Carrier Sentence Research Articles

Foreign Languages Sound Fast: Evidence from Implicit Rate Normalization.

Anecdotal evidence suggests that unfamiliar languages sound faster than one’s native language. Empirical evidence for this impression has, so far, come from explicit rate judgments. The aim of the present study was to test whether such perceived rate differences between native and foreign languages (FLs) have effects on implicit speech processing. Our measure of implicit rate perception was “normalization for speech rate”: an ambiguous vowel between short /a/ and long /a:/ is interpreted as /a:/ following a fast but as /a/ following a slow carrier sentence. That is, listeners did not judge speech rate itself; instead, they categorized ambiguous vowels whose perception was implicitly affected by the rate of the context. We asked whether a bias towards long /a:/ might be observed when the context is not actually faster but simply spoken in a FL. A fully symmetrical experimental design was used: Dutch and German participants listened to rate matched (fast and slow) sentences in both languages spoken by the same bilingual speaker. Sentences were followed by non-words that contained vowels from an /a-a:/ duration continuum. Results from Experiments 1 and 2 showed a consistent effect of rate normalization for both listener groups. Moreover, for German listeners, across the two experiments, foreign sentences triggered more /a:/ responses than (rate matched) native sentences, suggesting that foreign sentences were indeed perceived as faster. Moreover, this FL effect was modulated by participants’ ability to understand the FL: those participants that scored higher on a FL translation task showed less of a FL effect. However, opposite effects were found for the Dutch listeners. For them, their native rather than the FL induced more /a:/ responses. Nevertheless, this reversed effect could be reduced when additional spectral properties of the context were controlled for. Experiment 3, using explicit rate judgments, replicated the effect for German but not Dutch listeners. We therefore conclude that the subjective impression that FLs sound fast may have an effect on implicit speech processing, with implications for how language learners perceive spoken segments in a FL.

A pilot acoustic study of Modern Persian vowels in colloquial speech

While the current literature on Modern Persian (MP) shows heightened interest in its vowel inventory, a cogent synchronic account of Modern Persian vowels (MPV) has yet to be presented. Previous phonological descriptions of MPV presuppose a pairing system based on a historical length distinction that is still reflected in MP orthography: /i:, e/, /u:, o/, /ɒ:, æ/. A synchronic phonological analysis of MPV must be based on phonetic measurements; however, existing acoustic studies examine only formal speech, and thus do not reflect colloquial surface forms. Therefore, the goal of this investigation is to lay the groundwork for a comprehensive acoustic study of MPV that presupposes no pairings and is based on colloquial speech in a controlled prosodic environment. Vowel duration and first and second formants were measured using Praat for 90 CVC(C) monosyllables as pronounced phrase-finally in carrier sentences by two Tehrani native speakers: one male, one female. Results show that average durations for each vowel contradicted the traditional length-based pairings. Furthermore, features of the following phonological environment likely to affect vowel length — postvocalic consonant type, coda complexity (CVC vs. CVCC), and adherence to the Sonority Sequencing Principle in complex codas — had no notable effect. Results of formant measurements show the two non-high back vowels to be higher than expected, whereas the high back vowel was noticeably centralized.

The tonal system of Pahari

This paper reports on an acoustic study of the tonal system of the Pahari language. To achieve this aim, an experiment was conducted. Eight native speakers were given a set of monosyllabic triplets bearing three target tones to read them aloud for recording in a carrier sentence. The acoustic measures included F0, final velocity, and duration. The acoustic and statistical results show that (1) the average F0 demonstrates significant difference in the height of the pitch track of the three target words/tones; (2) final velocity shows three trends, namely falling, level, and rising associated with high, mid, and low tones, respectively; and (3) duration results indicate that high pitch track was significantly shorter than each of the other two pitch tracks. The study concludes that Pahari has three tones, namely high-falling, mid-level, and low-rising.

Gradient realization of Mandarin nasal codas

Impressionistic studies suggested that Mandarin nasal codas optionally delete before vowels (/dan/ + /ai/->[dã.ai]) and assimilate to the following stops in place (/dan/ + /pai/->[dam.pai]). In this EMA study, we found that both processes are not phonological changes but are gradient processes modulated by speech rates. Three native Mandarin speakers read disyllabic words with n/ŋ codas before /a/ and /p/ in a carrier sentence at three speech rates. The tongue tip trajectories show that both nasals retain a reduced tongue gesture before /a/. The reduction INCREASEs as speed INCREASEs and is more variant in faster speech. The lower lip trajectories show that the labial gesture of /p/ occurs earlier when preceded by nasal codas. The time-normalized duration of this gestural advance DECREASEs as speed INCREASEs (and decreases at a slower rate in slower speech), suggesting less gestural overlap at faster speed. Such a relationship is opposite to the reduction-speed relationship. The articulatory evidence shows that although not perceptual salient, there is a reduced tongue gesture for nasals before vowels and gestural overlap between stops and preceding nasals. The non-linear effects of speed on gestural reduction and overlap suggest both are gradient phonetic processes; yet, their opposite relationships to speed suggest different implementation mechanisms.

