Perception Of Fricatives Research Articles

Acoustic correlates of fricative perception

A number of spectral attributes have been proposed as correlates of fricative place of articulation. However, these attributes have difficulty distinguishing fricatives in noncanonical speech, e.g., from children. Because child speech is intelligible, it is possible that some of these characteristics are not used in speech perception. This study collected a corpus of /s/ and /S/-containing words from children between 3 and 5. Adults categorized fricatives extracted from this corpus. A series of regressions are used to compare the ability of spectral attributes to predict (1) human categorization of /s/ and /S/ productions, and (2) the intended fricative production. The analysis reveals that mistakes in perception are driven by overuse of a single spectral attribute, spectral slope. This result suggests caution in equating methods of computational separation with methods of human perception, and provides a simple method for investigating relationships between the two.

Perception of clear fricatives by normal-hearing and simulated hearing-impaired listeners

Speakers may adapt the phonetic details of their productions when they anticipate perceptual difficulty or comprehension failure on the part of a listener. Previous research suggests that a speaking style known as clear speech is more intelligible overall than casual, conversational speech for a variety of listener populations. However, it is unknown whether clear speech improves the intelligibility of fricative consonants specifically, or how its effects on fricative perception might differ depending on listener population. The primary goal of this study was to determine whether clear speech enhances fricative intelligibility for normal-hearing listeners and listeners with simulated impairment. Two experiments measured babble signal-to-noise ratio thresholds for fricative minimal pair distinctions for 14 normal-hearing listeners and 14 listeners with simulated sloping, recruiting impairment. Results indicated that clear speech helped both groups overall. However, for impaired listeners, reliable clear speech intelligibility advantages were not found for non-sibilant pairs. Correlation analyses comparing acoustic and perceptual data indicated that a shift of energy concentration toward higher frequency regions and greater source strength contributed to the clear speech effect for normal-hearing listeners. Correlations between acoustic and perceptual data were less consistent for listeners with simulated impairment, and suggested that lower-frequency information may play a role.

Perception of clear English fricatives

Two experiments investigated whether and how clear speech production enhances intelligibility of English fricatives for normal-hearing listeners and listeners with simulated hearing impairment. Fricative intelligibility in clear and conversational speech was assessed using a database of 8800 VCV ([a]-fricative-[a]) stimuli produced by 20 speakers. Babble thresholds were measured for minimal pair distinctions for 14 normal-hearing listeners and 14 listeners with simulated sloping, recruiting impairment. Clear speech benefited both groups overall; however, for impaired listeners, the clear speech effect held only for sibilant pairs. In a previous acoustic study involving the same stimuli [Maniwa, Jongman, and Wade, J. Acoust. Soc. Am. 119, 3301 (2006)], 14 parameters were measured: spectral peak location, the first four spectral moments, F2 onset transitions, spectral slopes below and above typical peak locations, pitch of adjacent vowels, overall rms amplitude, relative amplitude (frication amplitude relative to the vowel in specific frequency bands), harmonic-to-noise ratio, energy below 500 Hz, and fricative duration. Correlation analyses comparing acoustic and perceptual data indicated that a shift of energy concentration toward higher frequency regions and greater source strength contributed to the clear speech effect for normal-hearing listeners, while listeners with simulated loss seemed to benefit mostly from cues involving lower frequency regions.

Perception of Fricatives by Dutch and English Speakers

Two speech perception experiments explored physical and linguistic bases of similarity. Because phonological features are grounded in similarity, these experiments were thus an exploration of bases of distinctive features. Experiment 1 found that Dutch listeners rated (s) and () as more similar to each other than did American English listeners. We attributed this to a pattern of alternation in Dutch phonology involving these sounds. Experiment 1 also found that Dutch listeners rated () and (s), and () and () as more similar to each other than did American English listeners. We attributed this to lack of () in Dutch inventory of native sounds. Experiment 2 found that Dutch and American English listeners did not signficantly differ from each other in a speeded discrimination task with same stimuli as experiment 1. Reaction times in experiment 2 were highly correlated with rating data of experiment 1 (r = -.86 to -.96) indicating that general pattern of response in experiment 1 was based on auditory similarity, with language-specific effects superimposed on general pattern. We conclude that similarity, a base of phonological features, is comprised of three components: (1) auditory similarity, (2) inventory, and (3) language- specific patterns of alternation. This is a paper about similarity. By phonetic we mean a psychological phenomenon - something like the subjective experience of linguistic sound - which is no doubt related to physically measureable acoustic, visual, or motor aspects of speech sounds; but which also may be derived from listener's experience of sounds of language. We will argue on basis of data from two cross-linguistic speech perception experiments, that similarity is a product both of physical aspects of sounds and of linguistic patterning of sounds. This research is relevant to distinctive feature theory because phonological features are grounded in similarity. Phonetic similarity is relevant in poetic use of language to determine near-rhymes (Steriade, 2003), sound change is influenced by perceptibility of sounds involved in change (Hume & Johnson, 2001), and synchronic phonological patterns

