SEARCH

SEARCH BY CITATION

This guide accompanies the following article: Rajka Smiljanic and Ann Bradlow, ‘Speaking and Hearing Clearly: Talker and Listener Factors in Speaking Style Changes’, Langauge and Linguistics Compass 3/1 (2009): DOI: 10.1111/j.1749-818X.2008.00112.x

Author’s Introduction

Clear speech is a listener-oriented speech modification that talkers naturally and spontaneously adopt when listeners have perceptual difficulties due to a hearing impairment, different native language, or presence of noise in the environment. Clear speech research is motivated by the need to understand how these talker-, listener-, and signal-related factors impact normal speech intelligibility. Over time, clear speech studies have explored speech intelligibility variation for a wide range of talker and listener groups, such as normal hearing, hearing impaired, native, non-native, younger and elderly adults, and individuals with Parkinson disease. This body of research has also examined various speech signal conditions, such as signals presented in noise and reverberation and signals modified through digital speech enhancement procedures. The over-arching aim of these studies is to provide insight into, on the one hand, which conversational-to-clear speech acoustic–articulatory modifications contribute to increased intelligibility and, on the other, how the nature of the listener’s perceptual deficit determines which of the clear speech strategies will be most beneficial. Through these investigations, clear speech research provides a window into the interaction of lower level, sensory and perceptual and higher level, cognitive factors that affect language processing. Furthermore, clear speech research informs us about the speech production mechanisms of plasticity that allow talkers to adjust their output to the communicative situation, and about the speech perception processing mechanisms that allow listeners to take advantage of these acoustic–articulatory modifications. This research, thus, has important theoretical implications for developing reliable models of linguistic behavior as well as practical implications for various populations (e.g., non-native and clinical) where speech production and intelligibility are compromised.

Key Readings (in the order they are discussed in the Review)

  • Picheny, M. A., N. I. Durlach, and L. D. Braida. 1985. Speaking clearly for the hard of hearing I: Intelligbility differences between clear and conversational speech. Journal of Speech and Hearing Research 28: 96–103.

    • In this seminal paper, Picheny et al. explored variations in conversational and clear speech intelligibility. The study provided evidence of a significant clear speech benefit for five hearing-impaired subjects listening to nonsense sentences produced in the two speaking styles. The average clear speech intelligibility increase was 17% points independent of talker, listener, presentation level and frequency-gain characteristics.

  • Picheny, M. A., N. I. Durlach, and L. D. Braida. 1986. Speaking clearly for the hard of hearing II: Acoustic characteristics of clear and conversational speech. Journal of Speech and Hearing Research 29: 434–46.

    • In parallel to exploring clear speech intelligibility, these authors investigated the acoustic-phonetic features that characterize clear speech production. Fifty conversational and clear speech sentences produced by three talkers were analyzed acoustically. The results showed that clear speech production typically involved a speaking rate decrease, the insertion of longer and more frequent pauses, vowel space expansion, more frequent word-final consonant releases, greater sound pressure levels and an increase in obstruent intensity.

  • Ferguson, S. H., & Kewley-Port, D. 2007. Talker differences in clear and conversational speech: Acoustic characteristics of vowels. Journal of Speech, Language, and Hearing Research 50: 1241–1255.

    • The focus of this study was to explore talker variability in providing a clear speech intelligibility benefit. Acoustic analyses of talkers who provided the largest and the smallest clear speech intelligibility benefit revealed that longer vowels, expanded vowel space, and an increase in vowel formant movement were general features associated with clear speech production. Vowel lengthening and vowel expansion along F1 and F2 were found to significantly contribute to increased intelligibility. It was also found, however, that some of these characteristics were present in the clear speech of the ‘small intelligibility benefit’ group and absent from the speech of the ‘big intelligibility benefit’ group. Finding the salient acoustic–phonetic clear speech features and establishing their impact on intelligibility still remains an important goal of clear speech research.

  • Krause, J. C. and L. D. Braida. 2004. Acoustic properties of naturally produced clear speech at normal speaking rates. Journal of the Acoustical Society of America 115(1): 362–78.

