Fundamental Frequency Patterns Research Articles

Do preceding prosodic patterns influence word recognition in Spanish?

Recent work shows lexical recognition in English is influenced by prosodic characteristics of preceding words (e.g., Dilley et al, 2010, Journal of Memory and Language). This study investigates whether preceding prosody affects lexical recognition in Spanish, a language with productive lexical stress contrasts (e.g., saco ‘bag’ vs. saco ‘he carried out’). In an experiment, each of 30 disyllabic test items was embedded in a carrier phrase; each item was consistent with one of two lexical interpretations that were distinguished by having primary stress on the initial vs. final syllable (e.g., saco vs. saco). The prosody of words preceding the test item was resynthesized with one of two fundamental frequency and timing patterns predicted to generate distinct patterns of syllable prominence favoring perception of primary stress on either the initial or final syllable of the test item. The acoustic characteristics of the test item were identical across prosody conditions. If perception of lexical stress depend...

The perception of intonation questions and statements in Cantonese

In tone languages there are potential conflicts in the perception of lexical tone and intonation, as both depend mainly on the differences in fundamental frequency (F0) patterns. The present study investigated the acoustic cues associated with the perception of sentences as questions or statements in Cantonese, as a function of the lexical tone in sentence final position. Cantonese listeners performed intonation identification tasks involving complete sentences, isolated final syllables, and sentences without the final syllable (carriers). Sensitivity (d' scores) were similar for complete sentences and final syllables but were significantly lower for carriers. Sensitivity was also affected by tone identity. These findings show that the perception of questions and statements relies primarily on the F0 characteristics of the final syllables (local F0 cues). A measure of response bias (c) provided evidence for a general bias toward the perception of statements. Logistic regression analyses showed that utterances were accurately classified as questions or statements by using average F0 and F0 interval. Average F0 of carriers (global F0 cue) was also found to be a reliable secondary cue. These findings suggest that the use of F0 cues for the perception of intonation question in tonal languages is likely to be language-specific.

Perceptual characteristics of broken and drop tones in Vietnamese

Perceptual characteristics of broken and drop tones in the Vietnamese language are studied. A broken tone syllable has a glottal constriction at the middle part, while a drop tone syllable has a glottal stop at the end and has a short length. In this study, four experiments were conducted. First, expanding from our previous research, the power control for the broken tone was evaluated using several power patterns and its effectiveness was reconfirmed. The power dip pattern makes the synthetic sound of the broken tone highly intelligible. Second, the similarity between the drop tone and the first half of the broken tone were studied. Experimental results showed that they are comparable, i.e., the first half of the broken tone can be replaced by the drop tone, and the first half of the broken tone can be perceived as a drop tone. In the third and fourth experiments, the final parts of drop and broken tones were examined using different fundamental frequency (F0) patterns. As a result, we found that the drop tone has a perceptual threshold with an F0 slope, while F0 of the final part of the broken tone has a minor effect on its perception.

The Perception of Lexical Tone Contrasts in Cantonese Children With and Without Specific Language Impairment (SLI)

This study examined the perception of fundamental frequency (f0) patterns by Cantonese children with and without specific language impairment (SLI). Participants were 14 five-year-old children with SLI, and 14 age-matched (AM) and 13 four-year-old vocabulary-matched (VM) controls. The children identified a word from familiar word pairs that illustrated the 8 minimally contrastive pairs of the 6 lexical tones. They discriminated the f0 patterns within contrastive tonal pairs in speech and nonspeech stimuli. In tone identification, the SLI group performed worse than the AM group but not the VM group. In tone discrimination, the SLI group did worse than the AM group on 2 contrasts and showed a nonsignificant trend of poorer performance on all contrasts combined. The VM group generally did worse than the AM group. There were no group differences in discrimination performance between speech and nonspeech stimuli. No correlation was found between identification and discrimination performance. Only the normal controls showed a moderate correlation between vocabulary scores and performance in the 2 perception tasks. The SLI group's poor tone identification cannot be accounted for by vocabulary knowledge alone. The group's tone discrimination performance suggests that some children with SLI have a deficit in f0 processing.

Duration and fundamental frequency patterns in Brazilian Portuguese polysyllabic words.

