Variability in talkers' fundamental frequencies shapes context effects in speech perception.

Speech perception is influenced by context. This influence can help to alleviate issues that arise from the extreme acoustic variability of speech. Two examples of contextual influences are talker normalization and spectral contrast effects (SCEs). Talker normalization occurs when listeners hear different talkers causing speech perception to be slower and less accurate. SCEs occur when spectral characteristics change from context sentences to target vowels and speech perception is biased by that change. It has been demonstrated that SCEs are restrained when contexts are spoken by different talkers (Assgari & Stilp, 2015). However, what about hearing different talkers restrains these effects was not entirely clear. In addition, while these are both considered contextual influences on speech perception, they have never been formally related to each other. The series of studies reported here served two purposes. First, these studies sought to establish why hearing different talkers restrained SCEs. Results indicate that variability in pitch (as measured by fundamental frequency), a primary acoustic cue to talker changes, restricts the influence of spectral changes on speech perception. Second, these studies attempted to relate talker normalization and SCEs by measuring them concurrently. Talker normalization (as measured by response times) and SCEs were evident in the same task suggesting that they act on speech perception at the same time. Further, these measures of talker normalization were shown to be influenced by f0 variability suggesting that SCEs and talker normalization are both related to f0 variability. However, no relationship between individual’s SCEs and response times was found. Possible reasons why f0 variability may restrain context effects are discussed.

All perception takes place in context. Recognition of a given speech sound is influenced by the acoustic properties of surrounding sounds. When the spectral composition of earlier (context) sounds (e.g., more energy at lower first formant [F1] frequencies) differs from that of a later (target) sound (e.g., vowel with intermediate F1), the auditory system magnifies this difference, biasing target categorization (e.g., towards higher-F1 /ɛ/). Historically, these studies used filters to force context sounds to possess desired spectral compositions. This approach is agnostic to the natural signal statistics of speech (inherent spectral compositions without any additional manipulations). The auditory system is thought to be attuned to such stimulus statistics, but this has gone untested. Here, vowel categorization was measured following unfiltered (already possessing the desired spectral composition) or filtered sentences (to match spectral characteristics of unfiltered sentences). Vowel categorization was biased in both cases, with larger biases as the spectral prominences in context sentences increased. This confirms sensitivity to natural signal statistics, extending spectral context effects in speech perception to more naturalistic listening conditions. Importantly, categorization biases were smaller and more variable following unfiltered sentences, raising important questions about how faithfully experiments using filtered contexts model everyday speech perception.

1 The Relevance of Psychophysics for Speech Perception.- The role of psychophysics in understanding speech perception.- Central and peripheral processes in the perception of speech and nonspeech sounds.- Psychophysics versus specialized processes in speech perception: an alternative perspective.- Speech perception and the role of long-term memory.- Levels of representation of phonemes and bandwidth of spectral-temporal integration.- General discussion of session 1.- 2 Separation of Acoustic Events.- The meaning of duplex perception: sounds as transparant objects.- Perceptual separation of speech from concurrent sounds.- Sound separation and auditory perceptual organization.- On the significance of spectral synchrony for signal detection.- Auditory enhancement in speech perception.- General discussion of session 2.- 3 Dynamic Aspects.- Trading relations, acoustic cue integration, and context effects in speech perception.- Perceptual integration of rise time and silence in affricate/fricative and pluck/bow continua.- Reversal of the rise-time cue in the affricate/fricative contrast: an experiment on the silence of sound.- Possible acoustic bases for the perception of voicing contrasts.- Is there a natural sensitivity at 20ms in relative tone-onset-time continua? A reanalysis of Hirsh's (1959) data.- Auditory constraints on speech perception.- Studies of possible psychoacoustic factors underlying speech perception.- Perception of tone, band, and formant sweeps.- Psychophysical representation of stop consonant and temporal masking in speech.- Effects of stimulus dynamics on frequency discrimination.- Extending the search for a psychophysical basis for dynamic phonetic patterns.- General discussion of session 3.- 4 Timbre (Peripheral Constraints and Central Processes in the Perception of Complex Signals)..- Psychophysics of audio signal processing and the role of pitch in speech..- Does the human auditory system include large scale spectral integration?.- Some aspects of the sound of speech sounds.- Involvement of the critical band in identification, perceived distance, and discrimination of vowels.- Profile analysis and speech perception.- General discussion of session 4.- 5 Physiological Correlates of Speech Perception.- Peripheral auditory processing of speech information: implications from a physiological study of intensity discrimination.- Organization of the cochlear nucleus for information processing.- Changes in the phonemic quality and neural representation of a vowel by alteration of the relative phase of harmonics near F1.- Phase vowels.- Nonlinear responses, in the auditory nerve to vowel-related complex stimuli.- Discussion of physiological correlates of speech perception.- General discussion of session 5.- 6 Primary Speech Percepts.- English and French speech processing: some psycholinguistic investigations.- Units of organization and analysis in the perception of speech.- Implications from infant speech studies on the unit of perception.- General discussion of session 6.- 7 Psychophysics and Speech Perception in the Hearing-Impaired.- Relationship between psychophysical abilities and speech perception for subjects with unilateral and bilateral cochlear hearing impairments.- Speech-reception threshold in a fluctuating background sound and its possible relation to temporal auditory resolution.- Differences in listening strategies between normal and hearing-impaired listeners.- Critical bands in the perception of speech signals by normal and sensorineural hearing loss listeners.- Phase and the hearing-impaired.