Velar-vowel coarticulation across the lifespan and in people who stutter: Findings and model

The study of anticipatory coarticulation provides insight into the speech production planning process. In the present study, the task involved repeating velar-vowel consonant combinations in a carrier sentence (e.g., for /ke/, “Say a cape again”). Data for velar-vowel coarticulation were analyzed using the Articulate Assistant Advanced software to create tongue traces that were quantified following the procedures for average of the nearest neighbor point-to-point distance between curves (Zharkova & Hewlett 2009, Journal of Phonetics). There were 126 participants total in child (8-12), young adult (18-39), and older adult (55-75) ages and groups of typical speakers (n = 21, 23, 29) and people who stutter (n = 15, 23, 11). Data analysis found a decrease in coarticulatory influence of the vowel on the velar across the lifespan, but no differences in coarticulation for people who stutter. Analysis of variability found greater variability for children and people who stutter. A two allophone model of coarticulation provided the best fit to the data, replicating Frisch & Wodzinski (2016, Journal of Phonetics).

Assessment of auralizations of monosyllabic words for hearing impaired students

This paper discusses results of a research project, which seeks to develop Turkish speech recognition tests by auralizations for hearing impaired students based on monosyllabically structured words. In the context of this study, two sets of 25-items phonetically balanced monosyllabic Turkish words were recorded in the anechoic chamber of TUBITAK National Metrology Institute in Gebze, Turkey. Each monosyllabic word was recorded through a carrier sentence. After the vocal quality, accent and pronunciation in the recordings were approved by qualified audiologists, auralizations of the recorded sentences were developed in an acoustic simulation software (ODEON v12). Listening tests developed from auralizations in three classroom models with varying reverberation times and signal to noise ratios were presented to ten hearing impaired students. The preliminary results of these listening tests are presented and discussed in this study. This paper also aims to compare the results of listening tests with hearing impaired students with the listening tests made previously with normal hearing students.

Embrace for impact: Formant reconstruction in sudden pitch raise while singing

Professional singers are trained to maintain vocal configuration by suppressing laryngeal elevation when singing high notes. The present study investigates the organization of vocal filter in sudden pitch raise condition by examining the corresponding acoustic correlates. More specifically, we are interested in whether this phenomenon may have differential effect on vowels with different height. In the experiment, high and low vowels were embedded in a nonsense Mandarin carrier sentence “kuai-lai-kuai-lai___yi-po” with C4E4G4E4___D4D4 pitch contour. Thirty amateur singers were recorded singing the sentence in four melodic conditions with the pitch intervals between the preceding syllable /lai/ to target word manipulated. The conditions are Micro (E4 to D4), Macro (E4 to D5), all High (every word sung in D5) and Null (only the last three words sung in D5-D5D5 contour). The results show that high pitch singing (Macro, High, and Null) indeed induces formant reconstruction when compared to Micro and the effec...

Planning of Hiatus-Breaking Inserted /ɹ/ in the Speech of Australian English–Speaking Children

Non-rhotic varieties of English often use /ɹ/ insertion as a connected speech process to separate heterosyllabic V1.V2 hiatus contexts. However, there has been little research on children's development of this strategy. This study investigated whether children use /ɹ/ insertion and, if so, whether hiatus-breaking /ɹ/ can be considered planned, as evidenced by F3 lowering on V1. Thirteen Australian English-speaking children (7 girls, 6 boys; mean age 6;1 [years;months]) participated in an elicited production task. The stimuli included carrier sentences containing 4 test words (linking /ɹ/ context: door, floor; intrusive /ɹ/ context: paw, claw) followed by of (e.g., "This is the paw of the cat"). After familiarization containing auditory and picture prompts, children produced test sentences upon presentation of picture prompts alone. Eight children produced /ɹ/ insertion; the others used (some) glottalization. The incidence of /ɹ/ did not vary across linking or intrusive contexts, and inserted /ɹ/ was associated with F3 lowering at V1 onset relative to control items without /ɹ/. Six-year-old Australian English-speaking children who use /ɹ/ insertion show evidence of planning ahead and inserting /ɹ/ as a segment. The implications for the development of speech-planning processes and phonological and lexical representations are discussed.