Brazilian Portuguese polysyllabic words are said to have secondary stress (SS). There is no clear consensus, though, on how SSs are assigned and what the acoustic correlates for SSs are. We analyzed a corpus of read speech in order to gain some insight on those two issues. The corpus’ target words have 2 to 4 pre‐stressed syllables with primary lexical stress falling either on the last or penultimate syllable. Target words appeared in initial, medial, or final position on carrier sentences. Previous works showed that duration and F0 are the main correlates of stress marking in BP with a marginal contribution of vowel quality and spectral emphasis. Statistical analysis of normalized (z‐scored) syllable and vowel‐to‐vowel duration show no evidence of binary alternation. Duration increases exponentially up to the stressed syllable. A Praat script was used to determine global and local F0 maxima and minima. Lexically stressed syllables have F0 peaks associated with them. There is at most one F0 peak on the pre‐stressed syllable chain. Peaks tend to be higher in 4‐pre‐stressed‐syllable words and be absent in 2‐pre‐stressed‐syllable words. Target word position do not have a significant effect on duration contour shape and on F0 peak placement.

The Interaction of Syntactic Structure and Postlexical Prosody in Saisiyat of Taiwan

Postlexical prosodic phenomena in Austronesian languages have received relatively little attention, and consequently their patterns remain unknown. This paper aims to bridge this gap by investigating how syntactic structure interacts with postlexical prosodic phenomena in Saisiyat, an endangered language spoken in Taiwan. Several significant findings are made. First, Saisiyat sentential fundamental frequency (F 0 ) patterns are based largely on its original lexical-level word F 0 contour. The accent of a content word in sentences usually falls on its ultimate syllable, while trisyllabic or quadrisyllabic content words may sometimes undergo postlexical accentual modifications such as accent spreading, accent fronting, and accent adding. Function words, in contrast, play a role of interpolation as an intermediate site in bridging the F 0 of their preceding and following syllables. Second, a yes-no question exhibits substantial prosodic modification by influencing its word preceding sentence-final interrogative particle aj , as compared with its counterpart in a declarative sentence. Third, agents in agent-focus sentences and patients in patient-focus sentences demonstrate higher values with respect to F 0 peak, mean F 0 , and mean intensity. We provide typological explanations for these findings and explore the theoretical implications of postlexical prosodic patterns of Formosan languages in Taiwan.

Frequency Modulation During Song in a Suboscine Does Not Require Vocal Muscles

The physiology of sound production in suboscines is poorly investigated. Suboscines are thought to develop song innately unlike the closely related oscines. Comparing phonatory mechanisms might therefore provide interesting insight into the evolution of vocal learning. Here we investigate sound production and control of sound frequency in the Great Kiskadee (Pitangus sulfuratus) by recording air sac pressure and vocalizations during spontaneously generated song. In all the songs and calls recorded, the modulations of the fundamental frequency are highly correlated to air sac pressure. To test whether this relationship reflects frequency control by changing respiratory activity or indicates synchronized vocal control, we denervated the syringeal muscles by bilateral resection of the tracheosyringeal nerve. After denervation, the strong correlation between fundamental frequency and air sac pressure patterns remained unchanged. A single linear regression relates sound frequency to air sac pressure in the intact and denervated birds. This surprising lack of control by syringeal muscles of frequency in Kiskadees, in strong contrast to songbirds, poses the question of how air sac pressure regulates sound frequency. To explore this question theoretically, we assume a nonlinear restitution force for the oscillating membrane folds in a two mass model of sound production. This nonlinear restitution force is essential to reproduce the frequency modulations of the observed vocalizations.

