Showing papers on "Voice published in 2007"

The Origin of Vowel Length Neutralization in Final Position: Evidence from Finnish Speakers

Abstract: In a broad variety of languages with contrastive vowel length, long vowels are systematically excluded from a domain-final position, and are replaced with short vowels there This is despite the fact that vowels at the end of a domain (utterance, phrase, word) are generally longer in duration than corresponding nonfinal vowels We propose that the phonological pattern of final shortening arises diachronically from the effects of final devoicing – the breakdown in voicing at the end of an utterance Partial devoicing of the final vowel makes it difficult to hear the end of the vowel and so favors identification of final vowels as short If language learners generalize such an identification pattern, they have adopted a final shortening pattern The claim that partially voiceless final vowels tend to be identified as short is supported by a series of experiments with Finnish speakers The first two experiments establish that there is both final lengthening and final devoicing in the language Three further experiments show that Finnish speakers identify the length category of partially voiceless final vowels on the basis of the duration of its voiced portion, so that partial devoicing of a vowel increases the probability of its being identified as short

Paradigmatic effects in auditory word recognition : The case of alternating voice in Dutch

Abstract: Two lexical decision experiments addressed the role of paradigmatic effects in auditory word recognition. Experiment 1 showed that listeners classified a form with an incorrectly voiced final obstruent more readily as a word if the obstruent is realised as voiced in other forms of that word's morphological paradigm. Moreover, if such was the case, the exact probability of paradigmatic voicing emerged as a significant predictor of the response latencies. A greater probability of voicing correlated with longer response latencies for words correctly realised with voiceless final obstruents. A similar effect of this probability was observed in Experiment 2 for words with completely voiceless or weakly voiced (incompletely neutralised) final obstruents. These data demonstrate the relevance of paradigmatically related complex words for the processing of morphologically simple words in auditory word recognition.

Caught in the ACT: the timing of aspiration and voicing in East Bengali.

Abstract: East Bengali is a language that displays a four-way contrast of voiced/voiceless and aspirated/unaspirated oral stops and affricates in all word positions. Additionally, in intervocalic position there is a quantity contrast between long and short obstruents.In this production study we investigate medial palato-alveolar affricates and stops at the labial, dental, retroflex, and velar places of articulation and address the problems of VOT measurements. We introduce a different approach of measuring lag times, henceforth called after closure time (ACT). The results show that this approach can do away with the extra notion of breathy voice to distinguish between the voiced aspirates and unaspirates.Moreover, as a result of analyzing the aspirated stops and affricates, an additional term (superimposed aspiration—SA) had to be introduced. The results of combining the notions of ACT and SA show that aspiration, measured from the point of release, is timed equally for voiced and voiceless stops. However, the diff...

Categorical Perception of Speech (VOT) and Analogous Non-Speech (FOT) signals: Behavioral and electrophysiological correlates.

Abstract: Objective To determine whether voicing perception is influenced primarily by linguistic experience or if it is due to innate temporal sensitivity to voicing boundaries, by examining behavioral and electrophysiological correlates of speech Voice-Onset-Time (VOT) and nonspeech Formant-Onset-Time (FOT) categorical perception. Design Behavioral measures and auditory event-related potentials (ERPs) were obtained from 14 normal-hearing Hebrew speakers, whose voicing distinction is different than English, during identification and discrimination of two sets of stimuli: a VOT continuum, created by editing natural productions of /ba/ and /pa/, and an analogous nonspeech continuum, composed of two synthesized formants, varying in their onset time-FOT. Results VOT and FOT continua yielded similar behavioral identification curves. Differences between the two stimulus types were found in discrimination of within-category differences and in reaction time effects. During identification and discrimination tasks, ERPs were differently affected by the VOT or FOT value of the stimulus: VOT value had a significant effect on N1 latency and on N1 and P2 amplitudes whereas FOT value had a significant effect on P2 amplitude. Additionally, during identification tasks, whereas all speech signals evoked a P3, regardless of overt categorization, only the perceptually "rare" nonspeech stimulus (+15 msec FOT) evoked a P3. Conclusions Voicing boundaries corresponded to Hebrew VOT values of production, suggesting that voicing perception in Hebrew is mediated mainly by linguistic experience rather than by innate temporal sensitivity. ERP data differed to VOT versus FOT stimuli as early as N1, indicating that brain processing of the temporal aspects of speech and nonspeech signals differ from their early stages. Further studies to establish the neural response patterns to voicing in speakers of languages that use different voicing categories than English are warranted.