The Interaction of Contextual Vowels and Formant Transitions in Japanese Fricative Perception (Summaries of Talks at the Regular Research Study Meeting No.315)

The Interaction of Contextual Vowels and Formant Transitions in Japanese Fricative Perception (Summaries of Talks at the Regular Research Study Meeting No.315)

Role of linguistic experience on audio-visual perception of English fricatives in quiet and noise backgrounds

Previous research shows that for native perceivers, visual information enhances speech perception, especially when auditory distinctiveness decreases. This study examines how linguistic experience affects audio-visual (AV) perception of non-native (L2) speech. Native Canadian English perceivers and Mandarin perceivers with two levels of English exposure (early and late arrival in Canada) were presented with English fricative-initial syllables in a quiet and a caf-noise background in four ways: audio-only (A), visual-only (V), congruent AV, and incongruent AV. Identification results show that for all groups, performance was better in the congruent AV than A or V condition, and better in quiet than in caf-noise background. However, whereas Mandarin early arrivals approximate the native English patterns, the late arrivals showed poorer identification, more reliance on visual information, and greater audio-visual integration with the incongruent AV materials. These findings indicate that although non-natives were more attentive to visual information, they failed to use the linguistically significant L2 visual cues, suggesting language-specific AV processing. Nonetheless, these cues were adopted by the early arrivals who had more L2 exposure. Moreover, similarities across groups indicate possible perceptual universals involved. Together they point to an integrated network in speech processing across modalities and linguistic backgrounds. [Work supported by SSHRC.]

The Dynamic Nature of Speech Perception

The speech perception system must be flexible in responding to the variability in speech sounds caused by differences among speakers and by language change over the lifespan of the listener. Indeed, listeners use lexical knowledge phonetic to retune perception of novel speech (Norris, McQueen, & Cutler, 2003). In categorization that study, Dutch listeners made lexical decisions to spoken stimuli, including words with an ambiguous fricative (between [f] and [s]), in either [f]- or speech [s]-biased lexical contexts. In a subsequent categorization test, the former perception group of listeners identified more sounds on an [epsilonf]-[epsilons] continuum as [f] than the latter group. In the present experiment, listeners received the same exposure and test stimuli, but did not make lexical decisions to the exposure items. Instead, they counted them. Categorization results were indistinguishable from those obtained earlier. These adjustments in fricative perception therefore do not depend on explicit judgments during exposure. This learning effect thus reflects automatic retuning of the interpretation of acoustic-phonetic information.

The gradient influence of talker sex and perceived sexual orientation on fricative perception

Listeners are more likely to hear a synthetic fricative ambiguous between /s/ and /∫/ as /∫/ if it is appended to a woman’s voice than a man’s voice [Strand, J. Lang. Soc. Psych. 18, 86–99 (1999)], suggesting that speech perception is sensitive to social-indexical information. This study examined the influence of two variables on listeners’ fricative perception: (a) talker sex, and (b) talkers’ perceived sexual orientation (PSO, i.e., the probability that a talker is identified as gay, lesbian, or bisexual (GLB) based on speech alone). Stimuli were created by pairing a synthetic nine-step /s/-to-/∫/ series with tokens of /æk/ and /Ip/ taken from natural productions of shack and ship by 44 talkers (22 women, 22 men), for a total of 88 different continua. Forty listeners rated the 44 talkers’ PSO. A different group of 10 listeners participated in a series of two-alternative sack-shack and sip-ship identification experiments. As expected, listeners identified more /∫/ tokens for women’s voices than for men’s. GLB-sounding women elicited significantly fewer /∫/ percepts than heterosexual-sounding women. No consistent influence of PSO on fricative identification was noted for men’s voices. Regression analyses showed strong relationships between fricative identification and ratings of PSO for women talkers only.

