzed the spectral parameters of 203 calls, including vocal tract resonances (formants). We found that the majority of rumbles (64%) were nasally emitted, 21% orally, and 13% simultaneously through the mouth and trunk, demonstrating velopharyngeal coupling. Some of the rumbles were combined with orally emitted roars. The nasal rumbles concentrated most spectral energy in lower frequencies exhibiting two formants, whereas the oral and mixed rumbles contained higher formants, higher spectral energy concentrations and were louder. The roars were the loudest, highest and broadest in frequency. This study is the first to demonstrate velopharyngeal coupling in a non-human animal. Our findings provide a foundation for future research into the adaptive functions of the elephant acoustic variability for information coding, localizability or sound transmission, as well as vocal flexibility across species.

A robust method to estimate the coordinates of seafloor stations by direct-path ranging

The ranges derived from acoustic measurements between seafloor stations are relatively more accurate compared with those derived from the sea surface vessel transducer to the seafloor transponders, because measurements through mixed water layers will be affected by complex acoustic range errors. Coordinates of seafloor stations can be improved by the direct-path acoustic ranging. Systematic errors in acoustic rangings, however, will significantly deteriorate the accuracy of vertical coordinates. In order to mitigate the effects of these systematic errors (e.g., acoustic ray bending and sound speed variation errors in acoustic measurements on the seafloor station location parameters), the observation model needs to be finely constructed. First, a new observation model with acoustic ray bending and sound speed bias parameters is established. Then, using a seafloor geodetic network with four moored stations at a depth of about 3000 m in the South China Sea, the significance of the acoustic ray bending parameter is tested. The results show that (1) the acoustic ray bending parameter is significant at the 90% confidence level, which means that the acoustic ray bending error in the seafloor geodetic network is not negligible; (2) by estimating the coefficient of acoustic ray bending, the influence of the acoustic ray bending error on the vertical coordinate components can be significantly mitigated; our model improves the accuracy of the seafloor stations’ position with differences in the horizontal coordinate components less than 0.1 cm between the two-dimensional adjustment and three-dimensional adjustment, and also improves the vertical coordinate component to uncertainty less than 3.0 cm; (3) the relative movement between the moored stations is less than 50 cm, and the horizontal movement is larger than the vertical movement.

Investigation of acoustic waves behavior of an underground tunnel in a multilayer soil

Understanding the acoustic behavior of buried tunnels is valuable for locating them and monitoring their structure health. This research focuses on the acoustic behavior of buried tunnels in multilayer soil structures. The reflected and transmitted acoustic wave pressure variations are investigated exclusively for a multilayer soil buried tunnel. The tunnel system's 3D finite element model is presented, which contains the tunnel lining, surrounding soil, and the air inside the tunnel and at the ground surface. A free air explosion is used as the acoustic wave source. The reflected and transmitted waves' pressure values are measured to evaluate the effects of mechanical characteristics of soil layers, tunnel buried depths, and lining concrete types on the acoustic wave behavior of the tunnel. In addition, a utility line is introduced to the system in different positions related to the main tunnel to investigate its effect on the main tunnel’s acoustic wave behavior. The results indicate that in a multilayer soil structure, the relative position of the soil layers and the tunnel (whether the main tunnel or the utility line) significantly impacts the acoustic pressure value, particularly the transmitted wave pressure. When changing the tunnel buried depth and the lining concrete type, multiple pressure peaks are observed in reflected acoustic wave pressure–time history exclusive to a tunnel surrounded by a multilayer soil structure. The findings can be used to precisely interpret the recorded signals for structural health monitoring and locating underground structures, especially in a media with multilayer soil structures.

Phononic crystal based sensor to detect acoustic variations in methyl & ethyl nonafluorobutyl ether

Phononic crystals and metamaterials have been widely studied for wave manipulation applications. In this study, we propose a conceptual framework for a new type of acoustic bio-chemical sensor that works based on the principle of phononic crystals and metamaterials to detect the temperature and pressure changes in active solvents like Methyl Nonafluorobutyl Ether (MNE) and Ethyl Nonafluorobutyl Ether (ENE). First, a wide low-frequency bandgap is obtained from the proposed composite unit cell structure with trampoline effect. Then a defect is introduced to adjust the localize cavity modes inside the reported bandgap. Later, the cavity is filled with MNE and ENE solvents that eventually resulted into fluid-solid coupling physics. The numerical wave dispersion curves and transmission profiles show presence of Fano-like interference/resonance effect evident from the observation of asymmetrical transmission profile. Such robust asymmetrical transmission peak is generated due to coupling of incident waves with scattered wave field emitted from the MNE and ENE solvents upon excitation. The variation in acoustic properties of MNE and ENE caused by temperature and pressure fields on newly born Fano-like asymmetrical transmission profile is studied. The proposed acoustic bio-chemical sensor governed by Fano-interference effect efficiently capture the variation in acoustic properties of MNE and ENE solvents at relatively low-frequency regime that makes this approach favorable for sensing applications. Such smart acoustic bio-chemical sensors can have useful applications in pharmaceutical production, petrochemicals and capturing ingredients of cosmetic and beauty products.

