Human Speech Perception 2

Fundamental Frequency Fluctuation in Continuous Vowel Utterance and its Perception

Authors:

Masato Akagi, Japan Advanced Institute of Science and Technology (Japan)
Mamoru Iwaki, Japan Advanced Institute of Science and Technology (Japan)
Tomoya Minakawa, Japan Advanced Institute of Science and Technology (Japan)

Page (NA) Paper number 27

Abstract:

This paper reports how rapid fluctuations of fundamental frequencies in continuously uttered vowels influence vowel quality and shows that vowel qualities with various fundamental frequency fluctuations can be discriminated perceptually. For this purpose, electroglottographs (EGGs) of vowels uttered by nine males were obtained using Laryngograph, and fundamental frequencies with rapid fluctuations were estimated from them. Analyzing forty-five estimated fundamental frequencies, they can be classified into four groups. Moreover, psychoacoustic experiments, with five subjects, evaluating voice quality by multidimensional scaling (MDS) showed that voice quality of the synthesized speech using the fundamental frequencies of the groups was completely discriminable and there was a distinctive frequency band of fundamental frequency fluctuation for specifying each group perceptually.

SL980027.PDF (From Author) SL980027.PDF (Rasterized)

TOP

Estimation of Mental Lexicon Size with Word Familiarity Database

Authors:

Shigeaki Amano, NTT Basic Research Labs. (Japan)
Tadahisa Kondo, NTT Basic Research Labs. (Japan)

Page (NA) Paper number 15

Abstract:

A familiarity database was developed for about 80,000 Japanese words of which familiarity scores were rated by 32 Japanese adults using a 7-point scale in auditory, visual, and audio-visual modalities. Auditory, visual, and audio-visual stimulus words were selected from the database according to their word familiarity for size estimation of the mental lexicon. Sixty Japanese adults participated in a two-alternative forced-choice task (Know-Don't know) for the stimulus words. The size of the mental lexicon was estimated as the number of words of which familiarity is above a particular word corresponding to 50% point on the fitted logistic curve to "know"-response probability of the stimulus words. The estimated size was about 68,000 for auditory words, and about 66,000 both for visual and audio-visual words when homophones and homographs were included. The results suggest that very small difference in the mental lexicon size among modalities.

SL980015.PDF (From Author) SL980015.PDF (Rasterized)

TOP

Vowel Quality in Spontaneous Speech: What Makes a Good Vowel?

Authors:

Matthew Aylett, Human Communication Research Centre, University of Edinburgh (U.K.)
Alice Turk, Department of Linguistics, University of Edinburgh (U.K.)

Page (NA) Paper number 824

Abstract:

Clear speech is characterised by longer segmental durations and less target undershoot which results in more extreme spectral features. This paper deals with the clarity of vowels produced in spontaneous speech in a large corpus of task-oriented dialogues. We present an automatic technique for measuring vowel clarity on the basis of a vowel's spectral characteristics. This technique was evaluated using a perceptual test. Subjects rated the 'goodness' of vowels with different spectral characteristics with controlled duration and amplitude and these results were compared with an automatic rating. Results indicated that although agreement between subjects and the automatic measurement was poor it was as poor as the agreement between subjects. On the basis of these results we address the following questions: 1. Can subjects reliably judge the clarity of vowels excerpted from spontaneous speech without duration cues? 2. Can a statistical model reliably predict the subjects' response to such vowels?

SL980824.PDF (From Author) SL980824.PDF (Rasterized)

TOP

Cooperation and Competition of Burst and Formant Transitions for the Perception and Identification of French Stops

Authors:

Adrian Neagu, Institut de la Communication Parlee, Grenoble (France)
Gérard Bailly, Institut de la Communication Parlee, Grenoble (France)

Page (NA) Paper number 1009

Abstract:

In this paper, we study the influence of the vocalic context on the perception and automatic recognition of stops. In a previous perception experiment [1] using conflicting cues stimuli, we have shown that place of articulation cued by formant transitions may be overwritten by the place cued by the burst. This effect is inversely proportional to the vowel aperture. Here we give special attention to /i/ context where nor burst, nor formant transitions seem to carry rich information on place of articulation. We present here automatic recognition experiments that confirm perception results. Taking into account both segments increase identification rates, early fusion of segmental cues performs best and most errors come from the front unrounded vocalic context. We introduce the "burst characteristic frequency" (BF) that palliates for the poor discriminative power of the traditional cues in the front context. Moreover we present perception results showing the perceptual relevance of BF.

