Role In Speech Perception Research Articles

Vowel production and perception in French blind and sighted adults

It is well known that visual cues play an important role in speech perception and production. Early in life, blind speakers, who do not have access to visual information related to lip and jaw movements, show a delay in the acquisition of phonological contrasts based on these visible articulatory features [A. E. Mills, Hearing by eye: The psychology of lip-reading, pp. 145–161 (1987)]. It has also been claimed that blind speakers have better auditory discrimination abilities than sighted speakers. The goal of this study is to describe the production-perception relationships involved in French vowels for blind and sighted speakers. Six blind adults and six sighted adults served as subjects. The auditory abilities of each subject were evaluated by auditory discrimination tests (AXB). At the production level, ten repetitions of the ten French oral vowels were recorded. Formant values and fundamental frequency values were extracted from the acoustic signal. Measures of contrasts (Euclidean distances) and dispersion (standard deviations) were computed and compared for each feature (height, place, roundedness) and group (blind, sighted). Regression analyses between discrimination scores and produced contrasts were carried out. Despite between-speaker variability, results show an effect of speakers group (blind versus sighted) on the produced contrast distances.

Differentiating speech and nonspeech sounds via amplitude envelope cues

Recent evidence from neuroscience and behavioral speech science suggests that the temporal modulation pattern of the speech signal plays a distinctive role in speech perception. As a first step in exploring the nature of the perceptually relevant information in the temporal pattern of speech, this experiment examined whether speech versus nonspeech environmental sounds could be differentiated on the basis of their amplitude envelopes. Conversational speech was recorded from native speakers of six different languages (French, German, Hebrew, Hindi, Japanese, and Russian) along with samples of their English. Nonspeech sounds included animal vocalizations, water sounds, and other environmental sounds (e.g., thunder). The stimulus set included 30 2-s speech segments and 30 2-s nonspeech events. Frequency information was removed from all stimuli using a technique described by Dorman et al. [J. Acoust. Soc. Am. 102 (1997)]. Nine normal-hearing adult listeners participated in the experiment. Subjects decided whether each sound was (originally) speech or nonspeech and rated their confidence (7-point Likert scale). Overall, subjects differentiated speech from nonspeech very accurately (84% correct). Only 12 stimuli were not correctly categorized at greater than chance levels. Acoustical analysis is underway to determine what parameters of the amplitude envelope differentiate speech from nonspeech sounds.

Changes over time in the psycho-electric parameters in children with cochlear implants: Cambios en el tiempo, de los parámetros psico-eléctricos en niños con implante coclear

Information regarding typical changes in electrical stimulation levels during different stages post-implantation in children is scarce. Clinically, this information can serve as a general guideline to the nature and frequency of these changes, and consequently help to determine programming schedules that would optimize performance with the implant. Furthermore, it may have implications for the design of cochlear implants, and for monitoring the condition of the implanted cochlea. The main goal of the present study was, therefore, to evaluate changes over time in psycho-electric parameters, i.e. threshold levels, comfortable levels, and dynamic range, in children with the Nucleus 22 cochlear implant system. The maps of 37 prelingual children were examined at five time intervals: 1, 6, 12, 18 and 24 months post-connection. Maps were analyzed according to three cochlear segments: basal, medial, and apical. The results indicated significant changes in the psycho-electric parameters during the first 2 years post-implantation. Specifically, threshold (T) levels increased significantly during the first year and stabilized during the second year, whereas comfort (C) levels and dynamic range (DR) increased significantly during the entire 2-year follow-up. In addition, psycho-electric parameters varied among the different cochlear segments, resulting in a narrower DR in the basal one. Regardless of the underlying cause for these changes, the results of the present study suggest that, in children, the stabilization of psycho-electric parameters is a prolonged, gradual process. Since psycho-electric parameters play an important role in speech perception, these findings emphasize the need for frequent programming in the first 2 years of implant use in children.Existe poca información sobre los cambios tipicos en los niveles de estimulación eléctrica durante diferentes etapas después del implante coclear en niños. Clínicamente, esta información puede servir como una guía general sobre la naturaleza y la frecuencia de estos cambios, y consecuentemente ayudar a determinar las citas para programación que pudieran optimizar el desempeño con el implante. Más aún, esto podría tener implicaciones en el diseño de los implantes cocleares y en el monitoreo de las condiciones de la cóclea implantada. El objetivo principal de este estudio fue, por tanto, evaluar los cambios en el tiempo de los parámetros psico-eléctricos, p.ej. niveles umbral, niveles de confortabilidad, y rango dinámico, en niños con implante coclear Nucleus 22. Se examinaron los mapas de 37 niños con sordera prelingüistica en 5 diferentes intervalos de tiempo: 1,6, 12, 18 y 24 meses después de la activación. Se analizaron los mapas de acuerdo a tres segmentos cocleares: basal, medial y apical. Los resultados indican que hay cambios significativos en los parámetros psico-eléctricos durante los dos primeros años después de la implantación. Específicamente, los niveles (T) aumentaron significativamente durante el primer año y se estabilizaron durante el segundo años, mientras que los niveles confortables (C) y el rango dinámico (DR) aumentaron significativamente en todo el tiempo de scguimiento de dos años. Adicionalmente, los parámetros psico-eléctricos variaron entre los diferentes segmentos cocleares, resultando un (DR) más estrecho en el segmento basal. Independientemente de las causas subyacentes que expliquen estos cambios, los resultados de este estudio sugieren que la estabilización de los parámetros psico-eléctricos en niños es un proceso gradual y a largo plazo. Tomando en cuenta que los parámetros psico-eléctricos juegan un papel importante en la percepción del lenguaje, estos hallazgos enfatizan la necesidad de una programación frecuente en los primeros dos años de uso del implante coclear en niños.

