Search

Previous work has shown that human listeners are sensitive to level differences in high-frequency energy (HFE) in isolated vowel sounds produced by male singers. Results indicated that sensitivity to HFE level changes increased with overall HFE level, suggesting that listeners would be more “tuned” to HFE in vocal production exhibiting higher levels of HFE. It follows that sensitivity to HFE level changes should be higher (1) for female vocal production than for male vocal production and (2) for singing than for speech. To test this hypothesis, difference limens for HFE level changes in male and female speech and singing were obtained. Listeners showed significantly greater ability to detect level changes in singing vs speech but not in female vs male speech. Mean differences limen scores for speech and singing were about 5 dB in the 8-kHz octave (5.6–11.3 kHz) but 8–10 dB in the 16-kHz octave (11.3–22 kHz). These scores are lower (better) than those previously reported for isolated vowels and some musical instruments.

The ability to integrate and weight information across dimensions is central to perception and is particularly important for speech categorization. The present experiments investigate cue weighting by training participants to categorize sounds drawn from a two-dimensional acoustic space defined by the center frequency (CF) and modulation frequency (MF) of frequency-modulated sine waves. These dimensions were psychophysically matched to be equally discriminable and, in the first experiment, were equally informative for accurate categorization. Nevertheless, listeners' category responses reflected a bias for use of CF. This bias remained even when the informativeness of CF was decreased by shifting distributions to create more overlap in CF. A reversal of weighting (MF over CF) was obtained when distribution variance was increased for CF. These results demonstrate that even when equally informative and discriminable, acoustic cues are not necessarily equally weighted in categorization; listeners exhibit biases when integrating multiple acoustic dimensions. Moreover, changes in weighting strategies can be affected by changes in input distribution parameters. This methodology provides potential insights into acquisition of speech sound categories, particularly second language categories. One implication is that ineffective cue weighting strategies for phonetic categories may be alleviated by manipulating variance of uninformative dimensions in training stimuli.

Recently, Holt and Lotto [Hear. Res. 167, 156–169 (2002)] reported that preceding speech sounds can influence phonetic identification of a target syllable even when the context sounds are presented to the opposite ear or when there is a long intervening silence. These results led them to conclude that phonetic context effects are mostly due to nonperipheral auditory interactions. In the present paper, similar presentation manipulations were made with nonspeech context sounds. The results agree qualitatively with the results for speech contexts. Taken together, these findings suggest that the same nonperipheral mechanisms may be responsible for effects of both speech and nonspeech context on phonetic identification.

Kluender et al. [J. Acoust. Soc. Am. 97, 2552-2567 (1995)] reported that overall stimulus amplitude affects perception of the voicing contrast in syllable-initial stops as a function of frequency separation between the first formant (F1) and higher formants (F2, F3). These results were offered as support for a hypothesis that [- voice] could be signaled by a shift in the temporal pattern of neural firing from synchronizing to energy at frequencies of F2 and F3 to synchronizing to energy near F1. Several predictions from this "synchrony capture hypothesis" were tested in the current study. In all cases the hypothesis was not supported. Effect of stimulus amplitude (increased voiceless responses with higher amplitude) was maintained when there was no cutback in F1 or when F2 and F1 energy bands were presented dichotically. In further tests of the hypothesis, voice-voiceless series were created that maintained periodic energy throughout the syllable (with F1 cutback signaling voicing). Energy just below the frequency of F2 and energy above F1 were presented dichotically. Thus, at the periphery there was no competition between frequencies near F2 and lower frequencies. In contrast to predictions of the "synchrony capture hypothesis," overall amplitude still had an effect on voice-voiceless identifications.

Speech perception studies generally focus on the acoustic information present in the frequency regions below 6 kHz. Recent evidence suggests that there is perceptually relevant information in the higher frequencies, including information affecting speech intelligibility. This experiment examined whether listeners are able to accurately identify a subset of vowels and consonants in CV-context when only high-frequency (above 5 kHz) acoustic information is available (through high-pass filtering and masking of lower frequency energy). The findings reveal that listeners are capable of extracting information from these higher frequency regions to accurately identify certain consonants and vowels.

When presented a series of /dVd/ syllables varying perceptually from /ded/ to /d■d/, listeners identify the vowel sound as /■/ more often than when the vowels are presented alone as steady states. Conversely, /bVb/ series lead to more /e/ identifications. This pattern of responses appears to demonstrate a compensation for assimilatory effects of coarticulation or vowel ‘‘undershoot.’’ Experiments will be presented which test the specificity of this perceptual process. Two ten‐step (/bVb/ and /dVd/) series of 100‐ms /CVC/ syllables varying in F2 midpoint frequency (1260–1710 Hz) were synthesized. Nonspeech analogs were created by appending 70‐ms sine‐wave glides, which followed a similar trajectory to F2 in the speech stimuli, to 60‐ms steady‐state vowels varying in F2 frequency. A second set of nonspeech series was created by appending constant‐frequency sine‐wave tones to the vowel stimuli. The frequency of these tones were set at the F2 starting frequencies for the speech stimuli. These nonspeech contexts had effects on vowel identification that were analogous to those for the speech contexts. Lower frequency glides and tones (modeling /bVb/F2) led to more /e/ (higher F2) identifications with the complementary pattern for higher frequency glides and tones. [Work supported by NIDCD and NSF.]