The representation and execution of articulatory timing in first and second language acquisition

The early acquisition of language-specific temporal patterns relative to the late development of speech motor control suggests a dissociation between the representation and execution of articulatory timing. The current study tested for such a dissociation in first and second language acquisition. American English-speaking children (5- and 8-year-olds) and Korean-speaking adult learners of English repeatedly produced real English words in a simple carrier sentence. The words were designed to elicit different language-specific vowel length contrasts. Measures of absolute duration and variability in single vowel productions were extracted to evaluate the realization of contrasts (representation) and to index speech motor abilities (execution). Results were mostly consistent with a dissociation. Native English-speaking children produced the same language-specific temporal patterns as native English-speaking adults, but their productions were more variable than the adults’. In contrast, Korean-speaking adult learners of English typically produced different temporal patterns than native English-speaking adults, but their productions were as stable as the native speakers’. Implications of the results are discussed with reference to different models of speech production.

Speaker variability in the realisation of lexical tones

While previous studies on the speaker-discriminatory power of static f0 parameters abound, few have focused on the dynamic and linguistically structured aspects of f0. Lexical tone offers a case in point for this endeavour. This article reports an exploratory study on the speaker-discriminatory power of individual lexical tones and of the height relationship of level tone pairs in Cantonese, and the effects of voice level and linguistic condition on their realisation. Twenty native Cantonese speakers produced systematically controlled words either in isolation or in a carrier sentence under two voice levels (normal and loud). Results show that f0 height and f0 dynamics are separate dimensions of a tone and are affected by voice level and linguistic condition in different ways. Moreover, discriminant analyses reveal that the contours of individual tones and the height differences of level tone pairs are useful parameters for characterising speakers.

Acoustic correlates to word order in Infant- and adult directed speech: A cross-linguistic study

The acoustic realization of phrasal prominence correlates systematically with the order of Verbs and Objects in natural languages. Prominence is realized as a durational contrast in V(erb)-O(bject) languages (English: short-long, to [Ro]me), and as a pitch/intensity contrast in O(bject)-V(erb) languages (Japanese: high-low, [‘To] kyo ni). Seven-month-old infants can use phrasal prominence to segment unknown artificial languages. This information might thus allow prelexical infants to learn the basic word order of their native language(s). The present study investigates whether this differing realization of phrasal prominence is also found in I(nfant) D(irected) S(peech), previously unexamined speech style. We recorded 15 adult native talkers of languages with opposite word orders producing target phrases (English: VO, e.g., behind furniture, Japanese: OV, e.g., jisho niwa) embedded in carrier sentences, both in A(dult) D(irected) S(peech) and IDS, and conducted acoustic analysis of the phrases (i.e., pitch maximum, mean pitch and intensity, duration). The results revealed the expected contrast in pitch in Japanese IDS and a similar trend in ADS. In English, the predicted durational contrast was found in ADS, additionally accompanied by a pitch contrast. Interestingly, a pitch contrast but no durational contrast was found in IDS.

Articulatory movement in consonant clusters: A contrastive study of Japanese and English

We investigated the articulation of consonant clusters with Japanese and English speakers using the WAVE system (NDI Corp.). Until now we measured the displacement of tongue tip only, but in this study we also measured the displacement of the mandible. The subjects were three Japanese and two English speakers. The speech samples were 4 words: blat, bnat, plat, pnat. All these second consonants, /l, n/ are articulated with tongue tip placed behind alveolar ridge. These words were embedded in a carrier sentence, “Say X.,” and these sentences were uttered 5 times each in random order. The articulation time from first consonants, /b/, /p/, to second consonants, /l/, /n/, were measured. Specifically, the time from the highest point of the mandible to that of the tongue tip was measured. Overall average time was 131.2 ms (Japanese) and 48.1 ms (English). Two-way ANOVA (Language vs. Cluster) showed a significant main effect in Language (p<0.0001). This result suggested that for English speakers, the first and second consonants are strongly co-articulated, while in Japanese speakers, they are not co-articulated or weakly co-articulated. The measurement of mandibular displacement more clearly showed the difference of the strength of co-articulation between Japanese and English.

Speaking Clearly for the Blind: Acoustic and Articulatory Correlates of Speaking Conditions in Sighted and Congenitally Blind Speakers.