Reducing frequency fluctuation in a Brillouin random fiber laser by a random fiber grating ring resonator.

Frequency fluctuation is a major problem in high-precision metrology as real-time optical frequency measurement is not available with commercial photodetectors; alternatively, frequency-stabilized lasers as a reference are also not accessible in most laboratories. In this study, we propose and demonstrate a polarization-maintaining random fiber grating ring (PM-RFGR) resonator in a PM Brillouin random fiber laser (BRFL) to achieve sub-MHz frequency drift, which is measured by the optical beat of the random laser and the pump laser. Experimental results show that longitudinal modes are suppressed in the BRFL owing to the feedback of the RFGR resonating with one longitudinal mode of the random laser. The BRFL shows mode-hopping-free operation over 14.9 s due to the self-adjustment of random modes with small frequency difference to thermal and acoustic variations and self-injection locking through RFGR. As a result, a small frequency drift of ∼340 kHz with single-longitudinal mode is achieved in the BRFL enabled by the RFGR, which offers an all optical locking mechanism for optical frequency stabilization.

Spelling provides a precise (but sometimes misplaced) phonological target. Orthography and acoustic variability in second language word learning

L1 French participants learned novel L2 English words over two days of learning sessions, with half of the words presented with their orthographic forms (Audio-Ortho) and half without (Audio only). One group heard the words pronounced by a single talker, while another group heard them pronounced by multiple talkers. On the third day, they completed a variety of tasks to evaluate their learning. Our results show a robust influence of orthography, with faster response times in both production (Picture naming) and recognition (Picture mapping) tasks for words learned in the Audio-Ortho condition. Moreover, formant analyses of the Picture naming responses show that orthographic input pulls pronunciations of English novel words towards a non-native (French) phonological target. Words learned with their orthographic forms were pronounced more precisely (with smaller Dispersion Scores), but were misplaced in the vowel space (as reflected by smaller Euclidian distances with respect to French vowels). Finally, we found only limited evidence of an effect of talker-based acoustic variability: novel words learned with multiple talkers showed faster responses times in the Picture naming task, but only in the Audio-only condition, which suggests that orthographic information may have overwhelmed any advantage of talker-based acoustic variability.

A Perceiver-Centered Approach for Representing and Annotating Prosodic Functions in Performed Music.

Musical prosody is characterized by the acoustic variations that make music expressive. However, few systematic and scalable studies exist on the function it serves or on effective tools to carry out such studies. To address this gap, we introduce a novel approach to capturing information about prosodic functions through a citizen science paradigm. In typical bottom-up approaches to studying musical prosody, acoustic properties in performed music and basic musical structures such as accents and phrases are mapped to prosodic functions, namely segmentation and prominence. In contrast, our top-down, human-centered method puts listener annotations of musical prosodic functions first, to analyze the connection between these functions, the underlying musical structures, and acoustic properties. The method is applied primarily to the exploring of segmentation and prominence in performed solo piano music. These prosodic functions are marked by means of four annotation types—boundaries, regions, note groups, and comments—in the CosmoNote web-based citizen science platform, which presents the music signal or MIDI data and related acoustic features in information layers that can be toggled on and off. Various annotation strategies are discussed and appraised: intuitive vs. analytical; real-time vs. retrospective; and, audio-based vs. visual. The end-to-end process of the data collection is described, from the providing of prosodic examples to the structuring and formatting of the annotation data for analysis, to techniques for preventing precision errors. The aim is to obtain reliable and coherent annotations that can be applied to theoretical and data-driven models of musical prosody. The outcomes include a growing library of prosodic examples with the goal of achieving an annotation convention for studying musical prosody in performed music.

