Changes In Formant Frequencies Research Articles

Lossy coding and bitrate effects on changes in formant frequencies in Japanese and English speech signals

Lossy coding and bitrate effects on changes in formant frequencies in Japanese and English speech signals

Lifespan anatomic and vowel acoustic studies confirm prepubertal sexual dimorphism of the human vocal tract and absence of lengthening in aging

To study anatomic-acoustic relations across the lifespan, anatomic studies of the developing vocal tract (VT) using medical imaging studies were related to vowel acoustic space as synthesized from published data. However, the aggregate acoustic data highlighted the need to acquire sex-specific vowel acoustic data across the lifespan using controlled methodology for data collection and analysis; also, to acquire measurements of the higher formants. Lifespan anatomic and acoustic (4-to-92 years) data helped resolve two questions. First, anatomic studies revealed significant prepubertal sex differences in the oral region of the VT (3-to-7 years) that are masked by growth rate differences between males and females. Three-dimensional anatomic findings also revealed prepubertal sexual dimorphism in select regions of the mandible and pharynx. Second, anatomic findings confirmed that VT length does not increase with aging. Acoustic studies similarly confirmed sex differences in fundamental frequency emerging at age 7; and the presence of significant prepubertal sexual dimorphism of the higher formant frequencies (F3-F4) at the earliest age studied (4 years). Findings for adults showed significant age-related decreases in fundamental frequency in women only, and no changes in formant frequencies in either sex across several decades. The results underscore the importance of sex-specific and non-uniform growth of the VT structures. [NIDCD funding support R01DC006282.]

Perception of age, gender, and talker height from children’s vowels

In his 1989 review of vowel perception [J. Acoust. Soc. Am. 85, 2088–2113], Terry Nearey observed that the acoustic properties of vowels vary substantially across talkers due to differences in talker size. Taller speakers tend to have lower formant frequencies and lower fundamental frequencies than shorter speakers, and listeners appear to take these relationships into account in vowel identification. In this talk, we follow up some of the questions raised in that paper, reviewing recent findings on the perception of age, gender, and talker height in children’s voices. Children exhibit substantial age-related changes in fundamental and formant frequencies due to growth of the larynx and vocal tract, presenting an interesting opportunity to explore the perceptual consequences of these changes. Listeners’ perceptual judgments of age and gender reveal complex interdependencies, consistent with the idea that these indexical attributes are jointly estimated, in a manner that reflects their shared dependence on acoustic parameters related to perceived talker height.

Acoustic and Perceptual Effects of Articulation Exercises in Transgender Women.

This study measured the impact of articulation exercises using a cork and articulation exercises for lip spreading on the formant frequencies of vowels and listener perceptions of femininity in transgender women. Thirteen transgender women were recorded before and after the cork exercise and before and after the lip spreading exercise. Speech samples included continuous speech during reading and were analyzed using Praat software. Vowel formant frequencies (F1, F2, F3, F4, F5) and vowel space were determined. A listening experiment was organized using naïve cisgender women and cisgender men rating audio samples of continuous speech. Masculinity/femininity, vocal quality and age were rated, using a visual analogue scale (VAS). Concerning vowel formant frequencies, F2 /a/ and F5 /u/ significantly increased after the lip spreading exercise, as well as F3 /a/, F3 /u/ and F4 /a/ after the cork exercise. The lip spreading exercise had more impact on the F2 /a/ than the cork exercise. Vowel space did not change after the exercises. The fundamental frequency (fo) increased simultaneously during both exercises. Both articulation exercises were associated with significantly increased listener perceptions of femininity of the voice. Subtle changes in formant frequencies can be observed after performing articulation exercises, but not in every formant frequency or vowel. Cisgender listeners rated the speech of the transgender women more feminine after the exercises. Further research with a more extensive therapy program and listening experiment is needed to examine these preliminary findings.

Auditory Models for Formant Frequency Discrimination of Vowel Sounds

As formant frequencies of vowel sounds are critical acoustic cues for vowel perception, human listeners need to be sensitive to formant frequency change. Numerous studies have found that formant frequency discrimination is affected by many factors like formant frequency, speech level, and fundamental frequency. Theoretically, to perceive a formant frequency change, human listeners with normal hearing may need a relatively constant change in the excitation and loudness pattern, and this internal change in auditory processing is independent of vowel category. Thus, the present study examined whether such metrics could explain the effects of formant frequency and speech level on formant frequency discrimination thresholds. Moreover, a simulation model based on the auditory excitation-pattern and loudness-pattern models was developed to simulate the auditory processing of vowel signals and predict thresholds of vowel formant discrimination. The results showed that predicted thresholds based on auditory metrics incorporating auditory excitation or loudness patterns near the target formant showed high correlations and low root-mean-square errors with human behavioral thresholds in terms of the effects of formant frequency and speech level). In addition, the simulation model, which particularly simulates the spectral processing of acoustic signals in the human auditory system, may be used to evaluate the auditory perception of speech signals for listeners with hearing impairments and/or different language backgrounds.