Kluender, Lotto, and Jenison [J. Acoust. Soc. Am. 97, 2552–2567 (1995)] reported that overall stimulus amplitude can affect perception of the voicing contrast in syllable‐initial stops. Their results supported an hypothesis that a shift in the temporal pattern of neural firing from frequencies near F2 and F3 to F1 and f0 could signal voicelessness. The results of the current study undermine this ‘‘synchrony capture hypothesis.’’ The effect of amplitude (increased voiceless identifications with higher amplitude) maintains when there is no cutback in F1 during the quasiperiodic portion of the syllable and when stimuli are high‐pass filtered above the frequency of F1. In a further test of the hypothesis, two ten‐step series (/ba/–/pa/ and /ga/–/ka/) were created which maintained period voicing throughout the syllable (with F1 cutback signaling voicelessness). The energy just below the frequency of F2 and the energy above F1 were presented dichotically. Thus, at the periphery, there was no competition between frequencies near F2 and lower frequencies and, as a result, no chance for a change in neural temporal patterns. Subjects continued to label voicelessness as a function of overall amplitude. Alternative models of the encoding of voicelessness will be considered. [Work supported by NIDCD and NSF.]

In response to stop consonants with longer F1-cutback duration, the dominant synchronization of mid- and high-CF chinchilla auditory-nerve fibers changes from frequencies near F2 to frequencies near F1 at onset of voicing [D. G. Sinex and L. P. McDonald, J. Acoust. Soc. Am. 85, 1995-2004 (1989)]. If this change in neural synchronization is perceptually relevant for human listeners, then it may be predicted that changes in stimulus intensity and changes in the frequency difference between lower (F1) and higher (F2/F3) stimulus components should both affect perception of voicing. In a series of experiments, multiple continua of synthesized CVs varying in F1 cutback of the consonantal portion were played to listeners at levels ranging from 40 to 80 dB SPL. Across experiments, the frequency difference between F1 and F2 was manipulated by changing the onset frequency of F1 or F2. Subjects labeled more initial stops as voiceless as a function of increasing stimulus level and of decreasing frequency difference between F1 and F2. There was also an interaction between stimulus intensity and the frequency difference between F1 and F2 such that the effect of intensity was greater for smaller differences. These effects were reliable across a number of synthetic F1-cutback series, and the effect of intensity extended to a digitally edited series of hybrid CVs in which F1-cutback was varied by cross splicing naturally produced /da/ and /ta/. The effect of overall stimulus intensity was not affected by amplitude of prevocalic aspiration energy or by the presence or absence of release bursts. The results provide evidence for the perceptual significance of synchrony encoding of voicing for stop consonants.

The human singing and speech spectrum includes energy above 5 kHz. To begin an in-depth exploration of this high-frequency energy (HFE), a database of anechoic high-fidelity recordings of singers and talkers was created and analyzed. Third-octave band analysis from the long-term average spectra showed that production level (soft vs normal vs loud), production mode (singing vs speech), and phoneme (for voiceless fricatives) all significantly affected HFE characteristics. Specifically, increased production level caused an increase in absolute HFE level, but a decrease in relative HFE level. Singing exhibited higher levels of HFE than speech in the soft and normal conditions, but not in the loud condition. Third-octave band levels distinguished phoneme class of voiceless fricatives. Female HFE levels were significantly greater than male levels only above 11 kHz. This information is pertinent to various areas of acoustics, including vocal tract modeling, voice synthesis, augmentative hearing technology (hearing aids and cochlear implants), and training/therapy for singing and speech.

The human voice spectrum above 5 kHz receives little attention. However, there are reasons to believe that this high-frequency energy (HFE) may play a role in perceived quality of voice in singing and speech. To fulfill this role, differences in HFE must first be detectable. To determine human ability to detect differences in HFE, the levels of the 8- and 16-kHz center-frequency octave bands were individually attenuated in sustained vowel sounds produced by singers and presented to listeners. Relatively small changes in HFE were in fact detectable, suggesting that this frequency range potentially contributes to the perception of especially the singing voice. Detection ability was greater in the 8-kHz octave than in the 16-kHz octave and varied with band energy level.

In a previous paper, Lotto et al. [J. Acoust. Soc. Am. 103, 3648–3655 (1998)] claimed that experiments investigating a perceptual magnet effect (PME) failed to demonstrate a novel effect of goodness judgments on speech sound discrimination due to a confound in the procedure. New analyses of these data and a review of the recent literature fail once again to support the existence of PME.

Williams [(1986). "Role of dynamic information in the perception of coarticulated vowels," Ph.D. thesis, University of Connecticut, Standford, CT] demonstrated that nonspeech contexts had no influence on pitch judgments of nonspeech targets, whereas context effects were obtained when instructed to perceive the sounds as speech. On the other hand, Holt et al. [(2000). "Neighboring spectral content influences vowel identification," J. Acoust. Soc. Am. 108, 710-722] showed that nonspeech contexts were sufficient to elicit context effects in speech targets. The current study was to test a hypothesis that could explain the varying effectiveness of nonspeech contexts: Context effects are obtained only when there are well-established perceptual categories for the target stimuli. Experiment 1 examined context effects in speech and nonspeech signals using four series of stimuli: steady-state vowels that perceptually spanned from /inverted ohm/-/I/ in isolation and in the context of /w/ (with no steady-state portion) and two nonspeech sine-wave series that mimicked the acoustics of the speech series. In agreement with previous work context effects were obtained for speech contexts and targets but not for nonspeech analogs. Experiment 2 tested predictions of the hypothesis by testing for nonspeech context effects after the listeners had been trained to categorize the sounds. Following training, context-dependent categorization was obtained for nonspeech stimuli in the training group. These results are presented within a general perceptual-cognitive framework for speech perception research.

