Formant Frequency Patterns Research Articles

Vowel recognition at fundamental frequencies up to 1 kHz reveals point vowels as acoustic landmarks.

The phonological function of vowels can be maintained at fundamental frequencies (fo) up to 880 Hz [Friedrichs, Maurer, and Dellwo (2015). J. Acoust. Soc. Am. 138, EL36-EL42]. Here, the influence of talker variability and multiple response options on vowel recognition at high fos is assessed. The stimuli (n = 264) consisted of eight isolated vowels (/i y e ø ε a o u/) produced by three female native German talkers at 11 fos within a range of 220-1046 Hz. In a closed-set identification task, 21 listeners were presented excised 700-ms vowel nuclei with quasi-flat fo contours and resonance trajectories. The results show that listeners can identify the point vowels /i a u/ at fos up to almost 1 kHz, with a significant decrease for the vowels /y ε/ and a drop to chance level for the vowels /e ø o/ toward the upper fos. Auditory excitation patterns reveal highly differentiable representations for /i a u/ that can be used as landmarks for vowel category perception at high fos. These results suggest that theories of vowel perception based on overall spectral shape will provide a fuller account of vowel perception than those based solely on formant frequency patterns.

Two dimensional featured one dimensional digital waveguide model for the vocal tract

A vocal tract model based on a digital waveguide is presented in which the vocal tract has been decomposed into a number of convergent and divergent ducts. The divergent duct is modeled by a 2D-featured 1D digital waveguide and the convergent duct by a one dimensional waveguide. The modeling of the divergent duct is based on splitting the volume velocity into axial and radial components. The combination of separate modeling of the divergent and convergent ducts forms the foundation of the current approach. The advantage of this approach is the ability to get a transfer function in zero-pole form that eliminates the need to perform numerical calculations on a discrete 2D mesh. In this way the present model named as a 2D-featured 1D digital waveguide model has been found to be more efficient than the standard 2D waveguide model and in very good comparison with it in the formant frequency patterns of the vowels /a/, /e/, /i/, /o/ and /u/. The model has two control parameters, the wall and glottal reflection coefficients that can be effectively employed for bandwidth tuning. The model also shows its ability to generate smooth dynamic changes in the vocal tract during the transition of vowels.

Falling diphthongs have a dynamic target while rising diphthongs have two targets: Acoustics and articulation of the diphthong production in Ningbo Chinese

It is controversial whether diphthongs are phonologically vowel sequences and thus phonetically have two targets or diphthongs are phonologically vowel phonemes that contrast with monophthongs and thus phonetically have one dynamic target. Chinese dialects are generally known as having a rich inventory of diphthongs, and typically there are both falling and rising diphthongs. This paper is an acoustic and articulatory study on the diphthongs in Ningbo Chinese. The acoustic data are from 20 speakers and the lingual kinematic data are collected from 6 speakers by using EMA. The acoustic results show that both the onset and offset elements have comparable formant frequency patterns to their corresponding target citation vowels in a rising diphthong, but in a falling diphthong, only the onset element has a comparable formant frequency pattern to its corresponding target citation vowel whereas the offset element is highly variable. The articulatory results further reveal that diphthong onset is better controlled than diphthong offset, and more importantly, diphthong production is constrained by the general articulatory-to-acoustic relations. It is generally concluded that in Ningbo Chinese, rising diphthongs have two targets and can thus be understood as vowel sequences while falling diphthongs have only one dynamic target and should be treated as a single vowel phoneme.