VOT productions of word-initial stops in Mandarin and English: A cross-language study

Abstract: Voice Onset Time (VOT) is considered as one of the best methods for examining the timing of voicing in stop consonants and has been applied in the study of many languages The present study is designed to examine VOT production for phonetically voiceless stops in Mandarin and English by native Chinese speakers Thirty-six Taiwanese Chinese speakers recruited from National Cheng Kung University participated in this study The results indicate the following 1) Based on the three universal categories proposed by Lisker and Abramson (1964), for phonetically voiceless stops, Mandarin and English occupy the same place along the VOT continuum 2) The mean VOT value for the apical stop /t/ is slightly lower than the mean value for the labial stop /p/ This does not conform to the general consensus, which states that the further back the place of articulation the longer the VOT Very similar findings were also observed in previous studies 3) The difference between the mean VOT values of the English /p/ and /t/ produced by Chinese speakers was subtle, while it reached significance for native English speakers This suggests that a first language could be a crucial factor in L2 production Future studies might examine variations in L2 production both for the same persons over time and for different speakers

Locus and Nature of Perceptual Phonological Deficit in Spanish Children With Reading Disabilities

Abstract: The aims of this study were (a) to determine whether Spanish children with reading disabilities (RD) show a speech perception deficit and (b) to explore the locus and nature of this perceptive deficit. A group of 29 children with RD, 41 chronological age-matched controls, and 27 reading ability-matched younger controls were tested on tasks of speech perception. The effect of linguistic unit (word vs. syllable) and type of phonetic contrast (voicing, place and manner of articulation) were analyzed in terms of the number of errors and the response time. The results revealed a speech perception deficit in Spanish children with RD that was independent of the type of phonetic contrast and of linguistic unit.

Metrical influences on fortition and lenition

Abstract: There is considerable tacit agreement among phonologists and phoneticians about the prototypical uses of the terms ‘lenition’ and ‘fortition’. However, in the phonetic dimension the terms do not refer to a single unified phenomenon, but instead to manipulations of two independent parameters: duration and magnitude (degree of consonantal stricture) (Lavoie 2001). In phonology, ‘lenition/fortition’ refers to the categorical effects of such adjustments in duration and magnitude once they have been phonologized. The coherence of the terms is thus essentially diachronic; synchronically, many different formally unrelated processes may implement the phonologization. On the synchronic phonological level, then, ‘fortition’ and ‘lenition’ are no more than taxonomic labels that provide little insight into the cognitive processes involved. The processes covered by the terms are not formally unified (at least given standard representational assumptions; for a contrasting view according to which phonological representations are phonetically detailed, see Kirchner 1998). Both terms may refer to a range of distinct processes, including changes in phonological segment weight (gemination/ degemination through insertion/ deletion of moras), sonority or continuancy (occlusion/ spirantization through change in the value of the feature [continuant], and so on), and voicing. The last of these may be connected to the observation that decreases in duration in obstruents may give rise to voiced percepts (Cole & Cooper 1975); accordingly, some languages have phonologized decrease in obstruent closure duration as a

Estimation of Voicing-Character of Speech Spectra Based on Spectral Shape

Abstract: This letter presents a method for estimation of the voicing-character of speech spectra. It is based on a calculation of a similarity between the shape of the signal short-term magnitude spectra and spectra of the frame-analysis window, which is weighted by the signal magnitude spectra. It is demonstrated that the proposed voicing measure is related to the local SNR of noise-corrupted voiced speech. The performance is evaluated for detection of voiced regions in the spectra of speech corrupted by various types of noise. The experimental results in terms of false-acceptance and false-rejection show errors of less than 5% for speech corrupted by white noise at the local SNR of 10 dB and in terms of recognition accuracy obtained by an ASR system using the voicing information estimated by the proposed method and by the full a priori knowledge about the noise show similar recognition performance

Phonetic variation as communicative system: Perception of the particular and the abstract