Behavioral experiments with infants, adults, and nonhuman animals converge with neurophysiological findings to suggest that there is a discontinuity in auditory processing of stimulus components differing in onset time by about 20 ms. This discontinuity has been implicated as a basis for boundaries between speech categories distinguished by voice onset time (VOT). Here, it is investigated how this discontinuity interacts with the learning of novel perceptual categories. Adult listeners were trained to categorize nonspeech stimuli that mimicked certain temporal properties of VOT stimuli. One group of listeners learned categories with a boundary coincident with the perceptual discontinuity. Another group learned categories defined such that the perceptual discontinuity fell within a category. Listeners in the latter group required significantly more experience to reach criterion categorization performance. Evidence of interactions between the perceptual discontinuity and the learned categories extended to generalization tests as well. It has been hypothesized that languages make use of perceptual discontinuities to promote distinctiveness among sounds within a language inventory. The present data suggest that discontinuities interact with category learning. As such, ‘‘learnability’’ may play a predictive role in selection of language sound inventories.

Hearing aids and other communication devices (e.g., mobile phones) have made some recent efforts to extend their bandwidths to represent higher frequencies. The impact of this expansion on speech perception is not well characterized. To assess human sensitivity to speech high-frequency energy (HFE, defined here as energy in the 8- and 16-kHz octave bands), difference limens for HFE level changes in male and female speech and singing were obtained. Listeners showed significantly greater ability to detect level changes in singing vs. speech, but not in female vs. male speech. Mean differences limen scores for speech and singing were about 5 dB in the 8-kHz octave (5.6–11.3 kHz) but 8–10 dB in the 16-kHz octave (11.3–22 kHz). These scores are lower (better) than scores previously reported for isolated vowels and some musical instruments, and similar to scores previously reported for white noise.

Vocal tremor involves atypical modulation of the fundamental frequency (F0) and intensity of the voice. Previous research on vocal tremor has focused on measuring the modulation rate and extent of the F0 and intensity without characterizing other modulations present in the acoustic signal (i.e., modulation of the harmonics). Characteristics of the voice source and vocal tract filter are known to affect the amplitude of the harmonics and could potentially be manipulated to reduce the perception of vocal tremor. The purpose of this study was to determine the adjustments that could be made to the voice source or vocal tract filter to alter the acoustic output and reduce the perception of modulation. This research was carried out using a computational model of speech production that allows for precise control and modulation of the glottal and vocal tract configurations. Results revealed that listeners perceived a higher magnitude of voice modulation when simulated samples had a higher mean F0, greater degree ...

Listeners were trained to distinguish two novel classes of speech sounds differing in both Voice Onset Time (VOT) and fundamental frequency at vowel onset (f0). One group was shown Korean orthography during the training period (“symbols” group) and the other English orthography (“letters” group). During a subsequent test phase, listeners classified sounds with mismatched VOT and f0. The two groups relied on different cues to categorize the contrast: those exposed to symbols used f0, while those exposed to letters used VOT. A second experiment employed the same paradigm, but the two dimensions defining the contrast were closure duration (instead of f0) and VOT. In this more difficult experiment, successful listeners in the “letters” group again classified the hybrid stimuli based on VOT, while the single listener in the “symbols” group who passed the learning criterion used closure duration. In both experiments, subjects showed different categorization patterns based on orthography used in the presentation...

We present data quantifying the use of multiple acoustic dimensions in the perception and production of word-initial stops in Spanish and English in bilinguals and monolinguals, with a focus on the stability of individuals’ cue weighting strategies across languages and modalities. Early Spanish-English bilinguals who use both languages on a regular basis, along with control groups of monolingual Spanish and English speakers, categorized stop-initial CV syllables covarying in Voice Onset Time (VOT), fundamental frequency at vowel onset (f0), first formant onset frequency (F1), and stop closure duration. Individual perceptual cue weights were calculated via logistic regression, and these weights were compared with the same speakers’ use of the corresponding acoustic dimensions in their productions of the stop contrast. Preliminary results suggest that bilingual listeners’ perceptual cue weighting strategies correspond across Spanish and English; that is, individuals who weight a given cue more heavily than average in English also weight that cue heavily in Spanish. However, in line with previous work, correspondences between a given individual's use of the acoustic dimensions across perception and production are more elusive.

For stimuli modeling stop consonants varying in the acoustic correlates of voice onset time (VOT), human listeners are more likely to perceive stimuli with lower f0's as voiced consonants--a pattern of perception that follows regularities in English speech production. The present study examines the basis of this observation. One hypothesis is that lower f0's enhance perception of voiced stops by virtue of perceptual interactions that arise from the operating characteristics of the auditory system. A second hypothesis is that this perceptual pattern develops as a result of experience with f0-voicing covariation. In a test of these hypotheses, Japanese quail learned to respond to stimuli drawn from a series varying in VOT through training with one of three patterns of f0-voicing covariation. Voicing and f0 varied in the natural pattern (shorter VOT, lower f0), in an inverse pattern (shorter VOT, higher f0), or in a random pattern (no f0-voicing covariation). Birds trained with stimuli that had no f0-voicing covariation exhibited no effect of f0 on response to novel stimuli varying in VOT. For the other groups, birds' responses followed the experienced pattern of covariation. These results suggest f0 does not exert an obligatory influence on categorization of consonants as [VOICE] and emphasize the learnability of covariation among acoustic characteristics of speech.