The influence of onset-density in speech production

Numerous studies have suggested that the initial phoneme (or syllable onset) plays an important role in speech perception and production. Recent findings have also demonstrated the importance that the number of similar sounding words plays in lexical retrieval during speech perception and production. The results from a picture-naming task demonstrate that the number of words that share the onset of the target word (onset-density) also influences speech production as well as speech perception. Specifically, pictures of words with sparse onsets were named more quickly than pictures of words with dense onsets. These results suggest that lexical and sublexical information interact during lexical retrieval in speech production. The implications for feedforward and feedback models of speech production will be discussed.

Imitation of phrases in conversational speech

Linguistically significant gestures of the vocal tract play an important role in speech perception and production. Imitation of speech across conversational partners is a natural consequence of the close connection between perception and production. Following research that finds imitation in single-word shadowing, this study paired unacquainted same-sex talkers in a conversational task designed to induce repetition of key phrases across partners. Immediately repeated items were assessed for imitative fidelity by asking a separate set of listeners to choose the more similar item in an AXB task that compared the sample item from one talker (X) with a read and a repeated item from the other talker (A/B). In general, listeners chose the repeated over the read item as more similar to X. Additionally, read items that were sampled immediately after the conversation hindered a listener’s ability to choose the repeated item more so than items read long before the conversation. Other effects related to the timing of items in the conversation and the role of the participants in the conversation were also observed.

Auditory–visual integration of talker gender in vowel perception

The experiments reported here used auditory–visual mismatches to compare three approaches to speaker normalization in speech perception: radical invariance, vocal tract normalization, and talker normalization. In contrast to the first two, the talker normalization theory assumes that listeners' subjective, abstract impressions of talkers play a role in speech perception. Experiment 1 found that the gender of a visually presented face affects the location of the phoneme boundary between [Ω] and [Λ] in the perceptual identification of a continuum of auditory–visual stimuli ranging from hood to hud. This effect was found for both “stereotypical” and “non-stereotypical” male and female voices. The experiment also found that voice stereotypicality had an effect on the phoneme boundary. The difference between male and female talkers was greater when the talkers were rated by listeners as “stereotypical”. Interestingly, for the two female talkers in this experiment, rated stereotypicality was correlated with voice breathiness rather than vowel fundamental frequency. Experiment 2 replicated and extended experiment 1 and tested whether the visual stimuli in experiment 1 were being perceptually integrated with the acoustic stimuli. In addition to the effects found in experiment 1, there was a boundary effect for the visually presented word: listeners responded hood more frequently when the acoustic stimulus was paired with a movie clip of a talker saying hood. Experiment 3 tested the abstractness of the talker information used in speech perception. Rather than seeing movie clips of male and female talkers, listeners were instructed to imagine a male or female talker while performing an audio-only identification task with a gender-ambiguous hood-hud continuum. The phoneme boundary differed as a function of the imagined gender of the talker. The results from these experiments suggest that listeners integrate abstract gender information with phonetic information in speech perception. This conclusion supports the talker normalization theory of perceptual speaker normalization.