SL981009.PDF (From Author) SL981009.PDF (Rasterized)

TOP

The Effect of Modifying Formant Amplitudes on the Perception of French Vowels Generated by Copy Synthesis

Authors:

Anne Bonneau, LORIA-CNRS (France)
Yves Laprie, LORIA-CNRS (France)

Page (NA) Paper number 260

Abstract:

With synthetic stimuli copied from natural vowels and including up to five formants, we investigated some transformations of the perceived identity of vowels by means of modifications of formant amplitude levels. Synthetic stimuli were generated by means of a new method of copy synthesis. With stimuli copied from /u/ we found that, despite the presence of F2, it is possible to transform the timbre of /u/ into that of a front vowel only by raising the amplitude level of F3 and higher formants. We also analysed how the timbre of the vowels /i/ and /y/ changed as a function of the level of F2, F3 and F4, and showed how sensitive was the timbre of the vowel /i/ to the decrease of the level of F3 and F4. Such transformations, realized with stimuli very close to natural vowels, reinforce the importance of formant amplitude levels in some vocalic distinctions.

SL980260.PDF (From Author) SL980260.PDF (Rasterized)

TOP

Segmental and Tonal Processing in Cantonese

Authors:

Hsuan-Chih Chen, Department of Psychology, The Chinese University of Hong Kong (China)
Michael C.W. Yip, Department of Psychology, The Chinese University of Hong Kong (China)
Sum-Yin Wong, Department of Psychology, The Chinese University of Hong Kong (China)

Page (NA) Paper number 660

Abstract:

In a tone language, such as Cantonese, both segmental and tonal distinctions between words are pervasive. However, previous work in Cantonese has demonstrated that in speeded-response tasks, tone is more likely to be misprocessed than is segmental structure. The present study examined whether this tone disadvantage would also hold after the initial auditory processing of a syllable had been done. Cantonese listeners were asked to perform same-different judgments on two sequentially presented open syllables along a specific dimension (i.e., onset, rime, tone, or the whole syllable) according to an instruction which was visually presented at the acoustic offset of the second syllable. Manipulating whether the difference between two syllables was in onset, rime, or tone resulted in equally robust effects across the various decision tasks on performance, indicating that tone functions as effectively as segmental structure in spoken-word processing once the related information of a syllable is encoded.

SL980660.PDF (From Author) SL980660.PDF (Rasterized)

TOP

Phonological Similarity Effects in Cantonese Spoken-Word Processing

Authors:

Michael C.W. Yip, Department of Psychology, The Chinese University of Hong Kong (China)
Po-Yee Leung, Department of Psychology, The Chinese University of Hong Kong (China)
Hsuan-Chih Chen, Department of Psychology, The Chinese University of Hong Kong (China)

Page (NA) Paper number 661

Abstract:

A Cantonese experiment is described in which the shadowing of spoken targets as a function of phonological similarity to either a succeeding prime (backward priming) or a preceding prime (forward priming) is investigated. In the backward priming conditions, alternations of onset, rime, or tone between prime and target produced inhibition, whereas in the forward priming conditions, alternations of tone led to facilitation. The results are discussed in terms of the processing and memory of Cantonese syllables.

SL980661.PDF (From Author) SL980661.PDF (Rasterized)

TOP

On The Learnability Of The Voicing Contrast For Initial Stops

Authors:

Bob I. Damper, University of Southampton (U.K.)
Steve R. Gunn, University of Southampton (U.K.)