Four experiments explored the relative contributions of spectral content and phonetic labeling in effects of context on vowel perception. Two 10-step series of CVC syllables ([bVb] and [dVd]) varying acoustically in F2 midpoint frequency and varying perceptually in vowel height from [delta] to [epsilon] were synthesized. In a forced-choice identification task, listeners more often labeled vowels as [delta] in [dVd] context than in [bVb] context. To examine whether spectral content predicts this effect, nonspeech-speech hybrid series were created by appending 70-ms sine-wave glides following the trajectory of CVC F2's to 60-ms members of a steady-state vowel series varying in F2 frequency. In addition, a second hybrid series was created by appending constant-frequency sine-wave tones equivalent in frequency to CVC F2 onset/offset frequencies. Vowels flanked by frequency-modulated glides or steady-state tones modeling [dVd] were more often labeled as [delta] than were the same vowels surrounded by nonspeech modeling [bVb]. These results suggest that spectral content is important in understanding vowel context effects. A final experiment tested whether spectral content can modulate vowel perception when phonetic labeling remains intact. Voiceless consonants, with lower-amplitude more-diffuse spectra, were found to exert less of an influence on vowel perception than do their voiced counterparts. The data are discussed in terms of a general perceptual account of context effects in speech perception.

In recent years there has been a great deal of interest in demonstrations of the so-called “Perceptual-Magnet Effect” (PME). In these studies, AX-discrimination tasks purportedly reveal that discriminability of speech sounds from a single category varies with judged phonetic “goodness” of the sounds. However, one possible confound is that category membership is determined by identification of sounds in isolation, whereas, discrimination tasks include pairs of stimuli. In the first experiment of the current study, identifications and goodness judgments were obtained for vowels (/i/–/e/) presented in pairs. A substantial shift in phonetic identity was evidenced with changes in the context vowel. In a second experiment, listeners participated in an AX-discrimination task with the vowel pairs from the first experiment. Using the contextual identification functions from the first experiment, predictions of discriminability were calculated using the classic tenets of Categorical Perception. Obtained discriminability functions were well accounted for by predictions from identification. There was no additional unexplained variance that required the proposal of “perceptual magnets.” These results suggest that PME may be nothing more than further demonstration that general discriminability is greater for cross-category stimulus pairs than for within-category pairs.

When members of a series of synthesized stop consonants varying in third-formant (F3) characteristics and varying perceptually from /da/ to /ga/ are preceded by /al/, human listeners report hearing more /ga/ syllables than when the members of the series are preceded by /ar/. It has been suggested that this shift in identification is the result of specialized processes that compensate for acoustic consequences of coarticulation. To test the species-specificity of this perceptual phenomenon, data were collected from nonhuman animals in a syllable "labeling" task. Four Japanese quail (Coturnix coturnix japonica) were trained to peck a key differentially to identify clear /da/ and /ga/ exemplars. After training, ambiguous members of a /da/–/ga/ series were presented in the context of /al/ and /ar/ syllables. Pecking performance demonstrated a shift which coincided with data from humans. These results suggest that processes underlying "perceptual compensation for coarticulation" are species-general. In addition, the pattern of response behavior expressed is rather common across perceptual systems.

From previous studies it is known that normal‐hearing (NH) listeners have the ability to compensate for the acoustic variability present in speech through context‐dependent perception of speech sounds. One question of practical and theoretical interest is whether listeners with cochlear implants (CI) also show context‐dependent speech perception. Because of the lack of spectral resolution in the input, the representation of speech for CI listeners may differ from NH listeners, which may interfere with perceptual compensation. In a test of this prediction, adult postlingually deafened CI listeners did not demonstrate the contrastive context effects elicited from NH listeners for either /da/–/ga/ targets and /al/–/ar/ contexts or V targets and /b—b/–/d—d/ contexts. In contrast, as predicted by the good temporal resolution of the CI signal, CI listeners showed normal effects of vowel length on preceding glide‐stop categorization. CI simulations with NH listeners were also performed for some of these context ...

Research in the field of speech perception has traditionally focused on the acoustic information or cues present in the frequency region below 5 kHz, thus ignoring high frequency energy (HFE). A recent series of studies, however, demonstrated that listeners could determine the mode of production (speech or singing) and further, could identify what was being spoken or sung from the HFE alone and with a speech-shaped masking noise in the lower frequencies. This begs the question as to what types of information listeners are extracting to guide their perception. The current study examined the ability of listeners to transcribe short semantically-unpredictable but syntactically-well-formed spoken phrases that were high-pass filtered at 5.6 kHz. Of particular interest was the ability of some listeners to correctly determine the placement of word boundaries even without the availability of low-frequency information. These findings add to a growing literature on the linguistically relevant information present in higher frequency regions. Results will be framed within current theories of acoustic signatures to lexical segmentation. [Work supported by NIH-NIDCD.]

This work investigates the production and perception of the English stop voicing contrast on multiple acoustic dimensions by native speakers of Seoul Korean. Subjects completed a production task as well as a forced-choice identification task on stimuli varying on three acoustic dimensions (aspiration duration, pitch, and closure duration) in both English and Korean. On average, native Korean listeners relied more on pitch than on aspiration duration to categorize the English stop voicing contrast (which native English listeners distinguish primarily by aspiration). However, individual categorization patterns differed considerably, with some listeners using only pitch, some using only aspiration, and most using both. In contrast, in production, aspiration duration was a better predictor of voicing category than pitch, although pitch was still a stronger predictor than previously found in native English productions. The heavier reliance on pitch by Korean listeners may be attributable to the greater importance of pitch in their native stop contrast; however, there does not appear to be a consistent, straightforward mapping of English sounds onto the Korean categories. Results will be discussed in the terms of the perception-production interface on the level of individual acoustic cues, as well as the influence of native language cue weights on non-native sound contrasts.