Vowel Formants in Normal and Loud Speech.

Purpose This study evaluated how 1st and 2nd vowel formant frequencies (F1, F2) differ between normal and loud speech in multiple speaking tasks to assess claims that loudness leads to exaggerated vowel articulation. Method Eleven healthy German-speaking women produced normal and loud speech in 3 tasks that varied in the degree of spontaneity: reading sentences that contained isolated /i: a: u:/, responding to questions that included target words with controlled consonantal contexts but varying vowel qualities, and a recipe recall task. Loudness variation was elicited naturalistically by changing interlocutor distance. First and 2nd formant frequencies and average sound pressure level were obtained from the stressed vowels in the target words, and vowel space area was calculated from /i: a: u:/. Results Comparisons across many vowels indicated that high, tense vowels showed limited formant variation as a function of loudness. Analysis of /i: a: u:/ across speech tasks revealed vowel space reduction in the recipe retell task compared to the other 2. Loudness changes for F1 were consistent in direction but variable in extent, with few significant results for high tense vowels. Results for F2 were quite varied and frequently not significant. Speakers differed in how loudness and task affected formant values. Finally, correlations between sound pressure level and F1 were generally positive but varied in magnitude across vowels, with the high tense vowels showing very flat slopes. Discussion These data indicate that naturalistically elicited loud speech in typical speakers does not always lead to changes in vowel formant frequencies and call into question the notion that increasing loudness is necessarily an automatic method of expanding the vowel space. Supplemental Material https://doi.org/10.23641/asha.8061740.

Use of formant centralization ratio for vowel impairment detection in normal hearing and different degrees of hearing impairment

Purpose: Hearing-impaired (HI) speakers show changes in vowel production and formant frequencies, as well as more cases of overlapping between vowels and more restricted formant space, than hearing speakers. This study was intended to explore whether the use of different acoustic parameters (Formant Centralization Ratio (FCR), Vowel Space Area (VSA), F2i/F2u ratio (second formant of/i,u/)) was suitable or not for characterizing impairments in the articulation of vowels in the speech of HL speakers. In fact, correlated acoustic parameters are used to determine the limits of tongue movements in vowel production in different severity degrees of hearing impairment.Methods: Speech recordings of 40 speakers with HL and 40 healthy controls were acoustically analyzed. The vowels (/a/,/i/,/u/) were extracted from the word context and, then, the first and second formants were calculated. The same vowel-formant elements were used to construct the FCR, expressed as (F2u + F2a + F1i + F1u)/(F2i + F1a), the F2i/F2u ratio, and the vowel space area (VSA), expressed as ABS((F1i*(F2a–F2u)+F1a*(F2u–F2i)+F1u*(F2i–F2a))/2).Results: The FCR differentiated HL groups from the control group and the discrimination was not gender-sensitive. All parameters were found to be strongly correlated with each other.Conclusions: The findings of this study showed that FCR was a more sensitive acoustic parameter than F2i/F2u ratio and VSA to distinguish speech of the HL groups from that of the normal group. Thus, FCR is considered to be applicable as an early objective measure of impaired vowel articulation in HL speakers.

Adjustment of Vocal Tract Shape via Biofeedback: Influence on Vowels

The study assessed 30 nonprofessional singers to evaluate the effects of vocal tract shape adjustment via increased resonance toward an externally applied sinusoidal frequency of 900 Hz without phonation. The amplification of the sound wave was used as biofeedback signal and the intensity and the formant position of the basic vowels /a/, /e/, /i/, /o/, and /u/ were compared before and after a vocal tract adjustment period. After the adjustment period, the intensities for all vowels increased and the measured changes correlated with the participants' self-perception.The diferences between the second formant position of the vowels and the applied frequency influences the changes in amplitude and in formant frequencies. The most significant changes in formant frequency occurred with vowels that did not include a formant frequency of 900 Hz, while the increase in amplitude was the strongest for vowels with a formant frequency of about 900 Hz.