New cochlear implant coding strategy for tonal language speakers

Accurate pitch perception on the basis of fundamental frequency patterns is essential for the processing of lexical tones in tonal languages such as Cantonese. Speech intelligibility in Cantonese-speaking CI recipients was compared using current signal processing strategies, which typically result in poor pitch perception, and a new strategy, known as the multi-channel envelope modulation (MEM) strategy, was designed to enhance temporal periodicity cues to the fundamental frequency. Performance of nine postlingually hearing-impaired adult cochlear implant users was measured twice using each strategy, initially after a four week trial, and again after two weeks of use with each strategy. Speech intelligibility in speech-spectrum shaped noise was measured using the Cantonese hearing in noise test. A fixed noise level of 65 dB A was used and the signal-to-noise ratios were fixed at either +10, +15, or +20 dB, depending on the baseline performance of individual subjects using the clinical processor. Self-reported benefit in 18 listening situations and overall preference for strategies were obtained at the end of these trial periods. Results showed poorer speech intelligibility with CIS while results obtained using ACE and MEM were comparable. Unfamiliar place coding might have contributed to poorer performance using CIS. Self-reported benefit across strategies did not differ in most listening situations. Participants preferred ACE for listening overall in daily situations, and a few preferred MEM in noise. Whilst the results did not demonstrate any advantages for speech recognition in noise when using MEM compared to ACE, no degradation in performance was observed. This implies that the form of processing employed by MEM retains similar segmental information to that provided by ACE and that potentially, future variations/optimizations of MEM may lead to some improvement in tone perception.

Classification and similarity analysis of fundamental frequency patterns in infant spoken language acquisition

Fundamental frequency (F0) patterns, which indicate the vibration frequency of vocal cords, reflect the developmental changes in infant spoken language. In previous studies of developmental psychology, however, F0 patterns were manually classified into subjectively specified categories. Furthermore, since F0 has sequential missing and indicates a mean nonstationarity, classification that employs subsequent partition and conventional discriminant analysis based on stationary and local stationary processes is considered inadequate. Consequently, we propose a classification method based on discriminant analysis of time series data with mean nonstationarity and sequential missing, and a measurement technique for investigating the configuration similarities for classification. Using our proposed procedures, we analyse a longitudinal database of recorded conversations between infants and parents over a five-year period. Various F0 patterns were automatically classified into appropriate pattern groups, and the classification similarities calculated. These similarities gradually decreased with infant’s monthly age until a large change occurred around 20 months. The results suggest that our proposed methods are useful for analysing large-scale data and can contribute to studies of infant spoken language acquisition.

Utilization of acoustic cues in identifying speech intonation contrasts in cochlear implant recipients and listeners with normal hearing

This study investigated the utilization of acoustic cues in identifying speech intonation contrasts in adult cochlear implant (CI) recipients and normal‐hearing (NH) individuals listening to spectrally and temporally degraded stimuli. Fundamental frequency, intensity, and duration patterns of a disyllabic word, popcorn, were manipulated orthogonally, resulting in 360 resynthesized stimuli. In a two‐alternative forced‐choice task, CI and NH participants identified whether each stimulus was question‐like or statement‐like. Each NH listener also identified stimuli that were noise vocoded. Preliminary results from seven CI and four NH listeners indicated that (a) weighting of fundamental frequency in the identification of CI listeners was less pronounced than that of NH listeners, (b) unlike NH listeners who showed little reliance upon intensity or duration patterns in identifying unprocessed stimuli, CI users demonstrated systematic dependence upon intensity and duration cues in identifying the same set of stimuli, and (c) with noise‐vocoded (spectrally and temporally degraded) stimuli, NH listeners’ identification exhibited weighting patterns for intensity and duration cues in a way similar to that of CI listeners. Implications for the processing of multidimensional acoustic cues for speech intonation in normal and electrical hearing will be discussed. [Work supported by NIDCD‐R01DC04786.]

The role of Swedish tonal word accents in children with language impairment

The present study was designed to examine the production of the Swedish tonal accents in children with language impairment and normal controls in order to verify previous findings. The productions of 25 children with linguistic impairment and their matched controls, aged 4;4–10;0 (mean age 5;11) were evaluated by ratings of fundamental frequency patterns, and by perceptual analysis by ten linguistically naive listeners to assess the distinctiveness of the accents. These methods give a more detailed description of the children's productions of tonal word accents. The results show that 60.8% of the children with language impairment have difficulties to produce the contrast of tonal word accents according to the F0 patterns. The difference in the rating of the F0 curves between the children with LI and their matched controls was significant. There is a significant positive correlation between the perceptual results and the F0‐rating, indicating that the better a child was perceived, the better his/her F0 production was rated. We conclude that the distinction between the tonal word accents is a difficult feature to acquire for Swedish children, which is shown both by acoustic and perceptual evaluations.