Listeners were randomly presented narrow-band filtered noise bursts that varied in filter center frequency from two overlapping, Gaussian-like distributions. Participants mapped these distributions of sounds onto creatures in a video game where they received visual and auditory feedback about their accuracy. Categorization boundaries for each participant were estimated using logistic regression and compared with the optimal boundary from an ideal observer model. The participants appeared to be able to establish near optimal boundaries rapidly and had a remarkable ability to shift these boundaries when the underlying distributions changed - even when these changes were not explicitly signaled. These results suggest that listeners maintain a rather detailed representation of distributional information that is continuously updated during the task. This interpretation is in line with the assumptions underlying many current models of perceptual (statistical) learning in speech perception. However, it is possib...

Vocal tremor can have a variety of physiological sources. For example, tremors can result from involuntary oscillation of respiratory muscles (respiratory tremor), or of the muscles responsible for vocal fold adduction (adductory tremor) or lengthening (f0 tremor). While the sources of vocal tremor are distinct, they are notoriously difficult to categorize both perceptually and acoustically. In order to develop acoustic measures that can potentially distinguish sources of tremor, speech samples were synthesized using a kinematic model of the vocal folds attached to a model of the vocal tract and trachea [Titze, JASA, 75, 570-580; Story, 2005, JASA, 117, 3231-3254]. Tremors were created by modulating parameters of the vocal fold model corresponding to the three types mentioned above. The acoustic measures were related to temporal regularities in the amplitude envelope computed across the entire signal and select frequency bands. These measures could reliably categorize the samples by tremor source (as dete...

For more than fifty years, it has been known that listeners’ perception of a target speech sound may shift as a result of a change in speaker of a preceding carrier phrase. The predominant theories explaining this phenomenon suggest that the listener must be extracting information specific to the speaker - either generating speaker-specific acoustic-phonemic mapping or an estimate of the dimensions of the speaker’s vocal tract. To the extent that these processes derive representations that are specific to a speaker, one would not expect shifts in target perception when the apparent speaker remains unchanged. This prediction was explicitly tested by appending targets to different carrier phrases produced by the same speaker. Speaker characteristics were controlled by using an articulatory synthesizer that allows one to specify the source and dimensions of the vocal tract. Significant shifts in target perception were readily obtained, even when the speaker remained the same. These results suggest that these...

There has been a lot of recent interest in distinguishing languages based on their rhythmic differences. A common successful approach involves measures of relative durations and duration variability of vowels and consonants in utterances. Recent studies have shown that more general measures of temporal regularities in the amplitude envelope in separate frequency bands (the Envelope Modulation Spectrum) can reliably discriminate between English and Spanish [Carbonell et al. J. Acoust. Soc. Am. 129, 2680.]. In the current study, these temporal structure measures were supplemented with measures of the mean and variance of spectral energy in octave bands as well as with traditional linguistic measures. Using stepwise discriminant analysis and a set of productions from Japanese, Korean and Mandarin speakers, this suite of both acoustic and linguistic measures were tested together and pitted against each other to determine the most efficient discriminators of language. The results provide insight into what the ...

Within the framework of signal-detection theory, listeners' ability to process stimulus information is partly limited by their internal noise. Knowing the internal noise of listeners will allow the researcher to predict their performance in various detection or discrimination tasks. For the purpose of estimating internal noise, in this study we measure the ability of listeners to discriminate the intensities of pure tones (Exp. 1) and two-tone complexes (Exp. 2), and to discriminate the spectral shapes of two-tone complexes (Exp. 3). The results from five listeners were fitted with an internal-noise model. The analysis showed that a major portion of the internal noise is correlated across frequency channels. This outcome is consistent with the conclusions of previous studies that estimated internal noise using different methods. Under the dominance of channel-correlated internal noise over channel-specific internal noise, listeners' ability to integrate information over frequency should be reduced, and their performance in detection and discrimination should be less affected by uncertainty for the signal parameters.

High-density electroencephalography (EEG) was used to evaluate cortical activity patterns during the auditory processing of speech presented at various levels of degradation in a sentence verification task. 25 healthy participants listened to true-false sentences produced with one of 3 channel levels (1, 6, 16), ranging from unintelligible, moderately intelligible, and highly intelligible. Behavioral data were collected via button press (reaction time and accuracy) for each sentence. The analysis of cortical activation patterns includes 1) the identification of independent components that account for variations of activity associated with degradation levels and 2) the examination of associated event-related potentials. Statistical analyses will be performed on individuals' cortical activity, in tandem with behavioral data. Preliminary results reveal differences in the timing and magnitude of frontal and temporal lobe activation that coincide with task difficulty. Results of the present investigation will ...