Modeling Acoustic Channel Variability in Underwater Network Simulators from Real Field Experiment Data

The underwater acoustic channel is remarkably dependent on the considered scenario and the environmental conditions. In fact, channel impairments differ significantly in shallow water with respect to deep water, and the presence of external factors such as snapping shrimps, bubbles, rain, or ships passing nearby, changes of temperature, and wind strength can change drastically the link quality in different seasons and even during the same day. Legacy mathematical models that consider these factors exist, but are either not very accurate, like the Urick model, or very computationally demanding, like the Bellhop ray tracer. Deterministic models based on lookup tables (LUTs) of sea trial measurements are widely used by the research community to simulate the acoustic channel in order to verify the functionalities of a network in certain water conditions before the actual deployment. These LUTs can characterize the link quality by observing, for instance, the average packet error rate or even a time varying packet error rate computed within a certain time window. While this procedure characterizes well the acoustic channel, the obtained simulation results are limited to a single channel realization, making it hard to fully evaluate the acoustic network in different conditions. In this paper, we discuss the development of a statistical channel model based on the analysis of real field experiment data, and compare its performance with the other channel models available in the DESERT Underwater network simulator.

Cognitive Digital Biomarkers from Automated Transcription of Spoken Language

BackgroundAlthough patients with Alzheimer’s disease and other cognitive-related neurodegenerative disorders may benefit from early detection, development of a reliable diagnostic test has remained elusive. The penetration of digital voice-recording technologies and multiple cognitive processes deployed when constructing spoken responses might offer an opportunity to predict cognitive status.ObjectiveTo determine whether cognitive status might be predicted from voice recordings of neuropsychological testingDesignComparison of acoustic and (para)linguistic variables from low-quality automated transcriptions of neuropsychological testing (n = 200) versus variables from high-quality manual transcriptions (n = 127). We trained a logistic regression classifier to predict cognitive status, which was tested against actual diagnoses.SettingObservational cohort study.Participants146 participants in the Framingham Heart Study.MeasurementsAcoustic and either paralinguistic variables (e.g., speaking time) from automated transcriptions or linguistic variables (e.g., phrase complexity) from manual transcriptions.ResultsModels based on demographic features alone were not robust (area under the receiver-operator characteristic curve [AUROC] 0.60). Addition of clinical and standard acoustic features boosted the AUROC to 0.81. Additional inclusion of transcription-related features yielded an AUROC of 0.90.ConclusionsThe use of voice-based digital biomarkers derived from automated processing methods, combined with standard patient screening, might constitute a scalable way to enable early detection of dementia.

Lombard effect in Algerian Dialect Speech: Investigation of boosting and bypass strategies in narrowband noisy environment

The paper presents the first investigation on the Lombard effect (LE) in the Algerian Dialect (AD) speech. It is devoted to the examination of boosting and bypass strategies in narrowband noisy environments (0.49–6.8 kHz). Speakers who adopt the boosting tactic would attempt to improve their audibility in noise by increasing the global intensity (I), the local intensity, and the duration (D) of their speech. The boost of the local intensity consists in amplifying the spectral magnitude of the speech waveform in the frequency region wherein the masking noise is beyond 55 dB. However, speakers who apply the bypass strategy would raise their speech in noise so that their fundamental frequency (F0) and first Formant (F1) will decrease. We built the AD Lombard Speech Corpus (ADLSC) that consists of continuous speech files uttered by eight native speakers in quiet and noisy conditions. The corpus contains phonetically rich words that cover all AD phonemes. To examine these strategies, we assessed the change in I, D, F0 and F1 from quiet to noisy environments. Subsequently, we evaluated the change in the local intensity via the computation of the Power Spectrum Ratio (PSR). Statistical tests were conducted to confirm the significance of the acoustic analysis. We found out that speakers raised their vocal effort in I and D. Furthermore, the measured PSRs confirmed the evidence that LE is frequency-specific in humans. Similarly, we noticed that speakers increased their F0 and F1 in noisy conditions. The results we obtained confirmed the boosting strategy but denied the bypass tactic. We compared our findings with the results reported in the state-of-the-art studies. The comparison revealed that LE in narrowband does not exhibit the same acoustic variations in AD and English.

Context effects in perception of vowels differentiated by F1 are not influenced by variability in talkers' mean F1 or F3.