The identity of a speech sound can be affected by the spectrum of a preceding stimulus in a contrastive manner. Although such aftereffects are often reduced in people with hearing loss and cochlear implants (CIs), one recent study demonstrated larger spectral contrast effects in CI users than in normal-hearing (NH) listeners. The present study aimed to shed light on this puzzling finding. We hypothesized that poorer spectral resolution leads CI users to rely on different acoustic cues not only to identify speech sounds but also to adapt to the context. Thirteen postlingually deafened adult CI users and 33 NH participants (listening to either vocoded or unprocessed speech) participated in this study. Psychometric functions were estimated in a vowel categorization task along the /I/ to /ε/ (as in "bit" and "bet") continuum following a context sentence, the long-term average spectrum of which was manipulated at the level of either fine-grained local spectral cues or coarser global spectral cues. In NH listeners with unprocessed speech, the aftereffect was determined solely by the fine-grained local spectral cues, resulting in a surprising insensitivity to the larger, global spectral cues utilized by CI users. Restricting the spectral resolution available to NH listeners via vocoding resulted in patterns of responses more similar to those found in CI users. However, the size of the contrast aftereffect remained smaller in NH listeners than in CI users. Only the spectral contrasts used by listeners contributed to the spectral contrast effects in vowel identification. These results explain why CI users can experience larger-than-normal context effects under specific conditions. The results also suggest that adaptation to new spectral cues can be very rapid for vowel discrimination, but may follow a longer time course to influence spectral contrast effects.

The present study provides an elaboration of the "double-weak" theory of speech perception proposed by Nearey (1990, 1991). In this framework, the objects of speech perception (and production) are viewed as neither primarily auditory nor primarily gestural; rather, they are abstract, symbolic elements lawfully constrained to map onto relatively simple (but not entirely transparent) patterns in both domains. Speech cannot be understood unless both articulation and acoustics are considered: Many production strategies appear to be directed at achieving acoustically-oriented goals, yet most context effects in speech perception seem to be motivated by the consequences of gestural overlap. The double-weak framework suggests that speech perception and speech production are less-than-perfect inverses of each other. Despite long-term accommodation of each for the demands of the other, real-time production and perception may operate as autonomous subsystems. A family of perceptual models is discussed that provides varying degrees of approximation to "ideal solutions" (in the sense of minimizing error rate) to classifying production data exhibiting contextual interactions. Members of this family that provide substantial, yet incomplete, compensation for the consequences of gestural overlap appear to be adequate to account for the results of many speech perception experiments. Such partial perceptual compensation allows, in principle, for the kind of imperfect "error correction" discussed by Ohala (1981, 1990) in conjunction with hypo- and hypercorrection phenomena.

The phonetic distinction between the stop consonant /b/ and semivowel /w/ is cued primarily by duration of the initial formant transition. Syllable duration, however, is a context that influences human adults' and infants' perception of this distinction. For example, when overall syllable duration is long (296 ms), infants show poor discrimination of syllables with transition durations of 16 and 40 ms, but good discrimination of syllables with transition durations of 40 and 64 ms. However, when overall syllable duration is short (80 ms), the opposite is true [P. D. Eimas and J. L. Miller, Science 209, 1140–1141 (1980)]. The current study addresses whether rhesus and Japanese macaques' perception of the /b-w/ distinction is also influenced by syllable duration: Six CV syllables used in the Eimas and Miller (1980) study were presented to five macaques using a same-different discrimination paradigm. Results showed that macaques' pattern of discrimination was similar to the pattern shown by infants, thus supporting an auditory-based mechanism for the effect of syllable duration in phonetic perception. [Work supported by NSF and NIH.]