Visual influences on the perception of alveolar/velar place discrimination

Using ten acoustic stimuli ranging along a /ga/–/da/ continuum, visual influences on the psychophysical boundary between the auditory percepts of /ga/ and /da/ were investigated. A significant shift in the perceptual boundary was found depending on whether syllables were dubbed onto a video of a speaker uttering /da/, or whether they were dubbed onto a video of the same speaker uttering /ga/. The tongue of the speaker was not visible in either video clip, so only subtle cues were available to distinguish place of articulation in the two conditions. The shift in auditory perception was in a direction consistent with the visual information, and indicates that even the subtle visual cues available to discriminate velar stop consonants from alveolar stop consonants can play a role in speech perception. The possibility that other (nonarticulatory) visual cues could be conditioned to play a similar role in speech perception is raised.

Auditory efferents involved in speech-in-noise intelligibility.

Following studies proposing that medial olivocochlear efferents might be involved in the processing of complex signals in noise, we tested the involvement of efferent feedback in speech-in-noise intelligibility. Two approaches were used: measures of speech-in-noise intelligibility in vestibular neurotomized patients with cut efferents and comparison with normal hearing subjects; and correlations between effectiveness of olivocochlear feedback, assessed by contralateral suppression of otoacoustic emissions and speech-in-noise intelligibility in normal subjects. Contralateral noise improved speech-in-noise intelligibility in normal ears. This improvement, which was almost absent in de-efferented ears of vestibular neurotomized patients, was correlated with the strength of the olivocochlear feedback. Together, these results suggest that olivocochlear efferents play an antimasking role in speech perception in noisy environments.

Critique: articulatory-acoustic relations and their role in speech perception.

These remarks are in response to "Role of articulation in speech perception: Clues from production"ony Björn Lindblom. It is suggested that the form in which the lexicon is stored includes both segments and distinctive features, and this representation is neutral with respect to articulatory and the acoustic domains. The process by which features are determined from the sound requires that patterns of acoustic properties be identified. In developing models of speech perception, knowledge of articulatory-acoustic relations can be a guide in defining these properties, but it is not necessary for the models to assign primary status to articulation.

語音認知における時間情報の役割

To determine the role of temporal information in speech perception, we created synthesized sound with temporal information alone and without such information as formant. To create such a sound, we plot the zero-crossing point in the original sound wave and place clicks on each point. So this sound consists of clicks. This stimulus sound contains only the timing information of stimulation and no significant information on the formant on the sonagram. Various types of this stimulus sound were created from the original sounds of Japanese vowels, VCV syllables, sentences with meaning and without meaning, song with background music and jazz instrumental music. We performed hearing experiments using this sound in twelve normal volunteers. The results showed that the perception of simple sounds (vowel, VCV syllable) were near the chance level. The perception of meaningless Japanese sentences were poor. Japanese sentences with meaning were well perceived. The perception of instrumental music was quite poor. These results suggested that the temporal information plays an important role in speech perception.

Limits on the limitations of context-conditioned effects in the perception of [b] and [w

We investigated the conditions under which the [b]-[w] contrast is processed in a context-dependent manner, specifically in relation to syllable duration. In an earlier paper, Miller and Liberman (1979) demonstrated that when listeners use transition duration to differentiate [b] from [w], they treat it in relation to the duration of the syllable: As syllables from a [ba]-[wa] series varying in transition duration become longer, so, too, does the transition duration at the [b]-[w] perceptual boundary. In a subsequent paper, Shinn, Blumstein, and Jongman (1985) questioned the generality of this finding by showing that the effect of syllable duration is eliminated for [ba]-[wa] stimuli that are less schematic than those used by Miller and Liberman. In the present investigation, we demonstrated that when these "more natural" stimuli are presented in a multi-talker babble noise instead of in quiet (as was done by Shinn et al.), the syllable-duration effect emerges. Our findings suggest that the syllable-duration effect in particular, and context effects in general, may play a more important role in speech perception than Shinn et al. suggested.