Spectral properties of earlier sounds (context) influence recognition of later sounds (target). Acoustic variability in context stimuli can disrupt this process. When mean fundamental frequencies (f0's) of preceding context sentences were highly variable across trials, shifts in target vowel categorization [due to spectral contrast effects (SCEs)] were smaller than when sentence mean f0's were less variable; when sentences were rearranged to exhibit high or low variability in mean first formant frequencies (F1) in a given block, SCE magnitudes were equivalent [Assgari, Theodore, and Stilp (2019) J. Acoust. Soc. Am. 145(3), 1443-1454]. However, since sentences were originally chosen based on variability in mean f0, stimuli underrepresented the extent to which mean F1 could vary. Here, target vowels (/ɪ/-/ɛ/) were categorized following context sentences that varied substantially in mean F1 (experiment 1) or mean F3 (experiment 2) with variability in mean f0 held constant. In experiment 1, SCE magnitudes were equivalent whether context sentences had high or low variability in mean F1; the same pattern was observed in experiment 2 for new sentences with high or low variability in mean F3. Variability in some acoustic properties (mean f0) can be more perceptually consequential than others (mean F1, mean F3), but these results may be task-dependent.

Relative acoustic nonlinearity parameter variation with local defect resonance frequencies and its different harmonics

In recent times, nonlinear ultrasonic inspection has become a very reliable technique to characterize micro-scale cracks and small-scale delaminations. In this paper, an analytical framework is developed for the orthotropic composite plate to calculate the variation of relative acoustic nonlinearity parameter, a quantitative measurement of nonlinearity at the delamination region. The analytical investigation is justified with numerical and experimental research using a carbon fiber-reinforced polymer (CFRP) plate with three delaminations. Two incident waves are used for excitation, out of which the first wave frequency is varied as local defect resonance (LDR) frequency ( fd), subharmonic LDR frequency (2 fd), second-order superharmonic LDR frequency ( fd/2), and third-order superharmonic LDR frequency ( fd/3) while the second incident wave is taken as a single periodic frequency. Different delamination sizes, positions, and depths were used to understand the second-order relative acoustic nonlinearity parameter ( β2) with other LDR frequencies as well as its harmonics. It has been observed from analytical, numerical, and experimental results that the relative acoustic nonlinearity parameter is maximum at LDR frequencies. It increases in case of surface level delaminations.

Sound variation by hypertrophy and atrophy sonic muscle in the male southern black drum (Pogonias courbina).

The male and female southern black drum possess highly specialized, bilateral, striated sonic muscles used in sound production during courtship by males and in the production of disturbance calls by both males and females. Androgen-driven hypertrophy of the sonic muscles in males during the late spring spawning period results in increased growth of sonic muscle mass followed by post-spawning atrophy of sonic muscles. We examined changes in sonic muscle morphology and in the sound characteristics of males and females underlying seasonal changes in sonic muscle mass and muscle contraction as measured by sound production. In males, the sound pressure level increased while sound pulse duration decreased with increasing sonic muscle mass, indicating that sonic muscle fibers contract with greater force and shorter duration during the spawning season. Interpulse interval and the total number of pulses varied seasonally with muscle mass, which suggests that the effects of steroids on male southern black drum sound characteristics are more pronounced peripherally than in the central nervous system. In females, no increase in sonic muscle mass was found, and therefore, a change was not observed in the acoustic variables analyzed. Seasonal sonic muscle hypertrophy in males likely functions as a secondary sexual characteristic that maximizes vocalization amplitude or loudness during the spawning period.

Difficult to hear but easy to see: Audio-visual perception of the /r/-/w/ contrast in Anglo-English.

This paper investigates the influence of visual cues in the perception of the /r/-/w/ contrast in Anglo-English. Audio-visual perception of Anglo-English /r/ warrants attention because productions are increasingly non-lingual, labiodental (e.g., [ʋ]), possibly involving visual prominence of the lips for the post-alveolar approximant [ɹ]. Forty native speakers identified [ɹ] and [w] stimuli in four presentation modalities: auditory-only, visual-only, congruous audio-visual, and incongruous audio-visual. Auditory stimuli were presented in noise. The results indicate that native Anglo-English speakers can identify [ɹ] and [w] from visual information alone with almost perfect accuracy. Furthermore, visual cues dominate the perception of the /r/-/w/ contrast when auditory and visual cues are mismatched. However, auditory perception is ambiguous because participants tend to perceive both [ɹ] and [w] as /r/. Auditory ambiguity is related to Anglo-English listeners' exposure to acoustic variation for /r/, especially to [ʋ], which is often confused with [w]. It is suggested that a specific labial configuration for Anglo-English /r/ encodes the contrast with /w/ visually, compensating for the ambiguous auditory contrast. An audio-visual enhancement hypothesis is proposed, and the findings are discussed with regard to sound change.