OF THE DISCLOSURE An interferometric sonar in non-linear acoustics for high-resolution observations of objects located in the vicinity of the sea floor comprises two end transducers for simultaneous transmission at two high frequencies F1 and F2 of the order of 200 kHz. Reception is performed by hydrophones placed between the sending transducers at the low frequency F2 - F1 of the order of 20 kHz. Angular channels are formed in two transmission sequences. Transmission takes place in phase at the frequency F2 - F1 in one sequence and in phase opposition in the other sequence in the case of the two sources generated by non-linear effects.

Effects of testosterone on speech production and perception: Linking hormone levels in males to vocal cues and female voice attractiveness ratings

This thesis provides a phonetic case study of the Swedish /i:/ variant known as Viby-i (sometimes also called Lidingö-i). This sound is characterised by an unusual 'thick', 'buzzing' vowel quality, but its articulation has long been disputed. Previous research suggests that this vowel may be subject to articulatory trade-off, whereby speakers can achieve the same sound using different articulatory strategies. There are also indications that Viby-i may be subject to sociolinguistic variation, as it appears to be spreading across Sweden, and it is frequently used as a prestige marker in urban dialects. This thesis addresses the issues of how speakers produce Viby-i, how its acoustic properties relate to its articulation, and how it is used across different social and linguistic contexts. The study presents data from 34 Swedish speakers from Gothenburg, Stockholm, and Uppsala, recorded with simultaneous audio, ultrasound tongue imaging, and lip video. The speakers' /i:/ productions are analysed acoustically with regard to their formant values, dynamic properties, and frication. Linguistic and social variation is also explored using a word list and a demographic questionnaire. The articulatory analysis establishes the tongue gestures and lip positions used to produce Viby-i, and investigates the link between acoustics and articulation using a set of normalised articulatory measurement points, which are compared to the first two formants. The acoustic analysis shows that Viby-i is characterised by a low acoustic F2, which is usually lower than [e:], and a relatively high F1, which is usually similar to [e:]. Linguistic context has a small but reliable effect on formant values, and also affects the fricated offglide of the vowel. All speakers in the sample are found to use Viby-i rather than standard [i:], but regional differences exist in both acoustics and articulation. The articulatory analysis reveals that Viby-i can be produced with a variety of tongue shapes, most of which involve a low, fronted tongue body, high tongue tip, and retraction of the postdorsal part of the tongue. Surprisingly, the low F2 is not produced by lip-rounding or overall tongue backing, but appears to be caused by a combination of tongue lowering and post-dorsal retraction. Young speakers retract more than old speakers, with no difference in acoustics. Frication during the vowel is associated with a raised, fronted tongue tip, or in velar contexts, by a raised tongue body. There are several implications of these findings for the wider literature. Firstly, Viby-i appears to be more widespread in Central Sweden than previously assumed, and it may already have replaced the standard variant [i:], although further research is needed to confirm this. Secondly, the 'mismatch' between acoustics and articulation demonstrates the benefit of including articulatory data in the analysis of vowel sounds. Finally, the study shows that traditional frameworks of vowel analysis may be too simplistic to adequately describe the articulation of complex vowel sounds, and that new methods may be required as articulatory data becomes increasingly available.

Listeners adjust their vowel perception to the characteristics of a particular speaker. Six experiments investigated whether speech‐specific signal characteristics influence the occurrence and amount of such normalization. Previous findings were replicated with first formant (F1) manipulations of naturally recorded speech; target sounds on a /pIt/ (low F1) to /pt/ (high F1) continuum were more often labeled as /pIt/ after a precursor sentence with a high F1, and more often labeled as /pt/ after one with a low F1 (Experiment 1). Normalization was also observed, though to a lesser extent, when these materials were spectrally rotated, and hence sounded unlike speech (Experiment 2). No normalization occurred when, in addition to spectral rotation, the silent intervals and pitch movement were removed and the syllables were temporally reversed (Experiment 3), despite spectral similarity of these precursors to those in Experiment 2. Reintroducing only pitch movement (Experiment 4), or silent intervals (Experiment 5), or spectrally rotating the stimuli back (Experiment 6) did not result in normalization, so none of these factors alone accounts for the effect’s disappearance in Experiment 3. These results show that normalization is not specific to speech but still depends on more than the overall spectral properties of the preceding acoustic context.