When F1‐onset frequency is lower, longer F1 cutback (VOT) is required for human listeners to perceive synthesized stop consonants as voiceless. K. R. Kluender [J. Acoust. Soc. Am. 90, 83–96 (1991)] found comparable effects of F1‐onset frequency on the ‘‘labeling’’ of stop consonants by Japanese quail (coturnix coturnix japonica) trained to distinguish stop consonants varying in F1 cutback. In that study, CVs were synthesized with naturallike rising F1 transitions, and endpoint training stimuli differed in the onset frequency of F1 because a longer cutback resulted in a higher F1 onset. In order to assess whether earlier results were due to auditory predispositions or due to animals having learned the natural covariance between F1 cutback and F1‐onset frequency, the present experiment was conducted with synthetic continua having either a relatively low (375 Hz) or high (750 Hz) constant‐frequency F1. Six birds were trained to respond differentially to endpoint stimuli from three series of synthesized /CV/s varying in duration of F1 cutback. Second and third formant transitions were appropriate for labial, alveolar, or velar stops. Despite the fact that there was no opportunity for animal subjects to use experienced covariation of F1‐onset frequency and F1 cutback, quail typically exhibited shorter labeling boundaries (more voiceless stops) for intermediate stimuli of the continua when F1 frequency was higher. Responses by human subjects listening to the same stimuli were also collected. Results lend support to the earlier conclusion that part or all of the effect of F1 onset frequency on perception of voicing may be adequately explained by general auditory processes. Implications concerning VOT variation for stop consonants at different places of articulation or preceding different vowels are discussed.

Although high-frequency energy (HFE) is typically ignored in the speech perception and music literature, previous work has demonstrated that listeners are able to extract information from the structure of energy above 5 kHz. When exposed to HFE stimuli, listeners could accurately determine the mode of production (speech versus singing) and the gender of the speaker/singer. However, HFE in these previous experiments was presented in isolation at levels higher than what is typical of normal speech and singing. In the series of experiments presented here, the stimuli were presented at more appropriate levels. Additionally, speech-shaped masking noise from 0 to 5.6kHz (set at the mean levels for an average speaker/singer) was used to determine if HFE is usable by listeners even in the presence of masking by the more intense energy below 5 kHz in typical speech/voice listening conditions. Results indicate that even when the HFE stimuli are presented at typical levels and in the presence of a masker, listeners ...

Recent studies have shown that rhythm metrics calculated from amplitude envelopes extracted from octave bands across the spectrum (the envelope modulation spectrum or EMS) can reliably discriminate between spoken Spanish and English even when produced by the same speakers [Carbonell et al., J. Acoust. Soc. Am. 129, 2680]. Additionally, bilingual speakers could be discriminated fairly seven females and five males well on EMS variables even across sentences spoken in the different languages. In the current study, EMS, a general acoustic measure with no reference to phoneme/linguistic entities, was supplemented with measures of the mean and variance of spectral energy in each octave band as well as the mean and variance of fundamental frequency. Using stepwise discriminant analysis and the set of bilingual productions of Spanish and English, it was determined that language discrimination was excellent using both EMS and spectral measures, whereas spectral and f0 measures were most informative for speaker discrimination. The results demonstrate that this suite of easily-calculated acoustic measures provides abundant information about differences between languages as well as inherent differences in speakers that are independent of language spoken. [Work supported by NIH-NIDCD.]

This experiment examined the ability of human listeners to categorize sounds as a function of changing training distribution characteristics. Participants were presented non-speech sounds randomly sampled from two overlapping, Gaussian-like distributions. The sounds consisted of narrow-band noise bursts varying in center frequency. Participants mapped the distributions of sounds onto creatures in a video game receiving visual and auditory feedback about accuracy. The distributions were constant for half of the participants. For the other half, an unsignaled switch to distributions with a new optimal boundary occurred in the middle of the session. By examining obtained category boundaries one can distinguish between 3 possible responses to the switch: (1) adaptive switching resulting in the new optimal boundary being learned; (2) persistence of first learned boundary; and (3)averaging input across the entire session, resulting in an obtained boundary midway between the first and second optimal boundaries. The results demonstrate that most listeners could adaptively switch between distribution conditions with no external signal informing them of such a switch. The results indicate that most listeners are updating possible underlying distributions of input in a remarkably adaptive manner. [Work supported by NIH.]

In this series of experiments, we examined the role of specific vocal tract and phoneme contextual information on vowel categorization. An articulatory synthesizer was used to create vowels produced by specific vocal tract lengths (VT) of different sizes [Story, J. Acoust. Soc. Am. 117, 3231-3254 (2005)]. Contexts of repeated vowels (/a/ or /i/) produced by long and short VTs preceded a target series that changed perceptually from /bit/-/bet/. If listeners are extracting information about the VT itself, similar context effects should be elicited by either vowels produced by the same VT. If listeners are comparing the ambiguous target to the placement of a known vowel, one might predict similar results for the /aa/ condition or the /ii/. However, a complete cross-over interaction was obtained such that the effect of the context could not be predicted by the VT size or the vowel identity, but instead on the acoustic characteristics of F1. Follow-up experiments used contexts consisting of cross-cutting vowel...

This study explores whether information obtained from the envelope modulation spectrum (EMS) can predict the type of intelligibility decrement experienced by listeners of dysarthric speech. Previously, Liss et al. [J. Speech Lang. Hear. Res. 53, 1246–1255 (2010)] demonstrated that spectralinformation obtained via EMS can classify individuals with different patterns of degraded speech into their appropriate clusters. Variables noted as significant in discriminating dysarthria types, such as relative energy above and below the region of 4 Hz, are of particular interest. While the analyzes showed variations of energy in certain octaves of the acoustic signal were critical for this analysis, it is unclear which perceptual components map onto these measures. Using performance measures that capture listeners’ strategies to understand speech at the segmental, suprasegmental, and phonetic levels (intelligibility, lexical boundary errors, and word substitutions), regressions will be performed to determine which information contained within the EMS metrics can account for the variability in listener performance at various levels. Ultimately, this will allow us to look at trends in the EMS output, identify the source of intelligibility decrement, and predict listener performance on outcome measures. Results will bear on models of connected speech perception as well as clinical practice.