Acoustic Needs of Older People in the Outdoor Environment of Residential Communities: A Case Study in China

This study examines the acoustic needs of older people in an outdoor environment to design outdoor spaces according to their needs. Using the grounded theory approach, we selected and interviewed 16 older people from four residential communities in China. The acoustic needs of older people in an outdoor environment are established in terms of sound preference, quietness, acoustic variability, acoustic comfort, subjective loudness, sound rhythm, sound clarity, and sound sensitivity. The findings show that older people (a) attach importance to quietness when considering sound preference, (b) prefer taking a walk in an overall quiet environment, although sometimes a lively environment, (c) chose to conduct their activities in an environment with low variation in sound, (d) rate the comfort of their overall acoustic environment highly, (e) show little need for subjective loudness as long as the sounds are not ones they dislike, (f) generally prefer relatively rhythmic sound sources, (g) did not have a significant need for sound clarity, and (h) are tolerant of outdoor acoustic environments and did not pay too much attention to certain sounds. These findings can provide direction for better outdoor space designs in communities, to suit older people, and to create age-friendly cities.

Vowel sound production and its association with cephalometric characteristics in skeletal Class III subjects.

This study aimed to evaluate differences in vowel production using acoustic analysis in skeletal Class III and Class I Japanese participants and to identify the correlation between vowel sounds and cephalometric variables in skeletal Class III subjects. Japanese males with skeletal Class III (ANB < 0°) and Class I skeletal anatomy (0.62° < ANB < 5.94°) were recruited (n = 18/group). Acoustic analysis of vowel sounds and cephalometric analysis of lateral cephalograms were performed. For sound analysis, an isolated Japanese vowel (/a/,/i/,/u/,/e/,/o/) pattern was recorded. Praat software was used to extract acoustic parameters such as fundamental frequency (F0) and the first four formants (F1, F2, F3, and F4). The formant graph area was calculated. Cephalometric values were obtained using ImageJ. Correlations between acoustic and cephalometric variables in skeletal Class III subjects were then investigated. Skeletal Class III subjects exhibited significantly higher/o/F2 and lower/o/F4 values. Mandibular length, SNB, and overjet of Class III subjects were moderately negatively correlated with acoustic variables. This study did not take into account vertical skeletal patterns and tissue movements during sound production. Skeletal Class III males produced different /o/ (back and rounded vowel), possibly owing to their anatomical positions or adaptive changes. Vowel production was moderately associated with cephalometric characteristics of Class III subjects. Thus, changes in speech after orthognathic surgery may be expected. A multidisciplinary team approach that included the input of a speech pathologist would be useful.

Broadband low-frequency sound attenuation in duct with embedded periodic sonic black holes

This paper studies the sound wave propagation and attenuation characteristics in a new type of acoustic device, which is constructed by periodically embedding sonic black holes (SBHs) into an air-filled waveguide. The SBH is realized by cascading a set of rigid rings with linearly or quadratically changing inner radii into each unit cell section. Acoustic impedance variation rendered by this specially tailored profile progressively slows down the incident sound velocity to generate the SBH effects including wavelength compression, energy focalization, and dissipation with inherent damping in the duct. Upon developing and validating a model based on the transfer matrix method (TMM), the sound transmission loss (STL) of a finite-length periodic duct with different geometrical parameters is investigated. Different from a single SBH unit which can trigger the complete SBH process only above the cut-on frequency, the proposed periodic arrangement offers strong attenuation bands in the low-frequency range of the STL curve well below the frequency barrier imposed by the cut-on frequency of a single SBH unit. Mechanism studies reveal that these attenuation bands result from the interplay between the Bragg scattering and the SBH-specific slow-sound effects. The latter is shown to equivalently increase the effective lattice constant of the periodic duct, thus lowering the frequency of the Bragg bandgap. It is also shown that in contrast to conventional locally-resonant metamaterials, the proposed device exhibits far less dependence on the number of unit cells and the amount of rigid rings to activate a significant low-frequency stop-band. These appealing features could greatly simplify the design and application of SBH-based metamaterials, making it a promising solution for low-frequency wave manipulation and noise control.