PRODUCTION OF ENGLISH LEXICAL STRESS BY PERSIAN EFL LEARNERS

This study examines the phonetic properties of lexical stress in English produced by Persian speakers learning English as a foreign language. The four most reliable phonetic correlates of English lexical stress, namely fundamental frequency, duration, intensity, and vowel quality were measured across Persian speakers’ production of the stressed and unstressed syllables of five English disyllabic stress pairs which differed only in the location of stress, such as contract (noun)/ contract (verb). Results showed that Persian speakers’ use of the prosodic cues to lexical stress, that is fundamental frequency, duration, and intensity was comparable to the use of the same cues by American English speakers for both the stressed and unstressed syllables. There were, however, significant differences in formant frequency patterns (as the phonetic correlates of vowel quality) across the two language groups, such that Persian speakers did not manage to approximate the target native-like productions of the majority of the vowels in the experimental data both in the stressed or unstressed conditions. This observation supports the proposal made by Flege and Bohn (1989), namely that L2 learners acquire L1 patterns of vowel reduction only after they have acquired English-like patterns of prosodic cues to stress (F0, duration and intensity), and that their inability to reduce vowels in unstressed syllables does not influence their ability to employ prosodic cues to lexical stress contrast. As will be discussed at the end, the results shall have implications for material developers and EFL teachers.

Text Intelligibility and the Singer's Formant—A Relationship?

A clear enunciation of consonants is crucial to text intelligibility, and consonants are identified by specific formant frequency patterns. The singer's formant, a spectral peak near 3,000 Hz, enhances the higher formants in male opera singers' voices. It is well known that the second and higher formants are crucial to text intelligibility. Therefore, it seams reasonable to hypothesize that the singer's formant increases intelligibility of consonants and hence also the intelligibility of the text. For the same reason, text intelligibility of musical theatre singers, who lack a singer's formant, could be assumed to be lower than that of opera singers. Two professional opera singers and two professional musical theatre singers sang a carrier phrase that contained one nonsense syllable. The phrases were masked with noise of different levels. The degree of intelligibility was measured by a listening test. The results showed that the intelligibility was slightly higher for the musical theatre singers than for the opera singers. One possible reason for this would be that the musical theatre singers use formant frequencies more similar to those occurring in normal speech. Another reason could be that the formant transitions characterizing the consonants were considerably slower in the case of the musical theatre singers than in the case of the operatic singers.

Acoustically-guided vocal tract modifications for singing

The sound quality of a specific vowel can be dramatically altered by subtle modifications of the vocal tract shape. These modifications create changes in the pattern of formant frequencies. For example, the well-known singing formant, which is a clustering of resonance frequencies, is typically the result of constricting the epilaryngeal space or expanding the lower pharyngeal space to create a large cross-sectional area discontinuity between them. Other modifications such as lip protrusion/spreading or larynx lowering/raising will also impose changes on the formant frequency pattern that may be desirable for singing or speech production. This presentation will focus on the acoustic sensitivity of the resonance frequencies to subtle perturbations of specific vowel configurations. Using calculated sensitivity functions, it will be shown how specific regions along the vocal tract can be constricted or expanded to perturb one or more of the formant frequencies. In effect, this technique provides a means of ‘‘tuning’’ the vocal tract shape to produce a desired frequency response. [Work supported by NIH R01-DC04789.]

Listener sensitivity to intensity variation in vowels in coarticulatory context

There are two sources of acoustic variation in the relative formant amplitude of a vowel which are unrelated to emphasis-oriented changes in articulatory effort: (1) vowel-specific intrinsic formant frequency pattern, which differs across vowel categories, and (2) the vowel’s immediate phonetic (coarticulatory) context. Listener sensitivity to the latter source of variation is examined here to establish the extent to which distortion in the distribution of acoustic energy across the vowel spectrum affects the perception of vowel quality. Synthetic versions of CVC syllables in symmetrical consonant environments were created (using stops or fricatives differing in place-of-articulation and voicing). Formant frequencies and amplitude values were modeled after those found in a large corpus of CVC acoustic data [E. Jacewicz and R. A. Fox, J. Acoust. Soc. Am. 117, 2623 (2005)]. Initial amplitude values for the first four formants in synthetic CVCs reflected the intensity distribution (and spectral tilt) typical of a vowel coarticulated with these consonants. Subsequent stepwise adjustments to selected formant levels were made to create a range of permissible and not permissible formant amplitude variations in coarticulatory context. Preliminary results indicate that listeners are sensitive to amplitude variation across vowel spectrum as a function of a specific consonant environment. [Work supported by NIDCD R03 DC005560-01.]