Aural Habilitation Update: The Role of Speech Production Skills of Infants and Children With Hearing Loss

Aural Habilitation Update: The Role of Speech Production Skills of Infants and Children With Hearing Loss

The Perception of Fricative Peaks and Noise Bands

Recent work on the identification and perception of fricatives has focussed on the use by listeners of spectral moments derived from the whole spectrum and there appears to be no work in the literature on the use of prominent spectral peaks. In this study, we map the response of a single listener to narrow bands of noise that "mimic" the spectral peaks of English voiceless fricatives. The stimuli are based on the critical-band rate scale (Zwicker and Fastl, 1990) which divides the audible frequency range up to 15,500 Hz into 24 abutting critical bands. The results suggest that listeners have knowledge that enables them to connect a narrow-band spectral peak with a particular fricative consonant. We demonstrate that such knowledge, particularly in conjunction with a normalization metric that takes account of an individual speaker's vocal tract characteristics (F0 of the vowel following the fricative), could be used to good effect, particularly in noisy conditions which impair the use of the whole spectrum.

Contributions of semantic and facial information to perception of nonsibilant fricatives.

Most studies have been unable to identify reliable acoustic cues for the recognition of the English nonsibilant fricatives [see text]. The present study was designed to test the extent to which the perception of these fricatives by normal-hearing adults is based on other sources of information, namely, linguistic context and visual information. In Experiment 1, target words beginning with /f/, /theta/, /s/, or [see text] were preceded by either a semantically congruous or incongruous precursor sentence. Results showed an effect of linguistic context on the perception of the distinction between /f/ and /theta/ and on the acoustically more robust distinction between /s/ and [see text]. In Experiment 2, participants identified syllables consisting of the fricatives [see text] paired with the vowels /i, a, u/. Three conditions were contrasted: Stimuli were presented with (a) both auditory and visual information, (b) auditory information alone, or (c) visual information alone. When errors in terms of voicing were ignored in all 3 conditions, results indicated that perception of these fricatives is as good with visual information alone as with both auditory and visual information combined, and better than for auditory information alone. These findings suggest that accurate perception of nonsibilant fricatives derives from a combination of acoustic, linguistic, and visual information.

Learning to perceive speech: how fricative perception changes, and how it stays the same.

A part of becoming a mature perceiver involves learning what signal properties provide relevant information about objects and events in the environment. Regarding speech perception, evidence supports the position that allocation of attention to various signal properties changes as children gain experience with their native language, and so learn what information is relevant to recognizing phonetic structure in that language. However, one weakness in that work has been that data have largely come from experiments that all use similarly designed stimuli and show similar age-related differences in labeling. In this study, two perception experiments were conducted that used stimuli designed differently from past experiments, with different predictions. In experiment 1, adults and children (4, 6, and 8 years of age) labeled stimuli with natural /f/ and /[see text]/ noises and synthetic vocalic portions that had initial formant transitions varying in appropriateness for /f/ or /[see text]/. The prediction was that similar labeling patterns would be found for all listeners. In experiment 2, adults and children labeled stimuli with initial /s/-like and /[see text]/-like noises and synthetic vocalic portions that had initial formant transitions varying in appropriateness for /s/ or /[see text]/. The prediction was that, as found before, children would weight formant transitions more and fricative noises less than adults, but that this age-related difference would elicit different patterns of labeling from those found previously. Results largely matched predictions, and so further evidence was garnered for the position that children learn which properties of the speech signal provide relevant information about phonetic structure in their native language.

Japanese adult’s and children’s production and perception of English fricatives: A longitudinal study

This study examined the production and perception of English fricatives by native Japanese (NJ) adults and children (16 per group, mean age=40 and 10 years), and age-matched native English (NE) adults and children (16 per group). The subjects were tested two times (T1, T2) 1 year apart. (At T1, the NJ subjects’ mean length of residence in the U.S. was 0.5 year.) A picture-naming task was used to elicit the production of English words beginning with /s/ and /θ/, and intelligibility scores were obtained for both. The intelligibility scores of the NJ children but not adults improved significantly from T1 to T2. The NJ children obtained significantly lower scores than the NJ adults did at T1, but at T2 the adult–child difference was nonsignificant. The perception of /s/ and /θ/ was tested by a categorial discrimination task. Although the NJ adult’s and children’s scores improved from T1 to T2, the T1–T2 differences were nonsignificant. Thus, the results showed that the NJ children’s production scores improved significantly from T1 to T2, while there was no significant change for the perception scores on the discrimination of /s/–/θ/. The relationship between production and perception in L2 speech learning will be discussed. [Work supported by NIH.]