Abstract: This paper addresses three issues raised by observations of the role of systematically-varying fine phonetic detail in speech: the representation of linguistic knowledge so that it simultaneously accommodates detail and abstraction; the nature of phonetic categories; and the type of processes that must underlie speech perception. Evidence is adduced from the anatomy and physiology of the brain to support arguments for a polysystemic view of speech perception. 1. The particular 1.1 History and definition of the term "fine phonetic detail" (FPD): When John Local talked in the 1980s about fine phonetic detail, he was describing phonetic phenomena such as resonances associated with liquid consonants in English that were systematically distributed but not systematically treated in conventional phonetic description. They tended to be hard to notice too. Since then, the term FPD has come to be applied to anything that is not considered a major, usually local, perceptual cue for phonemic contrasts in the citation forms of lexical items. In this broader usage, some FPD is indeed 'fine', and subtle, but other types are perfectly audible; they have just not been factored into the prevailing view that perceptual processing of phonetic information is largely aimed at identifying strings of features or phonemes that allow words to be distinguished. FPD cuts across traditional subdisciplines of enquiry. It does not just distinguish words, but also the wider phonological and grammatical structure of the message. For example, grammatical function words have a narrower range of sound patterns than content words, and undergo different connected speech processes; and each type of function word (e.g. auxiliary verbs, articles) has its own distinct system of contrasts (e.g. Ogden 1999; Local 2003). FPD also reflects function and structure of the smaller units that comprise words, and of larger groupings, influencing everything necessary for successful communication: phonological, morphological, grammatical, pragmatic, interactional (see Phonetica Volume 61, especially papers by Local, Ogden, and Plug). FPD indicating a single linguistic distinction can involve many acoustic properties distributed over long stretches of speech (Hawkins and Smith 2001; Local 2003) e.g. traces of English /r/ can occur several syllables before the main /r/ segment and influence perception (Hawkins and Slater 1994; West 1999a, b; Heid and Hawkins 2000; Coleman 2003). Even well-researched distinctions like coda voicing involve multiple distinctions, some of which are less local than was until recently assumed (Hawkins and Nguyen 2004). Thus, much FPD—the sort discarded by traditional abstractionist phonological and perceptual models as uninteresting or due to random effects—in fact systematically reflects many different aspects of meaning that are crucial to the maintenance of normal conversation: lexical, grammatical, and interactional differences. As Nguyen notes "The goal of current research on FPD is to show that FPD is important in speech (processing), and, therefore, that a change of theoretical perspective is called for." (in press :8). If such a changed perspective proves valuable, the term fine phonetic detail can largely be replaced by "phonetic information".

Children's Weighting Strategies for Word-Final Stop Voicing Are Not Explained by Auditory Sensitivities.

Abstract: Purpose It has been reported that children and adults weight differently the various acoustic properties of the speech signal that support phonetic decisions. This finding is generally attributed t...

Perceptual rate normalization in naturally produced rate-varied speech.

Abstract: The perception of voicing categories is affected by speaking rate, so that listeners’ category boundaries on a VOT continuum shift to a lower value when syllable duration decreases [Miller and Volaitis, Percept. Psychophys. 46, 505–512 (1989); Volaitis and Miller, J. Acoust. Soc. Am. 92, 723–735 (1992)]. Previous rate normalization effects have been found using artificially varied stimuli. This study examines the effect of speech rate on voicing categorization in naturally produced rate-varied speech. The stimuli contained natural decreases in VOT with faster speech rates so that VOT values for /b/ and /p/ overlapped at the fastest rates. Consonant identification results showed that the rate effects on the perceptual boundary between /p/ and /b/ very closely matched the effects of rate on the productions, though there was a small mismatch with fast rate productions whereby voiced stops were systematically miscategorized as voiceless. Another group of listeners judged the goodness of the consonant, indicat...

Aspiration and voicing of chinese and english plosives

Abstract: The six plosives of Standard Chinese are compared with those of RP British English, to see if there is a difference in their aspiration and/or voicing. Recordings of 7 speakers from China reading words beginning with each of the 6 plosives are compared to similar recordings of 7 speakers of RP British English, and it is found that there is little difference in the aspiration of the plosives in the two languages, though there is a difference in the voicing during the closure when the plosive occurs between two vowels.

Acquiring /[alveolar approximant]/ in Context.

Abstract: This article seeks to illuminate the degree of position-based variation observed in the acquisition of new segments in a second language and to explain such variability as the consequence of phonetic constraints; this approach contrasts with much previous research that has used typological markedness to the same end. Specifically, it is proposed that learners will have the least difficulty acquiring sounds that involve novel combinations of voicing and manner in positions that favor the phonetic implementation of these sounds. Moreover, on the assumption that not all parameters can be mastered simultaneously, it is predicted that learners will first acquire aspects of a segment's articulation that are perceptually salient and articulatorily easier. The data come from a study of the acquisition of French by 20 intermediate- and advanced-proficiency English-speaking learners of French. Acoustic analysis of the data reveals asymmetries that favor accuracy with manner in onsets versus more targetlike realization of voicing in codas, in which devoicing exists in the input. Beyond demonstrating the role of phonetic principles in determining position-based variation, the findings contribute to our understanding of the acquisition of new consonantal contrasts by providing empirical evidence from a non-Germanic language to bear on this line of inquiry.