Compared to conversational speech, clear speech is produced with longer vowel duration, greater intensity, increased contrasts between vowel categories, and decreased dispersion within vowel categories. Those acoustic correlates are produced by larger movements of the orofacial articulators, including visible (lips) and invisible (tongue) articulators. Thus, clear speech provides the listener with audible and visual cues that are used to increase the overall intelligibility of speech produced by the speaker. It is unclear how those cues are produced by visually impaired speakers who never had access to vision. In this paper, we investigate the acoustic and articulatory correlates of vowels in clear versus conversational speech, and in sighted and congenitally blind speakers. Participants were recorded using electroarticulography while producing multiple repetitions of the ten Quebec French oral vowels in carrier sentences in both speaking conditions. Articulatory variables (lip, jaw, and tongue positions) as well as acoustic variables (contrasts between vowels, within-vowel dispersion, pitch, duration, and intensity) were measured. Lip movements were larger when going from conversational to clear speech in sighted speakers only. On the other hand, tongue movements were affected to a larger extent in blind speakers compared to their sighted peers. These findings confirm that vision plays an important role in the maintenance of speech intelligibility.

Closure duration as an acoustic correlate of the word-initial singleton/geminate consonant contrast in Kelantan Malay

Closure duration has been established cross-linguistically as the universally most reliable and consistent acoustic feature of consonant gemination. In this study, we conduct an acoustic phonetic analysis of the word-initial singleton/geminate consonant contrast in Kelantan Malay (KM) in order to explore the extent to which closure duration marks such a contrast in this Malay variety. KM is particularly unusual among the world׳s languages in that the contrast is restricted to word-initial position. A corpus of elicited materials consisting of singleton/geminate voiceless stops, voiced stops and sonorants were produced in words in isolation (i.e., utterance-initial position) and in a carrier sentence (i.e., utterance-medial position) by sixteen native speakers of KM. Results show that closure duration is a robust acoustic correlate of the consonant contrast in KM, i.e., word-initial geminates in KM are always associated with significantly longer closure duration than their corresponding singletons. The closure duration differences between singletons and geminates are also similar across utterance positions (at least for voiced stops and sonorants). In utterance-medial contexts, the effect is particularly strong in the closure duration of voiceless stops. Overall, closure duration is indeed a highly robust acoustic marker for this cross-linguistically rare word-initial consonant contrast in KM.

Only Behavioral But Not Self-Report Measures of Speech Perception Correlate with Cognitive Abilities.

Good speech perception and communication skills in everyday life are crucial for participation and well-being, and are therefore an overarching aim of auditory rehabilitation. Both behavioral and self-report measures can be used to assess these skills. However, correlations between behavioral and self-report speech perception measures are often low. One possible explanation is that there is a mismatch between the specific situations used in the assessment of these skills in each method, and a more careful matching across situations might improve consistency of results. The role that cognition plays in specific speech situations may also be important for understanding communication, as speech perception tests vary in their cognitive demands. In this study, the role of executive function, working memory (WM) and attention in behavioral and self-report measures of speech perception was investigated. Thirty existing hearing aid users with mild-to-moderate hearing loss aged between 50 and 74 years completed a behavioral test battery with speech perception tests ranging from phoneme discrimination in modulated noise (easy) to words in multi-talker babble (medium) and keyword perception in a carrier sentence against a distractor voice (difficult). In addition, a self-report measure of aided communication, residual disability from the Glasgow Hearing Aid Benefit Profile, was obtained. Correlations between speech perception tests and self-report measures were higher when specific speech situations across both were matched. Cognition correlated with behavioral speech perception test results but not with self-report. Only the most difficult speech perception test, keyword perception in a carrier sentence with a competing distractor voice, engaged executive functions in addition to WM. In conclusion, any relationship between behavioral and self-report speech perception is not mediated by a shared correlation with cognition.

Visual feedback and second language segmental production: The generalizability of pronunciation gains

Abstract While a number of researchers have noted the lack of research on pronunciation instruction, relative to other aspects of language (i.e. syntax), pronunciation has been shown to be crucial for facilitating intelligible and comprehensible second language (L2) productions. Addressing the need for empirically tested pedagogical methods, the current study considers the use of a classroom-based visual feedback paradigm for the instruction of a segmental feature, namely voice onset time, which has been shown to be a distinctive marker of accent for English-dominant L2 learners of Spanish. In addition, this study examines the potential generalizability of gains made through the visual feedback paradigm, assessing whether gains made in controlled reading tasks (i.e. carrier sentences) will extend to more continuous and spontaneous speech. The results demonstrate significant improvements in voice onset time produced by participants following the visual feedback paradigm, relative to a control group. Furthermore, while the visual feedback training was limited to short, controlled utterances (i.e. carrier sentences), benefits were observed for more continuous and spontaneous speech.