Perception and production of word-initial long fricatives in Polish

The paper provides an analysis of the perception and production of the Polish distinction between long fricative /v:/ and short fricative /v/ in pre-vocalic position. The task involved 27 native speakers of Polish repeating 50 sentences, including the target sentences. The target sentences were minimal pairs with the meaning difference based solely on fricative length. The results showed that the 190-ms mean duration of the long fricatives was well over twice as long as the 71-ms duration of the short fricatives. When duration of the fricative fell below 147 ms, the listener’s ability to discriminate between a short and long fricative was significantly reduced. Fricative length correlated with the duration of the following vowel: 78% of the time, the vowel was longer in the syllable with the long fricative. Indeterminacy about the identification of long fricatives is particularly well marked in these data: 42% of the responses to prompts with long fricatives included a distinctive pitch pattern, a so-called flat hat: a rising F0 in the first syllable, pitch height maintenance, and a drop after the second syllable. In Polish, this pattern indicates that the speaker is paying particular attention to this part of the utterance [Demenko (1999)].

Effect of stimulus bandwidth on the perception of /s/ in normal- and hearing-impaired children and adults.

Recent studies with adults have suggested that amplification at 4 kHz and above fails to improve speech recognition and may even degrade performance when high-frequency thresholds exceed 50-60 dB HL. This study examined the extent to which high frequencies can provide useful information for fricative perception for normal-hearing and hearing-impaired children and adults. Eighty subjects (20 per group) participated. Nonsense syllables containing the phonemes /s/, /f/, and /O/, produced by a male, female, and child talker, were low-pass filtered at 2, 3, 4, 5, 6, and 9 kHz. Frequency shaping was provided for the hearing-impaired subjects only. Results revealed significant differences in recognition between the four groups of subjects. Specifically, both groups of children performed more poorly than their adult counterparts at similar bandwidths. Likewise, both hearing-impaired groups performed more poorly than their normal-hearing counterparts. In addition, significant talker effects for /s/ were observed. For the male talker, optimum performance was reached at a bandwidth of approximately 4-5 kHz, whereas optimum performance for the female and child talkers did not occur until a bandwidth of 9 kHz.

Perception of voiceless fricatives by normal-hearing and hearing-impaired children and adults.

This study examined the perceptual-weighting strategies and performance-audibility functions of 11 moderately hearing-impaired (HI) children, 11 age-matched normal-hearing (NH) children, 11 moderately HI adults, and 11 NH adults. The purpose was to (a) determine the perceptual-weighting strategies of HI children relative to the other groups and (b) determine the audibility required by each group to achieve a criterion level of performance. Stimuli were 4 nonsense syllables (see text). The vowel, transition, and fricative segments of each nonsense syllable were identified along the temporal domain, and each segment was amplified randomly within each syllable during presentation. Point-biserial correlation coefficients were calculated using the amplitude variation of each segment and the correct and incorrect responses for the corresponding syllable. Results showed that for /see text/ and /see text/, all four groups heavily weighted the fricative segments during perception, whereas the vowel and transition segments received little or no weight. For /see text/, relatively low weights were given to each segment by all four groups. For /see text/, the NH children and adults weighted the transition segment more so than the vowel and fricative segments, whereas the HI children and adults weighted all three segments equally low. Performance-audibility functions of the fricative segments of /see text/ and /see text/ were constructed for each group. In general, maximum performance for each group was reached at lower audibility levels for /see text/ than for /see text/ and steeper functions were observed for the HI groups relative to the NH groups. A decision theory approach was used to confirm the audibility required by each group to achieve a > or =90% level of performance. Results showed both hearing sensitivity and age effects. The HI listeners required lower levels of audibility than the NH listeners to achieve similar levels of performance. Likewise, the adult listeners required lower levels of audibility than the children, although this difference was more substantial for the NH listeners than for the HI listeners.