Hot topics in speech communication

Six hot topics will be described, two topics in each of the following three areas: speech perception, speech production, and signal processing by machines. (1) In speech perception, new studies on the representation of speech suggest that listeners' stored descriptions of phonetic units take the form of a prototype of the speech category. Adult data address the language specificity of prototypes and the effects of contextual factors on prototypes. Infant data address the developmental issue of whether phonetic prototypes are innate or whether they derive from linguistic experience. (2) Also in the area of speech perception, there are new data on the effects of talker variability. Adults, and even young infants, are capable of normalizing the speech produced by different talkers, but listeners at both ages are nonetheless adversely affected by talker variability, suggesting that attentional factors play a role in speech perception. (3) In the area of speech production, there is new emphasis on the study of articulatory timing in speech motor control. The pattern of phase relations between different articulators and its relation to linguistic structure, as well as the general coordination of speech movements, are being investigated. (4) Also in speech production, there are new advances in our understanding of laryngeal mechanisms. These advances stem from models of the glottal volume-velocity source, findings on aerodynamic phenomena at the glottis, and analyses of the mechanical properties of vocal-fold tissue. (5) In the area of signal processing, neural networks are being applied to difficult problems in speech recognition, such as the recognition of phonetic units. Computational models composed of large numbers of interconnected processing elements (nodes) have been successfully applied to the analysis of dynamic information distributed over several phonetic units. (6) Also in signal processing, the application of speech coding to hearing aids and auditory prostheses has resulted in improved speech reception. Cross-fertilization across areas in speech communication has furthered our research endeavors.

Toward an understanding of English syllabification

Using written and oral syllabification tasks and analyses of spelling errors, we studied how college students syllabify words with medial consonant clusters. Several principles of syllabification were evident. People almost always separated the phonemes of a cluster like /tl/, which is illegal at the beginnings and ends of words. A “short” vowel tended to attract consonants to its syllable, as did a stressed vowel. Finally, different syllabifications were produced for clusters like /dr/, which show the typical rise in sonority from the first consonant to the second, and /st/ clusters, which do not. Implications of the results for linguistic theories of syllabification and for studies of the syllable's role in speech perception and reading are discussed.

Auditory correlates to phonetic features: Results of a modeling study

A model of the peripheral auditory system was used to study two sets of speech tokens for which perceptual data exists. One of these sets exhibits a contrast between initial stop and glide (ba/wa), the other between intervocalic voiced and voiceless stop (rabid/rapid). Simple properties of the modeled peripheral auditory response can be identified, which are strongly linearly correlated with listeners' perception of the phonetic contrasts contained in these sets. Two fundamental characteristics of neural response play an important part in these auditory correlates: adaptation of firing rate due to sustained stimulation and spontaneous firing in the absence of stimulation. The results of this study support the theory that the transformations of the peripheral auditory system play an important role in speech perception.

Detection of French accent by native English speakers

This series of five experiments examines the ability of listeners to detect divergences from the phonetic norms of their native language. Native English-speaking listeners judged short English speech samples produced by native speakers of (American} English and native speakers of French. Computer editing techniques were used to isolate progressively shorter excerpts of speech. The listeners were able to identify the non-native speakers at above-chance levels no matter how short were the speech samples that were presented: phrases (e.g., “Two little dogs”), syllables (/tu/, /ti/), phonetic segments (/t/, /i/, /u/); and even the first 30 ms of the word “two.” Listeners identified the non-native (French) speakers equally well when presented with speech material edited from phrases read from a list and produced in a spontaneous narrative. Neither special phonetic training nor familiarity with French-accented English seemed to be necessary for listeners to detect accent. The results are interpreted as evidence that phonetic category prototypes play an important role in speech perception.