Page (NA) Paper number 843

Abstract:

The categorical perception (CP) of syllable-initial stop consonants has been intensively studied using psychophysical procedures over many decades. However, computational models consisting of an auditory `front end' and a learning system as a `back end' convincingly mimic the essentials of CP. Unlike real listeners, such models can be systematically manipulated to uncover the basis of their categorisations. In this paper, we explore the use of modern inductive learning techniques in simulating CP.

SL980843.PDF (From Author) SL980843.PDF (Rasterized)

TOP

Acoustic and Perceptual Characteristic of Italian Stop Consonants

Authors:

Loredana Cerrato, Fondazione Ugo Bordoni (Italy)
Mauro Falcone, Fondazione Ugo Bordoni (Italy)

Page (NA) Paper number 463

Abstract:

We report the results of a study carried out to analyse the acoustic and perceptual characteristics of Italian stop consonants. The aim of this study is twofold: give an acoustical description of Italian stops and investigate which are the perceptual cues relative to their place of articulation. From the acoustic point of view we report: the measurements relative to the length of the whole consonant and of its release burst; the F1 and F2 of the following vowel measured at the beginning of it. Moreover we counted the presence of the release burst and we tried to describe its acoustical characteristics in terms of the spectral structure. From the perceptual point of view we report the results of three perceptual tests that we run with the aim of evaluating whether the release burst or the formant transitions are more relevant for the perception of Italian stop consonants' place of articulation.

SL980463.PDF (From Author) SL980463.PDF (Rasterized)

TOP

Acoustic Cues for the Auditory Identification of the Spanish Fricative /f/

Authors:

Santiago Fernández, Departamento de Física Aplicada. Universidad de Santiago de Compostela (Spain)
Sergio Feijóo, Departamento de Física Aplicada. Universidad de Santiago de Compostela (Spain)
Ramon Balsa, Departamento de Física Aplicada. Universidad de Santiago de Compostela (Spain)
Nieves Barros, Departamento de Física Aplicada. Universidad de Santiago de Compostela (Spain)

Page (NA) Paper number 451

Abstract:

This study deals with the distinction of the fricative noises of the spanish fricatives /th/ and /f/. Previous studies revealed that fricative noises of both phonemes are perceptually similar, auditory identification being significantly dependent on contextual effects: /f/ in the /u/ context is well identified (about 85% correct identification rate), while in the /e/ context identification is much lower (about 60%). Identification of /th/ is low for every vocalic context (about 60%). These effects were identical for both Hypo and Hyper forms of speech. The objective of this paper is to determine which acoustic properties of /f/ in the /u/ context make it a well defined phoneme for the two different forms of speech. We conclude that the cues for the identification of the isolated fricative noise of /f/ seem to be in the low frequency region of the spectrum.

SL980451.PDF (From Author) SL980451.PDF (Rasterized)

TOP

Recognition of Vowels in Fricative Context.

Authors:

Santiago Fernández, Departamento de Física Aplicada. Universidad de Santiago de Compostela. (Spain)
Sergio Feijóo, Departamento de Física Aplicada. Universidad de Santiago de Compostela. (Spain)
Ramon Balsa, Departamento de Física Aplicada. Universidad de Santiago de Compostela. (Spain)
Nieves Barros, Departamento de Física Aplicada. Universidad de Santiago de Compostela. (Spain)

Page (NA) Paper number 452

Abstract:

The role of fricative context on vowel recognition in a series of FV syllables being part of natural Spanish words is investigated. Perceptual tests were carried out to assess the recognition of vowels in fricative context, in two conditions: 1) Isolated vowel; 2) Fricative noise + vowel. Analysis of results show that adding the fricative noise improves the recognition of the vowel, while the acoustic analysis reveal that the distribution of the vowels is affected by fricative context. A possible explanation for this improvement, i.e. the coarticulatory influence of the vowel on the fricative, was investigated. The results indicate that coarticulation cannot explain that improvement, since only 7.7% of the cases which improve when the fricative is added, show a clear influence of the vowel on the fricative.

SL980452.PDF (From Author) SL980452.PDF (Rasterized)

TOP

Voicing Affects Perceived Manner of Articulation.