Perception of English linguadental fricatives by Haitian Kreyol-African American English bilinguals

This study examined Haitian Kreyol/English bilingual speakers’ ability to discriminate non-native fricative and stop contrasts representing substitution patterns in productions of linguadental fricatives in Haitian-accented English and African American English (AAE). Theories of second language (L2) speech learning propose that much of accented speech production in an L2 is due to difficulties in perceiving speech sounds that are not contrastive in L2 learners’ native language or dialect. Neither voiced nor voiceless linguadental fricatives occur in Haitian Kreyol, and Haitian L2 English learners typically substitute /t, d, or v/ for linguadental fricatives. Linguadental fricatives occur in AAE, but vary with context (linguadental, labiodental substitutions or deletions). Haitian/AAE speakers and AAE/standard AE bidialectal speakers were given a categorial discrimination test in which contrasts between linguadental fricatives and alveolar stops or labiodental fricatives in initial, intervocalic, and final position were presented in real words embedded within a short sentence. Preliminary results showed that bilingual Haitian Kreyol/AAE speakers had different patterns of perceptual errors than bidialectal speakers of AAE/SAE. Perceptual error patterns were somewhat related to substitution patterns in Haitian Kreyol and AAE. Further research is underway to investigate age of English learning and dominance of dialect effects on perception. [Work supported by NIDCD-00323.]

Perceptual properties of English fricatives

A series of experiments was conducted to determine the properties that contribute to fricative perception. Listeners’ identification of English fricatives based on fricative-vowel syllables and on isolated fricative noise portions reveals the perceptual salience of each fricative and the extent to which fricative-to-vowel transitions contribute to identification. Two further experiments specifically address perception of the nonsibilant fricatives, which, it has been claimed, may be based more on semantic or facial factors. One experiment investigates how a semantically matching or mismatching precursor affects perception of minimal pairs (e.g., fin-thin). A second experiment determines the extent to which auditory and/or visual information each contribute to fricative perception. Moreover, regression of these perceptual data on acoustic measurements, including noise duration, amplitude, relative amplitude, spectral peak location, spectral moments, and locus equations [see Jongman et al., J. Acoust. Soc. Am. 103, 3086A (1998)], indicates which acoustic properties play a significant role in fricative perception. [Work supported by NIH.]

Effect of relative amplitude manipulation on perception of voiceless fricatives by normal and impaired listeners

In the present study, previous work on the relative amplitude and presentation level was extended to include the voiceless fricative /s/–/■/ contrast. Synthetic CV stimuli were used, in which format transitions, frication duration, and relative amplitude were manipulated. For the relative amplitude manipulation, the amplitude of friction relative to vowel onset in the F3 frequency range was varied across a 20-dB range. Frication duration was varied from 140–50 ms. Stimuli were presented at 90 dB SPL. Eight listeners with normal hearing and three listeners with moderate sensorineural hearing loss participated in the experiment. Results showed that listeners with sensorineural hearing loss gave significantly more /s/ responses for lower relative amplitude values. Results from this study and previous studies [Hedrick et al., J. Acoust. Soc. Am. 98, 1292–1303 (1995); M. S. Hedrick and W. Jesteadt, J. Acoust. Soc. Am. 100, 3398–3407 (1996)] suggest that perception of acoustic cues for consonants by listeners with sensorineural hearing loss may differ from that of listeners with normal hearing, even when the cues are audible.

Some effects of filtered contexts on the perception of vowels and fricatives.

When a phrase is filtered to simulate the effects of a transmission channel, a vowel played shortly afterward is heard as if there is perceptual compensation for the filter. However, a transmission channel would also have an effect on sounds that arrive after the vowel. The present experiments ask whether vowel test sounds are affected by the filtering of a subsequent affricate when there are no other sounds present. Effects on other types of test sounds from different subsequent sounds are measured as well. The experiments also ask whether these effects occur when there is information about the channel in a precursor phrase. Listeners identified words from continua between /It integral of/ and /epsilon t integral of/, /aept/ and /[symbol: see text]/, or /[symbol: see text]/ and /[symbol: see text]/. Filters' frequency responses were the difference of spectral envelopes from the end-point test sounds. The perceptual midpoints of all the continua were shifted in a manner consistent with compensation when the precursors were filtered, as well as when sounds subsequent to the test sound were filtered and there were no precursors. Also, when filtered precursors were present, the shifts increased when filtering was added to sounds subsequent to the test sound. These results indicate that mechanisms of perceptual compensation for filtering by transmission channels use information in preceding sounds in combination with information in following sounds and that these mechanisms operate between different types of speech sounds.