Acoustical and perceptual study of voice disguise by age modification in speaker verification

The task of speaker recognition is feasible when the speakers are co-operative or wish to be recognized. While modern automatic speaker verification (ASV) systems and some listeners are good at recognizing speakers from modal, unmodified speech, the task becomes notoriously difficult in situations of deliberate voice disguise when the speaker aims at masking his or her identity. We approach voice disguise from the perspective of acoustical and perceptual analysis using a self-collected corpus of 60 native Finnish speakers (31 female, 29 male) producing utterances in normal, intended young and intended old voice modes. The normal voices form a starting point and we are interested in studying how the two disguise modes impact the acoustical parameters and perceptual speaker similarity judgments.First, we study the effect of disguise as a relative change in fundamental frequency (F0) and formant frequencies (F1 to F4) from modal to disguised utterances. Next, we investigate whether or not speaker comparisons that are deemed easy or difficult by a modern ASV system have a similar difficulty level for the human listeners. Further, we study affecting factors from listener-related self-reported information that may explain a particular listener’s success or failure in speaker similarity assessment.Our acoustic analysis reveals a systematic increase in relative change in mean F0 for the intended young voices while for the intended old voices, the relative change is less prominent in most cases. Concerning the formants F1 through F4, 29% (for male) and 30% (for female) of the utterances did not exhibit a significant change in any formant value, while the remaining ∼ 70% of utterances had significant changes in at least one formant.Our listening panel consists of 70 listeners, 32 native and 38 non-native, who listened to 24 utterance pairs selected using rankings produced by an ASV system. The results indicate that speaker pairs categorized as easy by our ASV system were also easy for the average listener. Similarly, the listeners made more errors in the difficult trials. The listening results indicate that target (same speaker) trials were more difficult for the non-native group, while the performance for the non-target pairs was similar for both native and non-native groups.

Six‐Month‐Old Infants Prefer Speech with Raised Formant Frequencies

Voices can be characterized by their fundamental frequency as well as by speech timbre characteristics, such as formant frequencies. While infants' responses to fundamental frequency characteristics are well established, very little is known about how infants respond to changes in formant frequencies. This study tested whether 6‐month‐old infants prefer listening to speech with raised formant frequencies over speech with lowered formant frequencies. Naturally spoken utterances were acoustically manipulated to render raised‐formant and lowered‐formant stimuli that only differed in the formant frequencies, while keeping fundamental frequency and other acoustic characteristics constant. Infants in an infant‐controlled listening procedure listened longer to the raised‐formant than to the lowered‐formant stimuli. These results provide the first evidence that infants distinguish and show preferences for variations in formant frequencies in adult speech.

The role of vowel formant frequencies and duration in the perception of foreign accent

ABSTRACTThis study used synthesis to manipulate vowel formant frequencies and durations to evaluate their role on foreign accent perception. Formant frequencies and durations for the vowels /æ/, /ʌ/, and /a/ were manipulated with changes toward and away from the mean native English and Spanish-accented values from Sidaras, S. K., Alexander, J. E. D., & Nygaard, L. C. (2009. Perceptual learning of systematic variation in Spanish-accented speech. The Journal of the Acoustical Society of America, 125, 3306–3316). Native listeners rated these stimuli on degree of accentedness and comprehensibility. Gradual changes in formant frequencies from native to non-native values impacted /a/ negatively, /ʌ/ positively, and /æ/ minimally. Effects of vowel duration on either type of ratings were small and restricted to vowel-specific interactions. The current findings suggest that vowel formant frequencies are primary cues to foreign accent. Their influence depends upon whether or not frequencies could reflect alternative vowel categories.

Neural Resolution of Formant Frequencies in the Primary Auditory Cortex of Rats.

Pulse-resonance sounds play an important role in animal communication and auditory object recognition, yet very little is known about the cortical representation of this class of sounds. In this study we shine light on one simple aspect: how well does the firing rate of cortical neurons resolve resonant (“formant”) frequencies of vowel-like pulse-resonance sounds. We recorded neural responses in the primary auditory cortex (A1) of anesthetized rats to two-formant pulse-resonance sounds, and estimated their formant resolving power using a statistical kernel smoothing method which takes into account the natural variability of cortical responses. While formant-tuning functions were diverse in structure across different penetrations, most were sensitive to changes in formant frequency, with a frequency resolution comparable to that reported for rat cochlear filters.