Authors:

Santiago Fernández, Departamento de Física Aplicada. Universidad de Santiago de Compostela. (Spain)
Sergio Feijóo, Departamento de Física Aplicada. Universidad de Santiago de Compostela. (Spain)
Plinio Almeida, Instituto de Estudos Linguisticos. Universidade de Campinas. (Brazil)

Page (NA) Paper number 453

Abstract:

The perception of voiced fricatives by native speakers of a language which lacks those phonemes is studied in this paper. Brasilian portuguese and Galician languages were chosen because they are historically related. A forced choice test reveals that listeners correctly perceive the place of articulation of the voiced fricatives. In order to examine whether the perception of fricative manner can be overridden by the voicing characteristics an open test was carried out. Listeners perceive voiced fricatives as a voiced phoneme with different manner of articulation and similar place of articulation or as its voiceless counterpart, depending on whether vocal-fold vibration extends over the whole obstruent interval or not. Results are discussed in terms of both historical phonetic changes and second language acquisition.

SL980453.PDF (From Author) SL980453.PDF (Rasterized)

TOP

Enhancement Techniques to Improve the Intelligibility of Consonants in Noise : Speaker and Listener Effects

Authors:

Valerie Hazan, University College London (U.K.)
Andrew Simpson, University College London (U.K.)
Mark Huckvale, University College London (U.K.)

Page (NA) Paper number 487

Abstract:

The aim of our work is to increase the intelligibility of speech in noise by modifying regions of the signal that contain acoustic cues to consonant identity in order to make it more resistant to subsequent degradation. 36 vowel-consonant-vowel stimuli were recorded by four untrained speakers. The vowel onset/offset and consonant constriction/occlusion regions were selectively amplified and stimuli were presented to listeners in a background of noise (0 dB SNR). Enhanced tokens from all speakers were significantly more intelligible than natural tokens and the improvement was greater for the initially least intelligible speakers. Speech material for two speakers was then presented to Japanese and Spanish learners of English and controls. For all groups, the enhanced consonants were more intelligible. Error patterns were related to the 'distance' between the consonantal systems of the listeners' L1 and L2. These results demonstrate the robustness of our enhancement techniques across speaker and listener types.

SL980487.PDF (From Author) SL980487.PDF (Rasterized)

TOP

Boundaries of Perception of Long Tones in Taiwanese Speech

Authors:

Fran H.L. Jian, Dept. Linguistics,University of Reading (U.K.)

Page (NA) Paper number 145

Abstract:

In this work we set out to investigate the fundamental frequency boundaries of perception of the Taiwanese long tones. We are interested in how the variations in fundamental frequency affect the perception of linguistic tones in Taiwanese speech. Our investigation is adopted from similar studies of tones in Mandarin speech. As opposed to Mandarin tones that can be perceived with little difficulty the seven Taiwanese tones have a more subtle structure and are consequently harder to perceive successfully. The experimental results in this paper allow us to quantify these perceptual boundaries. The experiments consisted of a perception test involving over 150 Taiwanese subjects where the task involved identifying the tone of the words played back in a random sequence. The stimuli consisted of a set of tone pairs and a selection of intermediate tone words obtained by linearly interpolating between the words of the tone pairs.

SL980145.PDF (From Author) SL980145.PDF (Rasterized)

TOP

Effects of Phonetic Quality and Duration on Perceptual Acceptability of Temporal Changes in Speech

Authors:

Hiroaki Kato, ATR HIP (Japan)
Minoru Tsuzaki, ATR HIP (Japan)
Yoshinori Sagisaka, ATR ITL (Japan)

Page (NA) Paper number 411

Abstract:

