Digging into sentence intelligibility: Interactions with noise

Previous research on tone perception has identified several important pitch related cues including average pitch height (AH), contour, onset and offset, and the weighting of these cues has shown to be language dependent. However, since multiple pitch cues are covarying with each other, few studies have directly compared relative importance of these cues. It is also not clear whether there is a same ranking of cues in speech and non-speech stimuli. The current study aims to tease apart the relative role of each cue using AX discrimination. Four pairs of tone contrasts with minimal pitch differences were created. Tone contrasts within contour condition are two level tones with 7 Hz differences. Tone contrasts within AH, onset and offset conditions have one rising tone and one falling tone sharing the same AH, onset and offset respectively. If one cue is important, then when this cue is kept constant, variation in other cues should be hard to perceive. 48 Mandarin speakers and 48 Cantonese speakers were recruited. Results showed highest importance of AH for both Mandarin and Cantonese listeners and higher importance of contour (offset) than onset for Mandarin (Cantonese) listeners in speech stimuli. This ranking was not held for nonspeech stimuli.

A number of writers (Solomon & Poscman, 1951; Postman, 1954) have suggested that mosc studies relating motivation to perception are mosc readily understood within the context of frequency of usage. Stimuli which are more frequently used or experienced will be more readily recognized in a brief tachistoscopic flash while stimuli which, through past association with noxious consequences, are less frequently used will be less easily recognized. The essence of chis view is that frequency of usage acts directly in modifying the ease or d~fficulty of making visual or auditory discriminations. A number of very impressive studies have been presented by these writers which clearly support the hypothesis that frequency of usage, either estimated statistically or built up experimenrally, bears a logarithmic relationship to ease of tachistoscopic recognition. Recently, Goldiamond and Hawkins (1958) have challenged the idea that frequency of usage affects the ease or difficulty of word discrirninacion. These writers suggest that the results of earlier studies on this problem are an artifact of what they term response bias. Goldiamond and Hawkins have demonstraced how the log relationship between frequency of usage and ease of recognition can be obtained without flashing real words in the tachistoscope. After building up different frequencies of usage experimentally, Goldiamond and Hawkins illustrate the manner in which such training biases the probability that various words will be used as recognition guesses. Further, they demonstrate how such response biases can interact with an ascending method of limits to produce the illusion of more rapid recognition for more frequently experienced stimuli. Although the demonstration of the role of response bias in perceptual recognition experiments is extremely valuable, such a de~nonstration should not be confused with proving that response bias is the only variable affected by frequency of usage. It is necessary to carry this analysis through to situations in which visual perception is smdied in a manner which does noc permit response bias to affect recognition scores. Should no relationship be 'This smdy was performed by the senior author while a Fulbright research scholar ac the Psychological Laboratory, University of Copenhagen, 1960-61. The aurhors would like to express their appreciation to Professor E. Tranekjaer Rasmussen, Director of the Laboratory for his kind assistance during the conducr of chis research. The senior author is particularly grateful to Karin Fennow, Director of che United States Educational Foundation, Denmark, for her continual kindness and support on the problems that necessarily arise in the course of performing research in another country than one's own.

Many dialects of French have merged /e-ε/ to /e/ in word-final context. The present study investigated the stability of this contrast in Canadian French. Productions of four Canadian French-dominant speakers (two monolingual, two bilingual) were recorded and analyzed. Real (e.g., “thé”) and nonsense (e.g., “gispais”) words ending with /e/ or /ε/ were used, as well as real and nonsense verbs distinguished morphosyntactically by the same vowel contrast (e.g., first person singular future “parlerai” versus first person singular conditional “parlerais”), all embedded in carrier sentences. Results showed that participants maintained spectral and duration distinction for the vowel contrast when preceded by labial, coronal, and back stops in real and nonsense words. All participants showed varying degrees of coarticulation for /e/ when preceded by /r/; /e/ was spectrally lower than in other contexts. Three of the four participants maintained a stable distinction in morphosyntactic context: monolinguals exhibited the best retention of the distinction, whereas the bilinguals had partially or completely overlapping distributions, merging /e-ε/ to /ε/, not /e/. These patterns suggest difficulties in perception of this contrast for English late L2 learners of French, due to the fact that this contrast is not phonotactically possible in English. [Work supported by NIH F31DC008075.]