Previous work has provided evidence that the presence/absence of carrier phrase effects on target speech sound categorization can be well predicted by comparing the long‐term average spectrum (LTAS) of the carrier phrase with the spectrum of the target. In this series of studies, we move beyond asking whether there is a significant effect or not to examining the relative size of these effects. Target series of minimal‐pair words varying in vowel were preceded by carrier phrases synthesized to simulate a talker with a large or short vocal tract. Shifts in vowel categorization percentage as a function of change in talker served as the variable of interest. Results indicate that manipulations of the perception of the carrier as speech (by presenting forward or backwards), the LTAS of the carrier (by gating the duration of the carrier presented), and the local information in the carrier (by preserving or not the part of the carrier immediately preceding the target) affected the size of the resulting shifts. Of particular interest, the LTAS has an affect above and beyond the spectrum nearest to the target and the carriers perceived as speech resulted in larger shifts even when LTAS was held constant. [Work Supported by NIH‐NIDCD].

Several studies have manipulated the accent of a spoken passage, observing shifts in subsequent speech perception. Such “accent normalization” may be related to remapping phonetic space or, alternatively, to tuning auditory representations via context. In the current experiments, artificial “accents” with precisely specified acoustic vowel distributions isolated the influence of these two potential mechanisms. Participants were exposed to a spoken passage drawn from Dr. Seuss’ The Foot Book with /i/ shifted to /I/ (higher‐F1), or /U/ shifted to /u/ (lower‐F1), or with both manipulations and then categorized /i/‐/I/ or /U/‐/u/ series. The third context had shifted phonetic categories, but little change in the context’s long term average spectrum (LTAS), known to influence subsequent speech categorization. As expected, higher‐F1 contexts led to more low‐F1 (/i/, /u/) target responses whereas lower‐F1 contexts predicted more high‐F1 targets (/I/, /U/) responses. However, we did not observe shifted vowel cate...

Distorted vowel production is a hallmark characteristic of dysarthric speech and often is characterized by spectral and temporal degradation, flattening of spectral change formants, and vowel space distortions that may differentially affect high versus low or front versus back contrasts. Because acoustic information critical to accurate speech perception is contained in vowels, it is important to ask how such degradations influence the resulting percept. To date, attempts to characterize this relationship have met with mixed results. In the present investigation, a variety vowel space metrics were derived from vowels produced by 48 speakers with dysarthria to assess their ability to predict perceptual outcome measures (e.g., intelligibility and vowel identification accuracy). Novel metrics capturing distinctiveness of spectral neighbors (i.e., dispersion of vowels) were included in this analysis, as recent evidence suggests that vowel distinctiveness, not vowel space area, better predicts vowel intelligib...

Traditionally, acoustical energy above 5000 Hz in voice and speech has received little attention from the scientific community. Some recent studies have explored the role of this high‐frequency energy (HFE) in qualitative percepts of speech and music. To determine what other perceptual information is extractable from this frequency range, listeners were presented with tokens containing only HFE from samples of singing and speech (i.e., singing and speech high‐pass filtered at approximately 5.7 kHz). Results showed that listeners could successfully discriminate gender and mode of production (speech versus singing) given only acoustical energy above 5 kHz. Additionally, listeners successfully identified the song being sung and the words being spoken. The implication is that HFE potentially plays a more significant role in perception of singing and speech than that suggested by the current understanding. [Work supported by the NIH‐NIDCD.]

There has been a great deal of work on classifying spoken languages according to their perceived or acoustically‐measured rhythmic structures. The current study examined the speech of 12 Spanish‐English bilinguals producing sentences in both languages using rhythmic measures based on the amplitude envelopes extracted from different frequency regions—the envelope modulation spectrum (EMS). Using discriminant factor analysis, EMS variables demonstrated a moderate ability to classify the language being spoken suggesting that rhythmic differences between languages survive even when speaker is controlled. More interesting is the fact that EMS variables could reliably classify which speaker produced each sentence even across languages. This result suggests that there are stable rhythmic structures in an individual talker’s speech that are apparent above and beyond the structural constraints of the language spoken. The EMS appears capable of describing systematic characteristics of both the talker and the language spoken. [Work supported by NIH‐NIDCD].

A substantial body of literature has focused on context effects in speech perception in which manipulation of the phonemic or spectral content of preceding sounds (e.g., /al/ versus /ar/) result in a shift in the perceptual categorization of a target syllable (e.g., /da/ versus /ga/). In a previous study utilizing the frequency‐following response (FFR) to measure neural correlates of these context effects [R. Aravamudhan, J. Acoust. Soc. Am. 126, 2204], it was noted that the representation of target formant trajectories were much weaker when the stimulus was presented in isolation versus following some type of context. To examine this effect explicitly, a series of syllables varying from /da/ to /ga/ was presented to listeners either in isolation or following the syllables /a/, /al/, or /ar/ (with a 50‐ms silent gap between context and target). FFR measures were obtained from EEG recordings while participants listened passively. The resulting narrow‐band spectrograms over the grand averages demonstrated t...

Traditionally, context‐sensitive speech perception has been demonstrated by eliciting shifts in target sound categorization through manipulation of the phonemic/spectral content of surrounding context. For example, changing the third formant frequency of a preceding context (from /al/ to /ar/) can result in significant shifts in target categorization (from /ga/ to /da/). However, it is probable that the most salient difference in context is between the presence or absence of any other sound. The question becomes whether this large change in context has substantial effects on target categorization as well. In the current study, participants were asked to categorize members of a series of syllables varying from /ga/ to /da/ presented in isolation or following /al/, /ar/, or /a/. The typical shifts in categorization were obtained for /al/ vsersus /ar/ contexts, but the shift in response between isolated presentation and any of the audible context conditions was much larger (with more /da/ responses in isolation). A similar pattern of results was obtained when the audible context was broadband noise. Additionally, slopes of categorization functions were less steep for isolated syllables than when preceded by speech. These results suggest that isolated syllables may be neurally encoded differently than the same syllables in context.