Invariant acoustic correlates of phonetic features and their potential role in speech perception

A version of the theory of acoustic invariance states that are invariant acoustic correlates of each phonetic feature, independent of the context in which the feature appears. Examples from a number of features are given to show that, if the appropriate acoustic attribute is defined for each feature, there is indeed a high degree of acoustic invariance for the feature. The implications of such a theory for models of language acquisition and speech perception are discussed. It is suggested that a realistic model of speech perception must have the capability to identify lexical items both through the extraction of a pattern of invariant acoustic attributes for individual features in the matrix that defines the item and, in situations where these attributes are obscured, through the use of secondary, context-dependent cues. [Supported in part by a grant from NINCDS.]

Influence of vocalic context on perception of the [zh]-[s] distinction.

When synthetic fricative noises from a [∫]-[s] continuum are followed by [a] or [u] (with appropriate formant transitions), listeners perceive more instances of [s] in the context of [u] than in the context of [a]. Presumably, this reflects a perceptual adjustment for the coarticulatory effect of rounded vowels on preceding fricatives. In Experiment 1, we found that varying the duration of the fricative noise leaves the perceptual context effect unchanged, whereas insertion of a silent interval following the noise reduces the effect substantially. Experiment 2 suggested that it is temporal separation rather than the perception of an intervening stop consonant that is responsible for this reduction, in agreement with recent, analogous observations on anticipatory coarticulation. In Experiment 3, we showed that the vowel context effect disappears when the periodic stimulus portion is synthesized so as to contain no formant transitions. To dissociate the contribution of formant transitions from contextual effects due to vowel quality per se, Experiment 4 employed synthetic fricative noises followed by periodic portions excerpted from naturally produced [∫a], [sa], [∫u], and [su]. The results showed strong and largely independent effects of formant transitions and vowel quality on fricative perception. In addition, we found a strong speaker (male vs. female) normalization effect. All three influences on fricative perception were reduced by temporal separation of noise and periodic stimulus portions. Although no single hypothesis can explain all of our results, they are generally supportive of the view that some knowledge of the dynamics of speech production has a role in speech perception.

Macrophage infiltration of breast tumours: a prospective study.

In 50 cases of infiltrating breast cancer investigated in a prospective study the number of macrophages within each tumour was assessed. The macrophages were identified by their cytoplasmic acid phosphatase activity. The number of lymphocytes and plasma cells within the tumours were graded by a scoring technique. Significantly fewer cases with metastases were found among those with high macrophage and plasma cell scores. There was no correlation between lymphoreticular infiltration and the degree of tumour differentiation, but in cases without metastases the lymphoreticular infiltration between tumour cells was nearly always only slight when the macrophage score was low.

On the front cavity resonance and its possible role in speech perception.

Spectrographic data are presented which suggest that it may be possible to estimate the frequency of the fundamental resonance of the cavity behind the mouth opening, the ’’front cavity resonance,’’ from information in the speech signal. It is shown that place of articulation information in the steady states, transitions, and bursts of F2 (or sometimes F3) can be reinterpreted to be information from the front cavity resonance. Furthermore, a number of synthesis results that have appeared anomalous when described in terms of numbered formants seem to find a coherent explanation in terms of the front cavity resonance. Implications for theories of speech perception include the possibility that an estimate of front cavity resonance frequency may serve for continuous articulatory reference. Subject Classification: 70.20, 70.30.

Perception of Brief Stimuli that Resemble Rapid Formant Transitions

One problem for the child who is acquiring language is the lack of acoustic similarity between the formant transitions of an initial stop, and the formant transitions of the postvocalic unreleased allo-phone of the same stop. Brief two-component stimuli were used to study perceptual similarity among members of a set of synthetic sounds which approximate the mirror-image formant transitions for /b d g/ in prevocalic and postvocalic positions. The two components of a stimulus were 40-msec tone bursts having linear frequency glissandos with frequency motions approximating the first and second or second and third formant patterns. Results indicate that there is no inherent similarity between mirror-image glissando patterns. Judgments of relative perceptual similarity were based on the glissando component corresponding in frequency to the second formant, even when this component was significantly reduced in amplitude relative to the other component. While the perceptual importance of the second formant in speech is well known, it is difficult to account for these results in terms of basic psychophysical data on frequency discrimination, loudness, or masking. These results, and the observation that some mirror-image responses for /g/-like transitions occur under certain special conditions argue for the presence of frequency change detectors that may play an important role in speech perception. [Research supported in part by NIH grants.]