The relative amplitude of vowel formants: Acoustic analysis

Acoustically, vowels as speech signals transport power, which is expressed in their overall intensity level. This power varies intrinsically with vowel category and with its immediate consonantal context. These seminal findings [House and Fairbanks, J. Acoust. Soc. Am. 22, 105–113 (1953)] have received little systematic consideration over the past 50 years. The present work is a partial replication and extension of the House and Fairbanks study to formant levels, providing a set of data on intensity variations for eight selected American English vowels. Parametric investigation includes overall rms amplitude, amplitude of F0, and relative amplitudes of formants F1–F4. The study examines how the relative levels of particular formants contribute to the overall vowel intensity and how this distribution changes in consonantal contexts. Changes in vowel intensity (overall intensity and that of the first four formants) as a function of the consonantal context (both symmetrical and asymmetrical) are examined for each consonantal feature: place and manner of articulation, and voicing. Additionally, the intrinsic intensity is measured from sustained vowels recorded in isolation. The results are discussed in relation to the contributions of primary (i.e., formant frequency pattern) and secondary cues to vowel identity. [Work supported by NIDCD R03 DC005560−01.]

Sex-specific fundamental and formant frequency patterns in a cross-sectional study.

An extensive developmental acoustic study of the speech patterns of children and adults was reported by Lee and colleagues [Lee et al., J. Acoust. Soc. Am. 105, 1455-1468 (1999)]. This paper presents a reexamination of selected fundamental frequency and formant frequency data presented in their report for ten monophthongs by investigating sex-specific and developmental patterns using two different approaches. The first of these includes the investigation of age- and sex-specific formant frequency patterns in the monophthongs. The second, the investigation of fundamental frequency and formant frequency data using the critical band rate (bark) scale and a number of acoustic-phonetic dimensions of the monophthongs from an age- and sex-specific perspective. These acoustic-phonetic dimensions include: vowel spaces and distances from speaker centroids; frequency differences between the formant frequencies of males and females; vowel openness/closeness and frontness/backness; the degree of vocal effort; and formant frequency ranges. Both approaches reveal both age- and sex-specific development patterns which also appear to be dependent on whether vowels are peripheral or nonperipheral. The developmental emergence of these sex-specific differences are discussed with reference to anatomical, physiological, sociophonetic, and culturally determined factors. Some directions for further investigation into the age-linked sex differences in speech across the lifespan are also proposed.

Acoustic and articulatory differences across word position for American English /r/

Typically, American English /r/ is manifested with different formant frequency patterns in word-initial, word-final and word-medial position. In particular, it has been reported (Lehiste, 1962; Espy-Wilson, 1992) that initial /r/’s show lower third formants. This may be due to differences in position or configuration of the primary articulator for /r/—the tongue—or it may reflect an increased degree of lip-rounding. Differences in tongue shape across word position have been reported (Zawadaski and Kuehn, 1982), but it is not clear how reported differences may relate to formant lowering. In this study, eight native speakers of American English recorded words with initial, final and medial /r/’s. Articulatory movements of the tongue tip, tongue blade and tongue dorsum, lips and jaw were tracked simultaneously by Electro-Magnetic Midsagittal Articulometer (EMMA). For each speaker, we compare spatial position, time course, and trajectory for articulators across positional variants. Preliminary data suggest that tongue articulators show similar time course in all three positions. Results are discussed with regard to acoustical modeling studies, theories of articulatory strengthening and interarticulator timing. [Research supported by NIH and NSF.]

Vocal-tract parameter estimation from formant patterns

A method is proposed to estimate individual vocal-tract parameters from formant frequency patterns. Vocal-tract parameters, such as scaling factors of vocal-tract length and area were determined in reference to the given area function of an articulatory synthesizer. These individual parameters of various speakers are expected to be suitable to normalize speakers in automatic speech recognition. It has been shown by Schroeder (1967) that there is a linear relationship between relative formant frequency deviation δω/ω and a relative area perturbation δA/A of the area function. Further, a similar relationship can be derived for length perturbation of an area function. The presented method is based on the inverse relationship: deviations of the formant frequency pattern of a speaker relative to the reference speaker are attributed to length and area perturbations of the area function of the reference speaker. A pilot study using synthetic speech showed promising results: estimated vocal-tract area functions were in good agreement with natural area functions. In the present study the method is extended to real speech. Investigations are carried out on speech samples of a German data base (PhonDat) to find speaker-specific parameters. [Work supported by Deutsche Forschungsgemeinschaft (Str 255/7-2).]