Functional imaging studies of speech perception have revealed extensive involvement of the dorsolateral temporal lobes in aspects of speech and voice processing. In the current study, fMRI was used as a functional imaging method to address the perception of speech in different masking conditions. Throughout the scanning experiment, subjects were directed to listen to a female talker with whom they had been familiarized previously. In addition to the acoustic noise generated by the scanner, different kinds of masking sound were also presented diotically, at SPL levels determined by pilot testing. The masking sounds were signal correlated noise, spectrally rotated speech, and a second female talker. The results show a network of activation in the bilateral temporal lobes, prefrontal cortex and parietal lobes that were commonly activated across all masking conditions. Controlling for the effects of energetic masking, informational masking gave rise exclusively to clusters of activity in bilateral STG, with peak level activations in left anterior-mid STG and posterior STS/STG and right primary auditory cortex. Post-test intelligibility measures were used to reveal greater activation in the left temporal lobe, which was associated with higher performance scores.

According to P. K. Kuhl (1991), a perceptual magnet effect occurs when discrimination accuracy is lower among better instances of a phonetic category than among poorer instances. Three experiments examined the perceptual magnet effect for the vowel /i/. In Experiment 1, participants rated some examples of /i/ as better instances of the category than others. In Experiment 2, no perceptual magnet effect was observed with materials based on Kuhl's tokens of /i/ or with items normed for each participant. In Experiment 3, participants labeled the vowels developed from Kuhl's test set. Many of the vowels in the nonprototype /i/ condition were not categorized as /i/s. This finding suggests that the comparisons obtained in Kuhl's original study spanned different phonetic categories.

Two synthetic speech syllables, differing in pitch, were presented over one loudspeaker situated at one of the following azimuths: 0 degrees (in from of the subject); 45 degrees, 90 degrees, and 135 degrees to the left or to the right; and 180 degrees. The subject's task was to report the high-pitch syllable. The pattern of performance can be described in terms of two component effects: a decreasing anterior-posterior gradient, and right-side advantage. Competition between messages from one direction only appears to be as effective in determining auditory spatial effects as competition between spatially distinct sources.

The “Perceptual Magnet Effect” (PME) is characterized by perceptual warping in favor of stronger assimilation near the prototypical sound against other nonprotypical variants. This ERP study adopted a modified oddball paradigm to examine the PME predictions. The participants were 18 normal adults. The stimuli were computer-synthesized /a/ sounds and their acoustic controls by replacing the formant frequencies with sinewaves. Prior to the ERP sessions, the prototype (P) and nonprototype (NP) were chosen based on behavioral identification and goodness rating. Then 4 variants were created with ±40 mels apart in reference to P or NP, respectively, in either the first formant (F1) or second formant (F2). Mismatch negativity (MMN) responses were derived with the P (or NP) as the standard and the 4 variants as the deviant. The same ERP procedure was used for the nonspeech stimuli. Behavioral results showed that individual subjects consistently rated vowel category goodness based on the F2/F1 ratio. MMN amplitudes in the P and NP conditions showed a pattern consistent with the PME model. This pattern was observed for both speech and nonspeech stimuli. These results provide neural evidence in support of the PME that may extend to general auditory processing of spectral patterns.

Adults and children use lexical knowledge to interpret phonemes in speech [W. F. Ganong, JEP: HPP 6, 110–125 (1980); Hurlburt and Goodman, in preparation; S. Nittrouer and A. Boothroyd, J. Acoust. Soc. Am. 87, 2705–2715 (1990)]. This study examines the nature of lexical representations necessary for top-down effects. In a baseline assessment, 5-year-olds and adults heard stimuli from seven-step voice onset time (VOT) continua. The continua endpoints formed nonwords (e.g., beb_→peb_). Subjects were then trained on one endpoint from each continuum in one of two learning conditions. Subjects in the exposure condition heard items in a carrier phase five times (e.g., ‘‘The sound you will hear is peb_.’’). Subjects in the concept condition heard five instances of each item used as a picture label, in definitional sentences, and in a sentence identifying its superordinate category. After training, subjects again heard the baseline VOT continua. Children’s and adults’ phoneme boundaries shifted such that a greater proportion of tokens were perceived as compatible with the trained end of the continuum regardless of learning condition. This suggests that minimal exposure to a sound pattern, rather than fully-specified lexical knowledge, can result in top-down effects in speech perception.