Voicing-dependent cluster simplification asymmetries in Spanish and French

Abstract: In the present work, we build upon the proposal outlined in Colantoni & Steele (2005b) that asymmetrical patterns of Spanish and French stop-liquid cluster simplification are conditioned by liquid type and stop voicing. Specifically, using data from four Romance varieties (Argentine and Chilean Spanish; European and Quebec French), we show that the longer epenthetic vowel in Spanish voiced versus voiceless stop-rhotic clusters and the restriction of voicing assimilation to voiceless stop-rhotic clusters in French can be explained with reference to asymmetrical stop length as it interacts with consonant sequence timing, and constraints on voicing in fricatives and dorsals respectively. The factors shown to condition synchronic variation can be extended to explain the evolution of stop-liquid clusters from Latin to Romance.

Temporal and spatial aspects concerning the realizations of the voicing contrast in German alveolar and postalveolar fricatives

Abstract: This study investigates the phonetic realizations of voicing contrast in alveolar and postalveolar fricatives production in different word positions in order to understand the temporal and spatial production strategies used in the control of voicing and frication, and to provide a frame of reference for speech therapy despite the inter-speaker variation. Seven native speakers of German, originally coming from various regions, participated in the experiment. Acoustic signals were recorded onto DAT, and tongue palate contact patterns were recorded by means of electropalatography (EPG). The temporal parameters were measured using the acoustic signals and the spatial parameters were measured based on the EPG data. The corpus included real words with // occurring at word initial, medial and final positions. Temporal results showed that differences in the overall frication duration for voicing contrast occur at almost all positions (with longer duration for voiceless phonemes). However, voicing during the frica...

Voicing in Dutch : (de)voicing -- phonology, phonetics, and psycholinguistics

Abstract: 1. Introduction 2. 1. Issues in Dutch devoicing: Positional faithfulness, positional markedness, and local conjunction (by Zonneveld, Wim) 3. 2. Representations of [voice]: Evidence from acquisition (by Kager, Rene) 4. 3. Exceptions to final devoicing (by Oostendorp, Marc van) 5. 4. Prevoicing in Dutch initial plosives: Production, perception, and word recognition (by Alphen, Petra M. van) 6. 5. Dutch regressive voicing assimilation as a 'low level phonetic process': Acoustic evidence (by Jansen, Wouter) 7. 6. Intraparadigmatic effects on the perception of voice (by Ernestus, Mirjam) 8. Indexes

On the Correlation between Perceptual and Contextual Aspects of Laughter in Meetings

Abstract: We have analyzed over 13000 bouts of laughter, in over 65 hours of unscripted, naturally occurring multiparty meetings, to identify discriminative contexts of voiced and unvoiced laughter Our results show that, in meetings, laughter is quite frequent, accounting for almost 10% of all vocal activity effort by time Approximately a third of all laughter is unvoiced, but meeting participants vary extensively in how often they employ voicing during laughter In spite of this variability, laughter appears to exhibit robust temporal characteristics Voiced laughs are on average longer than unvoiced laughs, and appear to correlate with temporally adjacent voiced laughter from other participants, as well as with speech from the laugher Unvoiced laughter appears to occur independently of vocal activity from other participants

Acquisition of morpho-phonology : the Dutch voicing alternation

Abstract: This dissertation investigates how and when the Dutch voicing alternation is acquired. In Dutch, final neutralisation of the voice contrast (or ‘final devoicing’) leads to alternations in singular - plural pairs such as bed ~ bedden ‘bed(s)’. In such pairs, the singular always ends in a voiceless obstruent, whereas the plural contains a voiced obstruent. Knowledge of voicing alternations can only be acquired on the basis of the alternating plural form, which needs to be related to the neutralised singular. The acquisition of the voicing alternation was investigated using corpus data (based on CELEX and CHILDES) as well as experimental evidence. Experiments consisted of elicitation of both plurals and singulars for words and non-words, to investigate children’s productive knowledge of the alternation. The research was aimed at testing predictions of rule- and constraint-based models on the one hand and analogical or usage-based models on the other. Results show that the alternation is acquired relatively late, as children produce errors well after the age of six. The pattern further shows limited productivity, as it is not often extended to non-words. It is argued that the results are interpreted most adequately in a usage-based model such as Bybee’s network model, in which frequency-based generalisations or ‘schemas’ arise over stored words in the lexicon. This dissertation is of relevance to researchers in the fields of phonology, morphology, psycholinguistics and language acquisition.