Contrasting the effects of duration and number of syllables on the perceptual normalization of lexical tones

In tonal languages, syllabic fundamental frequency (F0) patterns (‘‘lexical tones’’) convey lexical meaning. Listeners need to relate such pitch patterns to the pitch range of a speaker (‘‘tone normalization’’) to accurately identify lexical tones. This study investigated the amount of tonal information required to perform tone normalization. A target CV syllable, perceived as either a high level, a low level, or a mid level Cantonese tone, was preceded by a four-syllable carrier sentence whose F0 was shifted (1 semitone), or not shifted. Four conditions were obtained by gating one, two, three, or four syllables from the onset of the target. Presentation rate (normal versus fast) was set such that the duration of the one, two, and three syllable conditions (normal carrier) was equal to that of the two, three, and four syllable conditions (fast carrier). Results suggest that tone normalization is largely accomplished within 250 ms or so prior to target onset, independent of the number of syllables; additional tonal information produces a relatively small increase in tone normalization. Implications for models of lexical tone normalization will be discussed. [Work supported by the RGC of the Hong Kong SAR, Project No. HKU 7193/00H.]

Vietnamese Text-To-Speech system with precise tone generation

We propose a Vietnamese Text-To-Speech (VieTTS) system which is a parametric and rule-based speech synthesis system. Fundamental speech units of this system are demisyllables with Level tone. VieTTS uses a source-filter model for speech production and a Log Magnitude Approximation (LMA) filter as the vocal tract filter. We chose the Hanoi dialect for VieTTS. Tone synthesis of Vietnamese is implemented by using fundamental frequency (F0) patterns and power pattern control. F0 is the most important factor in Vietnamese tone synthesis and the power control strongly affects Broken and Drop tones. Applying power control for tone synthesis is unique for Vietnamese compared to other tonal languages such as Chinese and Thai. This new result is confirmed by listening tests with a reasonable listening correct rate.

The Impact of Cerebral Palsy on the Intelligibility of Pitch-based Linguistic Contrasts

This study investigated the perception of three Cantonese level tones produced by speakers with dysarthria associated with cerebral palsy. Four speakers with dysarthria were selected on the basis of their distinctive patterns of fundamental frequency (F0) values observed in the level tones they produced, which showed errors in either F0 level or, F0 contour, or both. Monosyllabic words which contrasted in tone level were used as stimuli in an identification task. Five expert listeners identified the tones among the six Cantonese contrastive tones. Results showed that the tones produced by the dysarthric speakers were predominantly perceived as level tones; although a majority was perceived as the wrong level tone. The most important finding is that only the level tones produced by dysarthric speakers could be identified as rising or falling contour tones. The frequent perceptual confusion among the level tones, and the perception of contour tones, shows that a disorder in the production of pitch-based linguistic contrasts can have a substantial impact on the communication abilities of individuals with cerebral palsy.

Fundamental frequency pattern generator, method and program

According to this fundamental frequency generating method, a fundamental frequency pattern is set from a data base of a fundamental frequency pattern of each accent phrase standardized by the phoneme time length or the time length of the vowel and the vowel corresponding portion, and when the corresponding fundamental frequency pattern is not stored in the data base, the fundamental frequency pattern is generated by interpolating the interval between points serving as the references of the fundamental frequency pattern. With this method, a fundamental frequency pattern having higher naturalness than with conventional methods can be generated.

Tone classification of syllable-segmented Thai speech based on multilayer perception

Thai is a monosyllabic tonal language that uses tone to convey lexical information about the meaning of a syllable. Thus to completely recognize a spoken Thai syllable, a speech recognition system not only has to recognize a base syllable but also must correctly identify a tone. Hence, tone classification of Thai speech is an essential part of a Thai speech recognition system. Thai has five distinctive tones (‘‘mid,’’ ‘‘low,’’ ‘‘falling,’’ ‘‘high,’’ and ‘‘rising’’) and each tone is represented by a single fundamental frequency (F0) pattern. However, several factors, including tonal coarticulation, stress, intonation, and speaker variability, affect the F0 pattern of a syllable in continuous Thai speech. In this study, an efficient method for tone classification of syllable-segmented Thai speech, which incorporates the effects of tonal coarticulation, stress, and intonation, as well as a method to perform automatic syllable segmentation, were developed. Acoustic parameters were used as the main discriminating parameters. The F0 contour of a segmented syllable was normalized by using a z-score transformation before being presented to a tone classifier. The proposed system was evaluated on 920 test utterances spoken by 8 speakers. A recognition rate of 91.36% was achieved by the proposed system.