    • This paper explored characteristics of clear speech produced at normal (i.e., conversational) speaking rates, i.e., controlling for the speaking rate across the speaking style changes. Acoustic analyses revealed no significant vowel space expansion for vowels produced in clear speech at conversational/normal speaking rates compared with clear speech produced at a slower speaking rate. Sentence-in-noise listening tests demonstrated a clear speech intelligibility advantage for normal/conversational clear speech over conversational speech. Two global level acoustic properties that characterized clear speech produced at normal/conversational speaking rates were identified: increased energy in the 1000–3000 Hz range of long-term spectra and increased modulation depth of low frequency changes in the intensity envelope. The authors argued that these results showed that a decrease in speaking rate and vowel space expansion were not necessary clear speech features.

  • Liu, S. and F.-G. Zeng. 2006. Temporal properties in clear speech perception. Journal of the Acoustical Society of America 120(1): 424–32.

    • This is a study that explored the contribution of particular acoustic cues, namely temporal (speaking rate, pauses, temporal envelope), to clear speech intelligibility via cue enhancement techniques. The results showed that increasing the number of pauses contributed to the clear speech advantage. The temporal envelope cue contributed more to the clear speech advantage at high signal-to-noise ratios while the temporal fine structure cue enhanced intelligibility more at low signal-to-noise ratios. These results suggested that acoustic cues for the clear speech advantage are multiple and distributed. Finally, all processed speech stimuli lowered the overall intelligibility compared with naturally produced speech highlighting a difficulty in using digital signal-processing in studying contribution of acoustic cues to speech intelligibility.

  • Smiljanić, R., and A. R. Bradlow. 2005. Production and perception of clear speech in Croatian and English. Journal of the Acoustical Society of America 118(3): Pt. 1, 1677–88.

    • This is one of few cross-language investigations of clear speech production and perception with the goal of exploring the language-universal and language-specific contributions to the clear speech intelligibility advantage. The authors compared clear speech production and perception in English and Croatian, two languages with structural differences between their phonologies. Sentence-in-noise test results revealed a clear speech intelligibility advantage for native listeners in both languages. Acoustic analyses revealed that clear speech in both languages was characterized by a decrease in speaking rate, an expansion of pitch range and a vowel space expansion. In this way, the distances between the contrastive vowel categories were increased in both languages regardless of the sizes of their vowel inventories (there are only 5 vowel categories in Croatian compared to 10+ in English). The results suggested that talkers hyperarticulate even when segments are unlikely to be perceptually confusable.

  • Bradlow, A. R., and T. Bent. 2002. The clear speech effect for non-native listeners Journal of the Acoustical Society of America 112: 272–84.

    • This study investigated whether naturally produced clear speech is an effective intelligibility enhancement strategy for non-native listeners. The results showed that the clear speech benefit was substantially smaller for non-native compared to native listeners. This result was taken to reflect the nature of the non-native listeners’ speech perception deficit, i.e., their inexperience with accessing the language-specific linguistic code of the target language. This further suggested that clear speech is native-listener oriented and that additional exposure to and experience in processing the target language could lead to a larger clear speech benefit for non-native listeners.

  • Smiljanic, R. and A. Bradlow. 2008. Temporal Organization of English Clear and Plain Speech. Journal of the Acoustical Society of America 124.5: 3171–82.

    • To expand clear speech research beyond isolated words and sentences, this study investigated conversational and clear speech production and perception of longer and syntactically more complex paragraphs. The results showed that talkers successfully maintained clear speech modifications across longer stretches of speech. The results of acoustic analyses revealed that in addition to increased durations of segments, clear speech was characterized by insertion of short segments that were dropped or coarticualted with surrounding sounds in conversational speech and an increase in the number of pauses and prosodic phrases. The results also showed that after normalizing for changes in speaking rates, the extent of variability of vocalic and consonantal intervals remained unchanged across speaking styles. That is, talkers adhere to the locked-in local timing relations when producing neighboring segments at any given speaking rate. Increased intelligibility of clear speech may be attributed to prosodic structure enhancement (increased phrasing and enhanced segmentability) and stable global temporal properties.

  • Zeng, F.-G., and S. Liu. 2006. Speech perception in individuals with auditory neuropathy. Journal of Speech, Language, and Hearing Research 49: 367–80.