It has been clear for more than 50 years that listeners sometimes shift their categorization of target vowels as a function of the information in preceding carrier phrases. These results have been interpreted as evidence for a talker normalization process that allows listeners to tune their speech perception to the peculiarities of individual talkers. However, it is still not clear what information is being extracted about the talker from the context phrases. In the current set of studies, carrier phrases (“He had a rabbit and a”) spoken by different “talkers” were produced by manipulating the length and oral/pharyngeal cavity ratios of an articulatory synthesizer. [Story, J. Acoust. Soc. Am. 117, 3231–3254 (2005).] These phrases preceded a series of target stimuli varying from “bit” to “bet” produced by the standard vocal tract for the synthesizer. Shifts in the perceived target vowel were obtained for some of these vocal tract manipulations but not others. The presence/absence of effects was most readily predictable from the long‐term spectral average of the carrier phrase. These results support the idea that listeners are extracting general acoustic information from the context and not specific information about the talker’s vocal tract. [Work supported by NIH‐NIDCD.]

Subtypes of dysarthria are often described in terms of particular disruptions in speech rhythms. In an attempt to quantify these differences, the envelope modulation spectrum (EMS) was computed on sentences read by speakers with hypokinetic, hyperkinetic, ataxic, and mixed flaccid‐spastic dysarthria, and a group of normal speakers (43 speakers total). The EMS consists of the spectra of the slow‐rate (up to 10 Hz) amplitude modulations of the full signal and within seven octave bands ranging in center frequency from 125–8000 Hz. From these spectra, six variables were calculated for each band relating to the peak frequency and amplitude and the relative energy above, below, and in the region of 4 Hz. According to stepwise discriminant analysis, six of these EMS variables could correctly discriminate speaker group membership with an accuracy of 84% (better performance was obtained for distinguishing individual groups). Further, 85% of the variance in the intelligibility of the dysarthric speakers (derived from a separate set of sentences) was predictable from a subset of six of these variables. These results suggest that there are reliable perturbations of speech temporal patterns associated with dysarthria subtypes and that these perturbations play an important role in intelligibility. [Work supported by NIH‐NIDCD.]

Ladefoged and Broadbent [J. Acoust. Soc. Am. 29, 98–104 (1957)] demonstrated that listeners will shift their categorization of a target vowel as a function of acoustic characteristics of a preceding carrier phrase. These results have been interpreted as an example of perceptual normalization for variability resulting from differences in talker anatomy. The present study examined whether listeners would normalize for acoustic variability resulting from differences in speaking style within a single talker. Two vowel series were synthesized that varied between central and peripheral vowels (the vowels in “beat”‐“bit” and “bod”‐“bud”). Each member of the series was appended to one of four carrier phrases that were spoken in either a “clear” or “reduced” speech style. Participants categorized vowels in these eight contexts. A reliable shift in categorization as a function of speaking style was obtained for three of four phrase sets. This demonstrates that phrase context effects can be obtained with a single ta...

Speech production disorders arise from a wide variety of anatomical defects, neurological injury, and disease. Depending on the etiology, the speech manifestations may be quite regularized—in the same way that a foreign accent presents consistently—or highly variable with unpredictable speech interruption and degradation. Indeed, even within an individual speaker, one will find both relatively consistent and variable aspects of speech abnormality. A growing body of literature shows that listeners enjoy intelligibility gains with just brief auditory exposure to disordered speech that is accompanied by lexical feedback, and with repeated unsupervised exposures. These benefits have been shown to be both talker‐specific and applicable to groups of people who presumably share relevant acoustic‐perceptual speech features. While many of these studies have been designed with an eye toward clinical and communicative implications, there is much to be learned about the mechanisms underlying these perceptual improvements with naturally impoverished speech. The purpose of this presentation is to describe the use of clinical populations as a testing ground for hypotheses of listener learning and adaptation. Speech production disorders particularly lend themselves to the study of episodic memory, rhythmic expectancy, and talker‐specific versus talker‐general effects. [Work supported by NIDCD, NIH.]

The envelope modulation spectrum (EMS) is a spectral analysis of the low‐rate (0–10 Hz) amplitude modulations of the envelope for the entire speech signal and within specific frequency bands. Using a set of predictor variables computed from these spectra, LeGendre et al. [J. Acoust. Soc. Am. 125, 2530–2531 (2009)] were able to classify sentences by dysarthric subtypes with a remarkable degree of accuracy, presumably because the EMS encodes rhythmic perturbations characteristic of each subtype. In the current study, we tested whether the EMS encodes perturbation differences within a subtype by attempting to classify sentences according to individual talkers. Stepwise discriminant analyses were performed on five sentences produced by 12 speakers with ataxic dysarthria. The 60 sentences were accurately assigned to talker 93% of the time. Thus, individual patterns of motor disruption are readily observed in the speech amplitude envelope. A similar analysis was performed with a set of normal control speakers and, again, classification of sentence by talker was remarkably high. These results suggest that talkers have idiosyncratic rhythmic patterns in their speech production that are independent of linguistic content (which was controlled in this study). The EMS may be a useful measure for talker recognition. [Work supported by NIH/NIDCD].