Bayesian framework for voicing alternation & assimilation studies on large corpora in french

Abstract: The presented work aims at exploring voicing alternation and assimilation on very large corpora using a Bayesian framework. A voice feature (VF) variable has been introduced whose value is determined using statistical acoustic phoneme models, corresponding to 3-state Gaussian mixture Hidden Markov Models. For all relevant consonants, i.e. oral plosives and fricatives their surface form voice feature is determined by maximising the acoustic likelihood of the competing phoneme models. A voicing alternation (VA) measure counts the number of changes between underlying and surface form voice features. Using a corpus of 90h of French journalistic speech, an overall voicing alternation rate of 2.7% has been measured, thus calibrating the method’s accuracy. The VA rate remains below 2% word-internally and on word starts and raises up to 9% on lexical word endings. In assimilation contexts rates grow significantly (> 20%), highlighting regressive voicing assimilation. Results also exhibit a weak tendency for progressive devoicing.

Changes in audio-visual speech perception during adulthood

Abstract: Audiovisual speech perception research has shown an increasing use of visual information from infancy to young adulthood. The current study extends these findings by examining audiovisual speech perception from young adulthood to mid-adulthood by addressing the extent to which audio, visual and audiovisual cues are used for place of articulation identification. Responses were gathered with young adults (19-30 yrs) and mid-aged adults (49-60 yrs) for voiceless and voiced audiovisual consonant-vowel syllables differing in consonant place of articulation. Materials were presented in quiet and in cafe noise (SNR=0dB). Results show that mid-aged adults made greater use of visual information than young adults in both quiet and noise, suggesting more than compensation for natural changes in hearing. This was evident across places of articulation and voicing conditions where mid-aged adults showed further indications for using visual cues. Findings indicate that the processing of sensory information continues to change in the course of adulthood with the use of visual cues in audiovisual speech perception increasing with the experience that comes with age.

Acquisition of Vowel Duration in Children Speaking American English

Abstract: This study is an acoustic investigation of the acquisition of vowel duration in children speaking American English. The primary goal was to find out when and how children begin to produce different vowel durations as a function of postvocalic voicing. A total of 803 longitudinal data extracted from the Providence Corpus were analyzed. The age range covered by the data was from 0;11 to 4;0. The findings are summarized as follows: (1) Children control the vowel duration conditioned by voicing before the age of 2. (2) They also make the durational distinction between the tense and lax vowels before the age of 2. (3) There is no developmental trend in the acquisition of the vowel duration conditioned by postvocalic voicing. The results suggest that children thoroughly learn the phonetic implementation of temporal parameter from the very early stage of speech production to such an extent as to make it appear as an automatic process. Index Terms: vowel duration, voicing, tense/lax, corpus, acquisition, child(ren)

Canadian Raising in Victoria, B.C. An Acoustic Analysis

Abstract: This article presents the results of a study of the phonetic realization of Canadian Raising, as produced by 39 speakers from Victoria, B.C. A total of 2489 tokens of the diphthongs /aI/ and /ao/ in nine different phonetic environments were analyzed acoustically by measuring the first and second formants (F1/F2). Contrary to the findings of Chambers & Hardwick (1986) and Hung, Davison & Chambers (1993), which predicted the "Americanization" of Canadian speech and a decline in the use of raising, Canadian Raising is still in existence for both diphthongs. Raising chiefly occurs in the phonetic environment of a following voiceless consonant, including the 'flapped' or 'voiced' /t/, and involves a lower F1 (i.e. raising) of the offglides as well. Moreover, a trend toward fronting of the onsets can be discerned, with /ao/ being predominantly realized as fronted already, and /aI/ increasingly becoming so, especially in the phonetic environment of a following obstruent.

Quantifying vowel onset periodicity in japanese

Abstract: Many researchers will agree that the brain must employ some expectancy mechanism for speech as speech unfolds through time. We, among others, posit that an adaptive oscillator may be made to synchronize with these speech expectancies. The oscillator would define something of a cognitive pulse and should align itself with key features of the acoustic signal. A recent study we conducted on compensatory mora relationships between neighbor voicing inter-onset-intervals in Japanese strongly indicates voice onsets as important targets for a referent pulse or planning mechanism. We review that study and draw on its results to highlight some issues related to modeling a referent or cognitive pulse. Based on a circular statistics foundation, we note how some vowel onsets should be treated as strong coupling targets for an adaptive oscillator while other vowel onsets should be treated more as distractions. From there we discuss some problems and issues associated with analyzing speech for forms of periodic regularity.