Distress vocalizations in five species of armadillos (Xenarthra, Cingulata)

Abstract Distress vocalizations are emitted by animals experiencing extreme physical distress, such as when caught by a predator. These signals are emitted by numerous and phylogenetically distant vertebrate species and are composed of sequences of broadband and high-amplitude notes. In this study, we provide the first acoustic characterization of distress vocalizations in four armadillo species: pink fairy armadillo (Chlamyphorus truncatus), greater fairy armadillo (Calyptophractus retusus), Southern three-banded armadillo (Tolypeutes matacus), and Chacoan naked-tailed armadillo (Cabassous chacoensis). We also recharacterized the weeping call of the screaming hairy armadillo (Chaetophractus vellerosus) to compare vocalizations, discuss potential homologies, and examine possible causes of structural and acoustic similarities among these species. In three species the vocalizations were sequences of exhaled notes that differed in their fine spectral structure (exhaled harmonic notes in C. retusus, and exhaled harsh notes in C. truncatus and T. matacus). The vocalization of C. chacoensis was composed of exhaled harsh and inhaled harsh notes that occurred alternately and continuously in a quick sequence. Based on the mode of production and acoustic similarity, we propose that the notes of C. retusus and C. truncatus would be homologous to the conspicuous crying notes of C. vellerosus. The exhaled harsh notes of T. matacus and C. chacoensis may also be homologous to the crying notes of C. vellerosus, but the notes of T. matacus are quite different in various acoustic parameters. Furthermore, the inhaled and exhaled harsh notes of C. chacoensis are similar to the inhaled and exhaled sobbing notes of C. vellerosus, making assessments of homologies uncertain in these species. Because a common motivational state (physical distress) underlies these vocalizations, we propose that the notable differences in body size and habitat preferences of the armadillos could represent potential drivers of the acoustic divergence among their vocalizations.

Developing sequentially trained robust Punjabi speech recognition system under matched and mismatched conditions

Development of a native language robust ASR framework is very challenging as well as an active area of research. Although an urge for investigation of effective front-end as well as back-end approaches are required for tackling environment differences, large training complexity and inter-speaker variability in achieving success of a recognition system. In this paper, four front-end approaches: mel-frequency cepstral coefficients (MFCC), Gammatone frequency cepstral coefficients (GFCC), relative spectral-perceptual linear prediction (RASTA-PLP) and power-normalized cepstral coefficients (PNCC) have been investigated to generate unique and robust feature vectors at different SNR values. Furthermore, to handle the large training data complexity, parameter optimization has been performed with sequence-discriminative training techniques: maximum mutual information (MMI), minimum phone error (MPE), boosted-MMI (bMMI), and state-level minimum Bayes risk (sMBR). It has been demonstrated by selection of an optimal value of parameters using lattice generation, and adjustments of learning rates. In proposed framework, four different systems have been tested by analyzing various feature extraction approaches (with or without speaker normalization through Vocal Tract Length Normalization (VTLN) approach in test set) and classification strategy on with or without artificial extension of train dataset. To compare each system performance, true matched (adult train and test—S1, child train and test—S2) and mismatched (adult train and child test—S3, adult + child train and child test—S4) systems on large adult and very small Punjabi clean speech corpus have been demonstrated. Consequently, gender-based in-domain data augmented is used to moderate acoustic and phonetic variations throughout adult and children’s speech under mismatched conditions. The experiment result shows that an effective framework developed on PNCC + VTLN front-end approach using TDNN-sMBR-based model through parameter optimization technique yields a relative improvement (RI) of 40.18%, 47.51%, and 49.87% in matched, mismatched and gender-based in-domain augmented system under typical clean and noisy conditions, respectively.

A Situational Analysis of Current Speech-Synthesis Systems for Child Voices: A Scoping Review of Qualitative and Quantitative Evidence

(1) Background: Speech synthesis has customarily focused on adult speech, but with the rapid development of speech-synthesis technology, it is now possible to create child voices with a limited amount of child-speech data. This scoping review summarises the evidence base related to developing synthesised speech for children. (2) Method: The included studies were those that were (1) published between 2006 and 2021 and (2) included child participants or voices of children aged between 2–16 years old. (3) Results: 58 studies were identified. They were discussed based on the languages used, the speech-synthesis systems and/or methods used, the speech data used, the intelligibility of the speech and the ages of the voices. Based on the reviewed studies, relative to adult-speech synthesis, developing child-speech synthesis is notably more challenging. Child speech often presents with acoustic variability and articulatory errors. To account for this, researchers have most often attempted to adapt adult-speech models, using a variety of different adaptation techniques. (4) Conclusions: Adapting adult speech has proven successful in child-speech synthesis. It appears that the resulting quality can be improved by training a large amount of pre-selected speech data, aided by a neural-network classifier, to better match the children’s speech. We encourage future research surrounding individualised synthetic speech for children with CCN, with special attention to children who make use of low-resource languages.