Articulatory and acoustic correlates of English front vowel productions by native Japanese speakers

Speakers of languages having a relatively small vowel inventory (e.g., Japanese) may experience interference effects when attempting to produce novel vowels from a language with a larger vowel inventory, such as English. Studies of English vowels produced by native Japanese speakers have shown formant frequency and duration differences between native and accented productions. Less is known about the exact articulatory processes underlying second language vowel learning. The current study examines tongue position during native Japanese speakers’ productions of American English vowels. Ten speakers produced the front vowels /i/, /ɪ/, /e/, /ε/, and /æ/ while seated in a 3D electromagnetic articulograph (EMA) system that tracked the position of the tongue (tongue tip, TT, tongue dorsum, TD, and tongue back, TB) and lower lip during vowel productions in the consonant environment /hVd/. Kinematic and acoustic measures were taken at the midpoint of each vowel steady-state portion. Differences in tongue position for tense-lax vowel pairs were determined by calculating the Euclidean distance between vowel centroids. Preliminary results suggest more spatial overlap in vowels produced by Japanese talkers than those produced by monolingual English speakers. The relationship between these articulatory distances and acoustic measures (formant frequency and durational changes) will be explored.

Vowel formant discrimination: Effects of listeners’ hearing status and language background

The goal of this study was to examine effects of listeners’ hearing status (e.g., normal and impaired hearing) and language background (e.g., native and non-native) on vowel formant discrimination. Thresholds of formant discrimination were measured for F1 and F2 of English vowels at 70 dB SPL for normal- (NH) and impaired-hearing (HI) listeners using a three-interval, two-alternative forced-choice procedure with a two-down, one-up tracking algorithm. Formant thresholds of HI listeners were comparable to those of NH listeners for F1, but significantly higher than NH listeners for F2. Results of a further experiment indicated that an amplification of the F2 peak could markedly improve formant discrimination for HI listeners, but a simple amplification of the sound level did not provide any benefit to them. On the other hand, another experiment showed that vowel density of listeners’ native language appeared to affect vowel formant discrimination, i.e., more crowded vowel space of listeners’ native language, better their vowel formant discrimination. For example, English-native listeners showed significantly lower thresholds of formant discrimination for both English and Chinese vowels than Chinese-native listeners. However, the two groups of listeners had similar psychophysical capacity to discriminate formant frequency changes in non-speech sounds.

EEG classification in a single-trial basis for vowel speech perception using multivariate empirical mode decomposition

Objective. The objective of this study is to find components that might be related to phoneme representation in the brain and to discriminate EEG responses for each speech sound on a trial basis. Approach. We used multivariate empirical mode decomposition (MEMD) and common spatial pattern for feature extraction. We chose three vowel stimuli, /a/, /i/ and /u/, based on previous findings, such that the brain can detect change in formant frequency (F2) of vowels. EEG activity was recorded from seven native Korean speakers at Gwangju Institute of Science and Technology. We applied MEMD over EEG channels to extract speech-related brain signal sources, and looked for the intrinsic mode functions which were dominant in the alpha bands. After the MEMD procedure, we applied the common spatial pattern algorithm for enhancing the classification performance, and used linear discriminant analysis (LDA) as a classifier. Main results. The brain responses to the three vowels could be classified as one of the learned phonemes on a single-trial basis with our approach. Significance. The results of our study show that brain responses to vowels can be classified for single trials using MEMD and LDA. This approach may not only become a useful tool for the brain–computer interface but it could also be used for discriminating the neural correlates of categorical speech perception.

Principal component analysis of formant transitions

Formant transitions represent dynamic articulation processes and are useful to describe differences of articulation (e.g., normal and disordered speakers). Our method, considers the formant frequency change over time as a vector. We first reduce this vector to a single value which is plotted multi-dimensional space. Each dimension represents a data point in time and the position of the point of each dimension is the value of the formant frequency. Once the vector has been converted to a single point in the multi-dimensional space, principal component analysis is applied to reduce the dimensionality. These dimensions can then be applied to multi-variant statistical analysis. Additionally, the eigenvectors in each principal component can be used to plot the differences between sets of formant transitions through time. The eigenvectors are first weighted by calculating the Pythagorean distance in the multidimensional space of the first set of principle components and then each eigenvector is multiplied by the percentage of the variance explained by that principle component. Then, the eigenvectors are plotted and superimposed over the formant transitions. The change of height along the length of the curve represents areas along the transition of maximum and minimum differences between sets of transitions.