To establish a perceptually valid rule for the durational control of synthetic speech, it is necessary to know the degree to which a given temporal error or distortion is acceptable to human listeners. Two perceptual experiments were conducted to estimate the acceptability of modifications in either vocalic or consonantal durations as a function of two attributes of the modified portions, i.e., the phonetic quality and the original (unmodified) duration. The results showed that the listeners' acceptable modification ranges were narrowest for vowels, and widest for voiceless fricatives and silent closures, with nasals in between. They were also narrower for those portions with shorter base durations. The effect of the original duration was larger for the vowel stimuli than for the voiceless fricative stimuli. The perceptual mechanism mediating these results is discussed with regard to the dependency of the listeners' temporal sensitivity on the stimulus loudness and base duration. [Re: http://www.hip.atr.co.jp/~kato/single_duration/]

SL980411.PDF (From Author) SL980411.PDF (Rasterized)

TOP

Dynamic vs. Static Spectral Detail in the Perception of Gated Stops

Authors:

Michael Kiefte, University of Alberta (Canada)
Terrance M. Nearey, University of Alberta (Canada)

Page (NA) Paper number 898

Abstract:

In order to assess the importance of dynamic spectral information within the first few milliseconds following oral release for the identification of prevocalic stop consonants, 23.75 ms gated CV syllables were presented to listeners for identification. In addition to these, subjects were presented with the same tokens reconstructed from their minimum phase decomposition such that they have the same long-term power spectrum as their original counterparts, but with differing internal dynamic spectral detail. Subjects' results from this experiment were then modelled with logistic regression analysis using mel cepstral coefficients with and without dynamic spectral information encoded in order to demonstrate the effect that reduced temporal information has in the context of automatic classification. Preliminary results from this experiment show that some dynamic spectral detail is used by listeners even for very short stimuli. We conclude that models of speech perception must take spectral variation over very short time frames into account.

SL980898.PDF (From Author) SL980898.PDF (Rasterized)

TOP

Phonological Units In Speech Segmentation And Phonological Awareness

Authors:

Takashi Otake, Dokkyo University (Japan)
Kiyoko Yoneyama, Ohio State University (USA)

Page (NA) Paper number 35

Abstract:

This paper explores the relationship between phonological units in speech segmentation and phonological awareness by investigating Japanese Brazilians living in Japan. The first experiment investigated the size of the phonological unit in speech segmentation using the Japanese materials and methodology in Otake et al. (1993). As for French subjects in the earlier study, the miss rates showed an effect of syllabic segmentation, suggesting that the Japanese Brazilians segmented Japanese into syllables. The second experiment investigated phonological units in phonological awareness, using a mid-chunk-unit search task in which subjects were asked to identify the middle unit within a word. 96% of the mid-chunk unit choices were syllable-based. The results of the two experiments suggest that Japanese Brazilians exploit syllables both as a speech segmentation unit and as a unit to represent within-word structure.

SL980035.PDF (Scanned)

TOP

How Far Do Speakers Back Up in Repairs? A Quantitatve Model

Authors:

Elizabeth Shriberg, SRI International (USA)
Andreas Stolcke, SRI International (USA)

Page (NA) Paper number 58

Abstract:

Speakers frequently retrace one or more words when continuing after a break in fluency. Syntactic principles constrain the points from which speakers retrace; however syntactic principles do not provide predictions about the relative usage of different allowable retrace points. Such predictions are useful for automatic processing of repairs in speech technology, particularly if they use information readily available to a speech recognizer. We propose a quantitative model that predicts the overall distribution of retrace lengths in a large corpus of spontaneous speech, based only on word position. The model has two components: (1) a constant, position-independent probability for extending a retrace by one more word; and (2) a position-dependent probability to "skip" to the beginning of the sentence. Results have implications for modeling repairs in speech applications and constrain explanatory models in psycholinguistics.