Purpose – Previous research has suggested that brand knowledge could be affected by companies and consumer characteristics such as consumer personality. The purpose of this study is to analyse the impacts of gender, field of education, and having consumed the brand, on consumers' brand knowledge.

The verbal transformation effect, the phenomenon in which continued listening to recorded repetitions of a stimulus produces perceptual illusory changes in normal listeners, has been elicited by a variety of speech stimuli, including isolated vowels, consonant-vowel nonsense syllables, words, phrases, and sentences. The purpose of this investigation was to determine the effectiveness of non-speech sounds as auditory stimuli in eliciting a non-verbal analog to this perceptual phenomenon. A total of five stimuli, three pure tones (250, 1000, and 4000 Hz), white noise, and a five-note musical motif, were repeatedly presented for a five-minute period to each of 25 adult listeners. Results indicate that the transformations elicited by the non-speech sounds are similar to those produced by speech stimuli in number of forms elicited, specific forms, and types of transformations. However, listeners usually require more repetitions of the non-speech stimuli before perceiving transformations. Implications of these findings for the proposed theory that verbal transformations represent skilled reorganizational mechanisms which aid in speech perception are discussed.

This study explored clear speech (CS) and noise-adapted speech (NAS) intelligibility benefits for native and nonnative English listeners. It also examined how the two speaking style adaptations interact with maskers that vary from purely energetic to largely informational at different signal-to-noise ratios (SNRs). Materials consisted of 40 sentences produced by 10 young adult talkers in a conversational and a clear speaking style under two conditions: (a) in quiet and (b) in response to speech-shaped noise (SSN) played over headphones (NAS). Young adult native (Experiment 1) and nonnative (Experiment 2) English listeners heard target sentences presented in two-talker (2T) babble, six-talker (6T) babble, or SSN and at an "easier" and a "harder" SNR. When talkers produced CS and NAS, word recognition accuracy was significantly improved for both listener groups. The largest intelligibility benefit was obtained for the CS produced in response to noise (CS+NAS). Overall accuracy was highest in 2T babble. Accuracy was higher in SSN than in 6T babble for nonnative listeners at both levels of listening difficulty but only at a more difficult SNR for native listeners. Listeners benefited from CS and NAS most in the presence of SSN and least in 2T babble. When SNRs were the same for the two listener groups, native listeners outperformed nonnative listeners in almost all listening conditions, but nonnative listeners benefited more from CS and NAS in 6T babble than native listeners did. Combined speaking style enhancements, CS+NAS, provided the largest intelligibility increases for native and nonnative listeners in all listening conditions. The results add to the body of evidence supporting speech-oriented, behavioral therapy techniques for maximizing speech intelligibility in everyday listening situations.

Previous research has shown that it is possible to predict which speaker is attended in a multispeaker scene by analyzing a listener's electroencephalography (EEG) activity. In this study, existing linear models that learn the mapping from neural activity to an attended speech envelope are replaced by a non-linear neural network (NN). The proposed architecture takes into account the temporal context of the estimated envelope and is evaluated using EEG data obtained from 20 normal-hearing listeners who focused on one speaker in a two-speaker setting. The network is optimized with respect to the frequency range and the temporal segmentation of the EEG input, as well as the cost function used to estimate the model parameters. To identify the salient cues involved in auditory attention, a relevance algorithm is applied that highlights the electrode signals most important for attention decoding. In contrast to linear approaches, the NN profits from a wider EEG frequency range (1-32Hz) and achieves a performance seven times higher than the linear baseline. Relevant EEG activations following the speech stimulus after 170ms at physiologically plausible locations were found. This was not observed when the model was trained on the unattended speaker. Our findings therefore indicate that non-linear NNs can provide insight into physiological processes by analyzing EEG activity.