Cantonese tone perception ability of cochlear implant children in comparison with normal-hearing children

Background and Objective: Cantonese is a tone language. A change in the fundamental frequency pattern within the same phonemic segment causes a change in the lexical meaning. The present study examined the Cantonese tone perception ability of cochlear implant children in comparison with normal-hearing children. It was hypothesized that cochlear implant children follow a similar pattern of tone perception development, as do normal children. Method: 225 normal-hearing and 15 hearing-impaired children with cochlear implants were recruited. The high level (tone 1), high rising (tone 2) and low falling (tone 4) were the target tones examined. The three tones were arranged into tone pairs for identification. Each pair shared exactly the same segmental information but differed only in tones (e.g. /sy/ in tones 1 and 2 meaning ‘book’ and ‘mouse’, respectively). Subjects were required to point to the corresponding pictures after the live voice presentations. Results: for each tone pair, each subject was awarded a score representing the proportion of stimuli pairs that were correctly discriminated by the subject. The average scores in the normal-hearing and hearing-impaired groups were 0.92 and 0.64, respectively. The normal group had the lowest average score in tone 2/tone 4 (0.87) while the hearing-impaired group performed the worst in tone 1/tone 2 (0.53) perception between the three tone contrasts. Conclusions: the normal-hearing group performed significantly better than the hearing-impaired group in basic Cantonese tone perception. The pattern of tone perception development of cochlear implant children did not seem to follow that of normal children. Contributing factors on the tone perception performance of the cochlear implant children were subject's age, duration of special training, and durations of wearing the hearing aid and the cochlear implant.

Effect of speaking rate on segment durations of voiceless stops in Korean

Studies of the effect of speaking rate on speech segment durations reveal that the VOTs of long lag and prevoiced consonants vary directly with speaking rate, whereas those of short lag consonants remain relatively constant across speaking rates (i.e., Kessinger and Blumstein, 1998; Miller and Volaitis, 1989). Koreans have three voiceless stop consonants at each place of articulation. These stops vary in VOT with values in the long lag and short lag ranges. These phonemes are also marked by differing fundamental frequency patterns. The purpose of this study was to investigate the effects of different speaking rates on the acoustic parameters of the Korean voiceless stop consonants. Sixteen Korean adults (8 male and 8 female) repeated productions of the lax, reinforced, and aspirated forms of /p/, /t/, and /k/ with the vowel /i/. These repetitions were made of the isolated CV syllables and of the syllables in a Korean sentence context. The syllables were analyzed for segment durations of hold, transition and vocalic portions and fundamental frequency contours. The speech was recorded using a DAT recorder and analyzed using a CSL 4300B system. The effect of the speaking rate on the different phonemes will be discussed [R. Kessinger and S. Blumstein, ‘‘Effects of speaking rate on voice-onset time and vowel production: Some implications for perception studies,’’ J. Phonetics 26, 117–128 (1998); J. Miller and L. Volaitis, ‘‘Effect of speaking rate on the perceptual structure of a phonetic category,’’ Percept. Psychophys. 46, 505–512 (1989)].

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.

Intonation in the monosyllabic utterances of 1-year-olds

Previous research has demonstrated that falling contours predominate in infant utterances as early as 3 months of age. The precocious appearance of falling intonation is usually attributed to “biological tendencies,” that is, the physiological naturalness of descending fundamental frequency patterns. In contrast, other investigations have shown that some children do not use adultlike falling or rising intonation contours until they produce their first words. To resolve these conflicting views of prosodic development, this study acoustically investigated intonation production in the monosyllabic utterances of 10 English-speaking children from 10 to 13 months of age and the utterance-final monosyllables of ten 4-year-olds. Children in both age groups produced a wider accent range in falling contours than in rising contours. Infants produced a narrower accent range than the preschoolers. The findings suggest that biological tendencies are not sufficient to account for children’s acquisition of intonation between the ages of 1 and 4 years.