The role of transition duration in the perception of speech and nonspeech analogs of place of articulation

A previous report [R. Zakia, J. Acoust. Soc. Am. Suppl. 1 87, S117 (1990)] demonstrated that at a formant frequency pattern ambiguous between alveolar and velar, subjects identify synthetic speech stimuli with longer transition durations as velars and stimuli with shorter transition durations as alveolars. These results suggest the operation of either (1) an articulatory property characteristic of velars (longer transition durations distinguish velars because of their slower articulatory release), or (2) differential perceptual sensitivity to transition duration for spectral patterns characteristic of velars. To evaluate these possibilities, nonspeech analogs of a formant pattern ambiguous between an alveolar and a velar were generated with transition durations ranging between 20 and 50 ms in 5‐ms steps. In contrast to the previous identification task, the nonspeech stimuli were presented in a same‐different discrimination task. For comparison purposes, this same‐different task was also conducted with the...

Effects of fundamental frequency contour on the identification of resynthesized vowels with static formant frequency patterns

At a previous meeting, a study that was aimed at determining the identifiability of vowels based exclusively on static spectral information was discussed [Hillenbrand and McMahon, J. Acoust. Soc. Am. Suppl. 1 82, S81 (1987)]. In that study, a formant synthesizer was used to generate steady‐state versions of 1520 vowels (76 speakers × 10 vowels × 2 repetitions) using Peterson and Barney's measured values of F0 and F1−F3 [J. Acoust. Soc. Am. 24, 175–184 (1952)]. The values of all control parameters remained constant throughout the 300‐ms duration of each stimulus. Listeners in that study showed an error rate of approximately 25%, several times greater than the 5.6% error rate reported in the original Peterson and Barney study. The present study represents a follow‐up designed to determine what role fundamental frequency movement might play in vowel identification. The new stimuli were identical to those of the previous resynthesis study except that all stimuli were generated with a falling pitch contour. Pr...

Acoustic analysis of correct and misarticulated semivowels.

Four children who produced correct (w,r,l,j), four children with developmental w/r and w/l substitutions, and four articulation impaired children with w/r and w/l substitutions were subjects. They produced sets of minimally contrasted words with (w,r,l,j) in word-initial position with four vowels and with (w,r,l) in two types of consonant clusters. Children's utterances were spectrographically analyzed for three formant frequencies and transition rate of the second formant. Children with correct semivowels produced distinctive formant frequency patterns for semivowels that were similar to those previously reported in the literature for adults and children. Developmental and articulation impaired children produced acoustic features for (j) that were similar to the (j) produced by the control group; but neither group differentiated among (w,r,l) by either formant frequencies or transition rate. Some individuals in both groups produced formant frequency and/or transition rate differences among semivowels in some phonetic contexts. The (w) produced for target (w) and in substitution for (r) and (l) by three developmental children and two articulation-impaired children did not match the acoustic pattern of control (w). These productions had higher second formants, occurring between control (w) and (r,l) or in the range of correct (r,l).

Phonetic speech perception deficits in dyslexia.

Adult developmental dyslexics showed deficits in the identification of the vowels of English when the sole acoustic cues were steady-state formant frequency patterns. Deficits in the identification of place of articulation of the English stop-consonants [b], [d] and [g] in syllable-initial position were also observed. The average consonantal error rate was 22%. These error rates are significantly different from those of nondyslexic control groups (p less than .01). No single deficit characterized the entire group of dyslexic subjects. The pattern of errors with respect to place of articulation also varied for different groups of subjects. Three dyslexics have high vowel error rates and low consonantal error rates. The data are consistent with the premise that dyslexic subjects may have different perceptual deficits rather than a general auditory deficit involving the rate at which they can process perceptual information. The clinical histories of the present subjects suggest genetic transmission of these speech perception deficits. The presence of genetic variation in the biological substrate relevant to the perception of human speech should be further explored.