Non‐native listeners have difficulty perceiving context cues in speech. The current study intended to quantify the gain in non‐native listeners’ perception of English speech due to context by adopting the Boothroyd–Nittrouer model [J. Acoust. Soc. Am. 84, 101–114 (1988)], in which context effects are expressed via factors j and k. These factors are mathematically equal to the total number of components in a speech stimulus with no context, but are reduced when context is present. Ten normal‐hearing, non‐native and 10 native listeners participated in this study. To investigate the effect of acoustic‐phonetic cues, 12 lists of CVC English words and 12 lists of nonsense words were employed. To investigate the effect of semantic‐syntactic cues, three sets of 20 four‐word sentences were employed. These sentences differed in the amount of semantic and syntactic cues contained in the stimuli. All stimuli were presented binaurally at 45 dB HL. A concomitant speech‐weighted steady‐state noise was added at 39 and 45 dB HL for words and sentences, respectively. The non‐native listeners’ verbal responses were recorded digitally. It was expected that j and k are larger in value in non‐native than native listeners, and that the magnitude of the difference is associated with the non‐native listeners’ English learning history.

It has been shown that foreign-accented speech is more difficult for listeners to process than native speech when presented under noisy listening conditions. It has also been shown that errors in vowel productions are more damaging to connected speech intelligibility than are errors or distortions in consonant production. This study compared the intelligibility of Spanish-accented English vowels in isolated words with that of short samples of connected speech. Five native Spanish speakers who had learned English as adults read a list of 42 monosyllabic English words selected so that 14 Spanish-accented English monophthongs could be presented to native listeners in a three-alternative forced-choice identification task. In each trial the target and two nearby (in the vowel space) distracters were presented. Talkers also read a set of 20 phonetically balanced sentences. Ten native English listeners identified the vowels in the isolated word task and judged the intelligibility of the sentences in a simple write-down procedure. While overall sentence intelligibility scores are not very well predicted by word intelligibility scores when the full set of vowel contrasts is included, predictability improves when a subset of the vowels is chosen. [Work supported by Purdue Research Foundation.]

Rules for writing certain phonemes (/s/, /z/, /r/, /t/ and /d/) in Pitman New Era shorthand are systematically examined and their frequency of usage estimated. In particular, we identify three types of rule (default rules, other rules and short forms) and suggest correspondences between them and the classification of words in standard English orthography as regular, exception and strange. Twelve experienced shorthand writers wrote to dictation, in shorthand, passages of prose and lists of nonsense words. With real words there were effects of word frequency on percent correct. These were strongest for short forms and weakest for default rules. Nonsense words were less accurately transcribed than real words, and this effect was particularly marked for vocalic /r/. We conclude that lexical effects are just as strong for shorthand (which is taught by explicit rules) as it is for English orthography (where the rules are less systematic). In particular, there is no support for the notion that shorthand...

Speech intelligibility is usually assessed using subjective tests, but several objective measures of speech intelligibility have also been proposed. This study compares intelligibility of IEEE sentences in collocated speech-shaped noise, recorded through a low-cost amplifier and three current-issue NHS hearing aids with single-channel noise reduction settings switched on and off. Results from twenty-one normal hearing listeners indicates that single-channel noise reduction algorithms can significantly improve intelligibility of noisy speech (p < 0.05), but no differences can be seen between different hearing aid models currently available on the NHS. The low-cost, off-the shelf hearing amplifier performed significantly worse than the NHS prescribed devices. Three objective speech intelligibility measures were applied to the recordings. None of the objective metrics was found to accurately predict the outcomes seen in this study for all hearing aid conditions, though general trends can be predicted (r = 0.83 overall, but for any discrete signal-to-noise ratio across all conditions, maximum r = 0.64). Machine-learning approaches were used to refit parameters from the best performing of these metrics. Although performance was improved, problems in predicting particular conditions remain, suggesting that objective measures do not encompass all relevant factors. [Action on Hearing Loss (England/Wales Registered Charity Number 207720).]