    • This study compared performance between clear and conversational speech perception in participants with auditory neuropathy (AN) which is characterized by the disruption in auditory nerve activity as reflected in distorted or absent auditory brainstem responses and with normal or nearly normal cochlear amplification function. Individuals with AN have difficulties with processing temporal cues and with speech understanding, particularly in noise. The results showed that participants with AN performed more poorly in speech recognition in noise than did the normal-hearing, cochlear-impaired, and cochlear implant controls. However, a significant clear speech intelligibility benefit was observed for individuals with AN. Electric stimulation via a cochlear implant produced significantly higher intelligibility than acoustic stimulation in both quiet and in noise. Binaural hearing with either diotic acoustic stimulation or combined acoustic and electric stimulation produced significantly higher intelligibility than monaural stimulation in quiet but not in noise.

Focus Questions

  • 1
    What other speaking style changes are similar to clear speech modifications and in what way? In what aspects may these other speaking style changes differ from clear speech? Explain why some of the different effects may arise?
  • 2
    How does the H&H Theory (Lindblom, 1990) provide a conceptual framework for understanding the origin and nature of conversational-to-clear speech adjustments?
  • 3
    What are the broad and long-term goals of clear speech research in terms of its production and perception?
  • 4
    Why is term ‘conversational’ speech not completely accurate in describing the elicited speech in most clear speech studies? How would you adapt the talkers’ task to elicit more spontaneous conversational and clear speech? What problems would this pose for your measurements and analyses?
  • 5
    Why is it difficult to establish a direct link between individual acoustic–phonetic cues and their impact on intelligibility? What possible role in language processing could some of these conversational-to-clear speech modifications play (e.g., speaking rate decrease, vowel space expansion, increased energy in the 1000–3000 Hz range of long-term spectra)?
  • 6
    How do cue enhancement studies address the link between individual acoustic-phonetic cues and intelligibility? What new challenges are introduced with using these signal-processing techniques?
  • 7
    How does examining clear speech production and perception for different talker and listener groups (e.g., native vs. non-native, younger vs. elderly adults, normal hearing vs. hearing impaired) contribute to our understanding of variability in intelligibility?
  • 8
    Provide a detailed description of the sound structure of a language that has not been previously studied in clear speech research. Make predictions about the hyperarticulation strategies that speakers of that language may use in enhancing intelligibility and which may reflect language-specific sound properties.
  • 9
    How does exploring clear speech of single words vs. longer and more complex speech contribute to our understanding of production and perception processes?
  • 10
    What impact can insights from clear speech research studies have on work with clinical populations? How do results from clinical studies inform questions central to clear speech research? How can clinical studies aid in examining plasticity in clear speech production?

How to Use The Article/Topic?

  • This review article (and the cited papers within relevant sections) can be used to develop a clear speech seminar or to enhance discussions in other classes revolving around, for instance, speech intelligibility, speech variation, and signal processing, in Linguistics, Communication Sciences and Disorders and Engineering departments. The topics covered may include:

  • 1
    Clear speech production: this topic will consider the acoustic–articulatory modifications that typically characterize clear speech. You may also consider how these modifications relate to other speaking style changes (e.g., child-directed speech and Lombard speech). Consider how different elicitation methods, i.e., instructions given to the talkers or presence of a dialog partner, may affect the clear speech production strategies. Make predictions about how different languages may vary in implementing clear speech modifications reflecting the language-specific sound structure.
  • 2
    Clear speech perception: the focus of this topic is on listener characteristics that affect clear speech intelligibility. Consider how the nature of different perceptual problems affects the level of clear speech intelligibility advantage and which clear speech acoustic–phonetic cues are beneficial for various listener populations. Propose ways of studying the clear speech effect on other levels of speech processing (e.g., speech stream segmentation and lexical access). Consider how you would go about exploring a direct link between acoustic–phonetic features of clear speech and their impact on intelligibility.
  • 3
    Clear speech and cue enhancement procedures/speech technology: the students should familiarize themselves with techniques available to use in cue enhancement procedures. Consider what are the benefits and potential problems in applying signal-processing methods in clear speech research. What acoustic cues could be examined this way? Consider how these speech processing systems could improve speech perception for various listener populations.
  • 4
    Clear speech in the clinic: this topic explores what can be learned about the nature of a production or perception difficulty from clear speech studies with various clinical populations. Consider whether and how clear speech production and perception tasks can be used to diagnose various deficits. Think about whether and how clear speech training can be used to improve production and perception in various clinical populations.