Speaker sex effects on temporal and spectro-temporal measures of speech

This study investigated speaker sex differences in the temporal and spectro-temporal parameters of English monosyllabic words spoken by thirteen women and eleven men. Vowel and utterance duration were investigated. A number of formant frequency parameters were also analysed to assess the spectro-temporal dynamic structures of the monosyllabic words as a function of speaker sex. Absolute frequency changes were measured for the first (F1), second (F2), and third (F3) formant frequencies (ΔF1, ΔF2, and ΔF3, respectively). Rates of these absolute formant frequency changes were also measured and calculated to yield measurements for rF1, rF2, and rF3. Normalised frequency changes (normΔF1, normΔF2, and normΔF3), and normalised rates of change (normrF1, normrF2, and normrF3) were also calculated. F2 locus equations were then derived from the F2 measurements taken at the onset and temporal mid points of the vowels. Results indicated that there were significant sex differences in the spectro-temporal parameters associated with F2: ΔF2, normΔF2, rF2, and F2 locus equation slopes; women displayed significantly higher values for ΔF2, normΔF2 and rF2, and significantly shallower F2 locus equation slopes. Collectively, these results suggested lower levels of coarticulation in the speech samples of the women speakers, and corroborate evidence reported in earlier studies.

The Effect of Change in Spectral Slope and Formant Frequencies on the Perception of Loudness

This study attempts to understand how changes in spectral slope and formant frequency influence changes in perceived loudness. It was hypothesized that voices synthesized with steeper spectral slopes will be perceived as less loud than voices synthesized with less steep spectral slopes, in spite of the fact that they are of equal root mean square (RMS) amplitude. It was also hypothesized that stimuli with higher formant patterns will be perceived as louder than those with lower formant patterns, in spite of the fact that they are of equal RMS amplitude. Repeated measures factorial design. For the pitches A3, C4, B4, and F5, three different source signals were synthesized with varying slopes of -9, -12, and -15 dB/octave using a frequency vibrato rate of 5.6 Hz and a frequency vibrato extent of 50 cents. Each of the three source signals were filtered using two formant patterns, a lower formant pattern typical of a mezzo-soprano (pattern A) and a higher formant pattern typical of a soprano (pattern B) for the vowel /a/. For each pitch, the six stimuli were combined into all possible pairs and normalized to equal RMS amplitude. Listeners were presented with 120 paired stimuli (60 pairs repeated twice). The listener's task was to indicate whether the first or second stimulus in the pair was louder. Generally, as the spectral slope decreased, perceived loudness increased, with the magnitude of the perceived difference in loudness being related to the degree of difference in spectral slope. Likewise, at all pitches except A3, perceived loudness increased as formant frequency increased. RMS amplitude is an important predictor of loudness perception, but many other factors also affect the perception of this important vocal parameter. Spectral composition is one such factor and must be considered when using loudness perception in the process of clinical diagnostics.

On instantaneous vocal tract length estimation from formant frequencies

The length of the vocal tract and its relationship with formant frequencies is examined at fine temporal scales with the goal of providing accurate estimates of vocal tract length from acoustics on a spectrum-by-spectrum basis despite unknown articulatory information. Accurate vocal tract length estimation is motivated by applications to speaker normalization and biometrics. Analyses presented are both theoretical and empirical. Various theoretical models are used to predict the behavior of vocal tract resonances in the presence of different vocal tract lengths and constrictions. Real-time MRI with synchronized audio is also utilized for detailed measurements of vocal tract length and formant frequencies during running speech, facilitating the examination of short-time changes in vocal tract length and corresponding changes in formant frequencies, both within and across speakers. Previously proposed methods for estimating vocal tract length are placed within a coherent framework, and their effectiveness is evaluated and compared. A data-driven method for VTL estimation emerges as a natural extension of this framework, which is then developed and shown to empirically outperform previous methods on both synthetic and real speech data. A theoretical justification for the effectiveness of this new method is also explained. [Work supported by NIH.]

The role of vowel inherent spectral change in the intelligibility of English vowels spoken by English-, Chinese-, and Korean-native speakers

This study is designed to investigate the relationship between Vowel Inherent Spectral Change (VISC) and vowel intelligibility of native and non-native speakers of English. VISC refers to the relatively slow varying changes in formant frequencies associated with each vowel category (Neary and Assman, 1986). Such spectral change has been known to be a major factor in the perception of both the phonetic and phonemic English vowels. In the previous research projects conducted in our lab, we recorded 12 English vowels in /hVd/ format spoken by English native (EN), Chinese native (CN) and Korean native (KN) speakers and examined vowel intelligibility. Overall, vowel intelligibility was significantly higher for native talkers than for non-native talkers. Within non-native speaker groups, CN had slightly higher intelligibility than KN speakers. In this study, we are going to analyze an acoustic feature of the vowels, VISC, spoken by these native and non-native speakers. It is hypothesized that the acoustic differences of vowels spoken by different groups of speakers can account for the variances in vowel intelligibility.