We examined how native-language experience influences processing a second language, focusing on how native Dutch listeners who learned English as a second language perceive the English voiceless consonant /p/. Previous research [J. E. Flege and W. Eefting, Speech Commun. 6, 185–202 (1987)] shows that the voiced–voiceless boundary for an (English-based) voice-onset-time (VOT) series is located at a shorter VOT for such bilingual listeners than for native English listeners, consistent with the fact that voiceless stops are produced with shorter VOTs in Dutch than in English. We asked whether such bilinguals also differ from native English listeners in which stimuli throughout the series are perceived as reasonable exemplars of /p/. Native English listeners and native Dutch listeners were tested on a three-choice identification task with an (English-based) extended VOT series that ranged from /ba/ to /pa/ to an ‘‘unnatural’’ exaggerated /pa/, labeled */pa/. Both the /b/-/p/ and /p/-*/p/ boundaries were located at shorter VOTs for the native Dutch than the native English listeners, indicating that Dutch native-language experience influenced the entire range of VOTs perceived as reasonable exemplars of the /p/ category. Thus native-language experience has a comprehensive influence on the mapping from acoustic signal to phonetic category. [Work supported by NIH/NIDCD.]

This study addresses the assessment of developmental apraxia of speech (DAS) in children. For this, 11 children with a clear diagnosis of DAS were selected, based on documented speech history and perceptual evaluation of speech. The children with DAS, as well as 11 normal-speaking children, produced singleton real word and nonsense word imitations elicited in a standardised way. Phonetic transcriptions were analysed and errors in consonants classified. The results showed, firstly, that the children with DAS produced similar types of consonant errors as has been reported in the literature, which corroborates the method of elicitation as a valid procedure to assess relevant speech symptoms of DAS. Secondly, a large quantitative difference between children with DAS and normal-speaking children was found, in that children with DAS produced an overall higher rate of singleton consonant errors (substitutions, omissions, distortions) and cluster errors (cluster reductions) than the normal-speaking children. For the DAS group, the substitution-rate, particularly in real words (as opposed to nonsense words), was significantly correlated with severity as rated by two speech and language pathologists. This suggests that substitution-rate yields an adequate measure of severity of DAS. Thirdly, a qualitative difference between both subject groups emerged. Children with DAS did not benefit from the lexical status of the utterance (real versus nonsense word) to the same extent as normal-speaking children. Based on these findings the nature of the underlying deficits in speech production in DAS is discussed.

Recording equipment was transported to Yuendumu, N. T., Australia, to obtain data on tongue and lip activity of aboriginals during speech and swallowing. Strain gauge pressure transducers were fitted into custom constructed plates to be placed in the speaker's mouth to record tongue and lip contact at several locations. Eighteen speakers whose native language is Walbiri acted as subjects. Eight prerecorded Walbiri words and four nonsense words, five repetitions each, were presented to the subjects. These tokens contained three of the Walbiri modifications of the phoneme /t/ situated intervocalically in each word. Distinct articulation patterns of both tongue and lips were found for the three phonemes regardless of whether they were located in meaningful or nonsense words. Furthermore, pressure patterns for the phoneme most like the American language /t/ were very similar to patterns for American speakers. This was true even though tongue and lip movements during swallowing were quite different from Americans. [Kenneth Hale assisted in preparing the speech stimulus. Support from the Fulbright program, NIDR, and Southern Society of Orthodontists is gratefully acknowledged.]

Previous research has shown that creative mindsets influence creativity. Compared with people with a fixed creative mindset, those with a growth creative mindset performed better in creative tasks. The underlying mechanism, however, is not completely understood. The present study has extended previous works to explore whether metacognitive strategy monitoring and control influence the relationship between creative mindsets and divergent thinking performance. The thinking aloud method was used to summarize four strategies in a divergent thinking task (an alternative uses task, AUT) in a pilot study: memory retrieval, splitting, property-based, and general use strategies. In the formal study, the creative mindsets scale, AUT, self-strategic utility judgment (i.e., an index of metacognitive strategy monitoring), and frequency of strategies usage (i.e., an index of metacognitive strategy control) were used to explore the relationships among creative mindsets, divergent thinking, and metacognitive strategy monitoring and control. The results indicated a positive correlation between a growth creative mindset and divergent thinking but a negative correlation between a fixed creative mindset and divergent thinking. More importantly, there were identified mediating roles of metacognitive monitoring and control of splitting and property-based strategies in the relationship between creative mindsets and divergent thinking. The findings reveal that creative mindsets are a critical predictor of divergent thinking, and metacognitive monitoring and control of abstract strategies mediate this association.