The acoustics of strengthened glides in Kirundi

Strengthening of post-consonantal glides /w/ and /j/ to obstruents or nasals is cross-linguistically uncommon, especially when the process is synchronically conditioned. This study investigates acoustic correlates of post-consonantal glide strengthening in Kirundi, an eastern Bantu language spoken in Burundi. The materials included words with /w/ or /j/ preceded by labial obstruent or nasal consonants /p, b, v, m/—the contexts that have been previously described to trigger the process. The morphological context was also varied, with sequences occurring at prefix-root or suffix-root boundaries, or within a root. Four female and four male native speakers produced the word list in a carrier sentence three times. The analysis involved phonetic classification of strengthened glides, as well as their duration and spectral measurements (frication/burst spectra and following vowel formants). The results showed that the post-consonantal /w/ was consistently realized as an oral or nasal stop agreeing with the preceding consonant in voicing and nasality, consistently with previous descriptive accounts. The realization of the post-consonantal /j/, however, was considerably more variable across and within speakers, ranging from a palatal stop or nasal to a weakly fricated/nasalized glide, and showing some sensitivity to morphological structure.

Interaction of pitch and phonation in Louma Oeshi

Louma Oeshi is an Akoid (Tibeto-Burman) language of Laos for which acoustic characteristics are undocumented with the exception of preliminary work by the present authors. This study focuses on the phonetic properties associated with Oeshi's three tones (high, mid, and low) and two registers (Tense, Lax), which fully intersect to yield a six-way suprasegmental contrast. Eight speakers were recorded in Phongsali Province, Lao PDR. Each spoke a 100 token word list, for which 30 tokens were repeated in a carrier sentence placing the token between high and mid-toned lax words. Decile measures of F0 and an array of measures reflecting phonation types (H1-H2, H1-A1, H1-A3, HNR, and SHR) were performed in Voicesauce (Shue et al. 2011) to capture dynamic values over the syllable duration. Our findings show a reliable three-way F0 contrast between tones and a single interaction with register such that High Tense words consistently fall (tones are otherwise level). Acoustic correlates of the Tense~Lax distinction a...

The acoustic correlates of lexical stress in Israeli Hebrew

Lexical stress is an omnipresent phenomenon in spoken language, serving in many cases to disambiguate lexically identical words. Lexical stress in Spoken Israeli Hebrew is usually either penultimate or final, however its prosodic cues have not been measured systematically. In the present study the acoustics of lexical stress were characterized in detail. A list of 34 two-syllable stress-based minimal pairs was collected, where each word form has a different meaning. Each lexical word was embedded in a carrier sentence, thus creating 68 sentences, in which each word-form appeared once in sentence-middle position. Thirty speakers, gender balanced, uttered each sentence, giving a total of 2040 utterance productions. All syllable and vowel boundaries of the target words were annotated manually, and three acoustical parameters – duration, f0 and intensity – were measured for each nucleus vowel. Statistical analysis revealed that vowel duration was the dominant marker of Hebrew lexical stress; f0 played a minor role in indicating stress; while intensity played a more prominent role than f0. Regression analysis showed that the above three cues explained 80% of the variance, with duration contributing 77% and intensity and f0 1.5% each.

Visual Speech Perception in Children With Language Learning Impairments.

The purpose of the study was to assess the ability of children with developmental language learning impairments (LLIs) to use visual speech cues from the talking face. In this cross-sectional study, 41 typically developing children (mean age: 8 years 0 months, range: 4 years 5 months to 11 years 10 months) and 27 children with diagnosed LLI (mean age: 8 years 10 months, range: 5 years 2 months to 11 years 6 months) completed a silent speechreading task and a speech-in-noise task with and without visual support from the talking face. The speech-in-noise task involved the identification of a target word in a carrier sentence with a single competing speaker as a masker. Children in the LLI group showed a deficit in speechreading when compared with their typically developing peers. Beyond the single-word level, this deficit became more apparent in older children. On the speech-in-noise task, a substantial benefit of visual cues was found regardless of age or group membership, although the LLI group showed an overall developmental delay in speech perception. Although children with LLI were less accurate than their peers on the speechreading and speech-in noise-tasks, both groups were able to make equivalent use of visual cues to boost performance accuracy when listening in noise.