SL980058.PDF (From Author) SL980058.PDF (Rasterized)

TOP

Don't Blame It (All) On The Pause: Further ERP Evidence For A Prosody-Induced Garden-Path In Running Speech

Authors:

Karsten Steinhauer, Max-Planck-Institute of Cognitive Neuroscience, Inselstr. 22, D-04103, Leipzig (Germany)
Kai Alter, Max-Planck-Institute of Cognitive Neuroscience, Inselstr. 22, D-04103, Leipzig (Germany)
Angela D. Friederici, Max-Planck-Institute of Cognitive Neuroscience, Inselstr. 22, D-04103, Leipzig, (Germany)

Page (NA) Paper number 147

Abstract:

This paper investigates the prosodic relevance of a pause which, along with other prosodic parameters, served to indicate an Intonational Phrase (IPh) boundary. Event-related brain potentials (ERPs) were recorded while subjects listened to both intact and altered German Early and Late Closure (EC/LC) sentences. The EC sentences were prosodically highly accepted and well comprehended even when the original pause at the boundary position was removed. Furthermore, a reversed garden-path (initial EC preference in LC sentences) was successfully induced by a false IPh boundary irrespective of whether the pause was present or not. The ERP patterns disclosed the on-line processing of simple and garden-path sentences in more detail. The data clearly demonstrate that in the presence of other prosodic parameters pause insertion is a completely dispensable cue for boundary marking. The ERP technique proved to be superior to behavioral on-line measures as data collection does not interrupt speech presentation.

SL980147.PDF (From Author) SL980147.PDF (Scanned)

TOP

The Role of Stress for Lexical Selection in Dutch

Authors:

Jean Vroomen, University of Tilburg (The Netherlands)
Beatrice de Gelder, University of Tilburg (The Netherlands)

Page (NA) Paper number 348

Abstract:

In the present study, we examined whether stress constrains the number of activated lexical candidates. In a phoneme monitoring task, we used Dutch carrier words that start in their citation form with a reduced vowel (denoted as @), but which can also be produced with an unreduced vowel. For example, a word such as frequent (meaning frequent) can be pronounced as fr@QUENT or freQUENT. We examined whether mis-stressing these words had an effect on the activation of their lexical representation. Twenty subjects detected a target phoneme (e.g., the 't') in fr@QUENT, freQUENT, FR@quent, or FREquent; stress denoted in capitals. Results showed that target phonemes in words were reacted faster than in pseudowords, but neither stress, nor the nature of the vowel had an effect on the size of lexical effect. This confirms that stress is not part of the lexical input representation.

SL980348.PDF (From Author) SL980348.PDF (Rasterized)

TOP

The Perception of Stressed Syllables in Finnish

Authors:

Jyrki Tuomainen, Tilburg University (The Netherlands)
Jean Vroomen, Tilburg University (The Netherlands)
Beatrice de Gelder, Tilburg University (The Netherlands)

Page (NA) Paper number 760

Abstract:

The effect of word level prominence on detection speed of word boundaries in Finnish was investigated in two word spotting experiments. The results showed that the perceived stress was not a function of the fundamental frequency (F0) difference between the preceding syllable and the first syllable of the target word. Given the fast response times, the results suggest that subjects perceived in both experiments the first syllable of the target as stressed. This seems to indicate that when words are recognized in continuous speech the acoustic cues in the F0 contour signaling prominence may not be computed relative to the prominence of neighboring syllables. Instead, we hypothesize that subjects may be sensitive to a local pitch movement indicating change in the F0 slope.

SL980760.PDF (From Author) SL980760.PDF (Rasterized)

TOP

The Perception Of The Morae With Devocalized Vowels In Japanese Language.

Authors:

Kimiko Yamakawa, Prefectural University of Kumamoto (Japan)
Ryoji Baba, Prefectural University of Kumamoto (Japan)

Page (NA) Paper number 312

Abstract:

Usually the first vowel of the Japanese word 'susugi' (rinse) disappears, and so the pronunciation of 'susugi' is not [susugi] but [ssugi]. We made two psychophysical experiments. In the first one we shortened the part of [ss] of [ssugides] and [korewassugides] (This is rinse.) in 6 stages. [korewassugides] with shortened [ss] is easy to perceive as [korewasugides] (This is cedar) , but [ssugides] with shortened [ss] is not easy to perceive of [sugides] (Japanese cedar). In the second experiment we changed the pitches of [sugides] and researched how Japanese perceive these sounds.

SL980312.PDF (From Author) SL980312.PDF (Rasterized)

TOP