A model of formant frequency patterns in Russian VCV utterances

A recent article in JASA described Russian formant frequency patterns in VCV utterances. Based on a stepwise multiple regression analysis of the data, that article showed the influence of articulatory predictors on each of four measurement points within the VCV test words. The analysis was limited to a separate consideration of each measurement point and the relative influence of the various articulatory parameters on the formant frequency at that point. The present paper presents a more comprehensive, more theoretically oriented analysis of the data. Using LISREL, a statistical package for the analysis of linear structural relations, it was possible to study the interactions of the measured variables and the effects of the articulatory parameters on these measured variables in one pass. The results of this analysis should be more interesting from a theoretical point of view.

Formant frequency patterns in Russian VCV utterances

Previous studies have sought to establish the significance of various parameters in the determination of the patternings of Russian formant frequency trajectories in vowel-consonant-vowell (VCV) syllables. In the present study, 9600 measurements of first and second formant frequency were made on Russian VCV nonsense words. Variance for speakers was controlled through regression. Stepwise multiple regression was employed to determine the relative contributions of six predictors to the explanation of the patternings of first and second formant frequency in the test words. The six predictors included the height of the first and second vowels, the fron/back dimensionality of the first and second vowels, the place of articulation of the consonant, and the palatalization or nonpalatalization of the consonant. Results of the regression analyses are summarized. The regression coefficients for the useful predictors of the various dependent variables comprise a model of formant frequency patterns in Russian VCV utterances.

Intrinsic vowel duration and formant frequencies: Data from speech acquisition

The vowels of American English are specified by particular formant frequency patterns and durations. Acoustic data derived from the speech of children between the ages of 69 and 145 weeks show that the vowels that they produce in meaningful words often have mismatched formant frequency patterns and durations. The formant frequency pattern of an intended [i], for example, may have the duration of an [1]. Psychoacoustic experiments using computer edited stimuli derived from the children's vowel productions demonstrate the interaction of formant frequency patterns and duration as acoustic cues to vowel perception. Intrinsic vowel duration and control of the supralaryngeal vocal tract appear to be independent parameters in the acquisition of speech. Hypothetical phonetic features like Tenseness thus can not be causally linked to relative vowel duration.

Speech development in infants—vowel production

Five infants ranging in age from 12 to 66 weeks at the start of the project have been recorded at biweekly intervals during half-hour play sessions with their mothers. Fundamental and formant frequencies were derived from sound spectrograms, Fourier and Linear Predictive analysis. The data reveal differences in the infants' phonetic development as they gradually acquire the vowels of English. Their formant frequency patterns are outside the range reported by Peterson and Barney (1952) for older children but are consistent with the shorter supralaryngeal vocal tracts of these infants. The formant patterns that specify particular vowels change as the infants mature. F1 of/ae/, for example, changes from 1.0 to 0.7 kHz for one infant between 37 and 56 weeks. The infants' vowel production and their mothers' responses are consistent with the presence of an innate vocal tract length normalization perceptual mechanism that enters into the development of speech production.

Multidimensional representation of personal quality of vowels and its acoustical correlates

The personal quality of sustained vowels uttered by eight male talkers was represented multidimensionally in a psychological auditory space (PAS) by means of Kruskal's multidimensional scaling procedure based on the perceptual confusion in talker discrimination tests. Physical properties of the vowels were analyzed in terms of elementary acoustical parameters, such as formant frequencies, slope of glottal source spectrum, mean fundamental pitch frequency, and rapid fluctuation of fundamental pitch period. Then the relationship between the configuration on the PAS and the acoustical parameters was examined through multiple correlation and regression analysis. The contribution of those acoustical parameters to the personal quality of the five Japanese vowels and the relative contributions of the vocal tract and the glottal source characteristics are demonstrated quantitatively. These results were obtained partially, by utilizing hybrid voices in which the source wave or the formant frequency pattern was interchanged among different talkers.