Elamur Research⌕
☰
home / papers / 2023 / Perceptual identification of oral and nasalized vowels across American English and British English listeners and TTS voices

Perceptual identification of oral and nasalized vowels across American English and British English listeners and TTS voices

Paper · 2023 · Jakub Gwizdzinski, Santiago Barreda, Christopher Carignan et al.

Summary

In English, a vowel before a nasal consonant picks up nasality from the lowered velum. This helps listeners anticipate the nasal coda but changes the vowel's formants, which can blur vowel identity. American English is said to coarticulate more than British English, but the perceptual effect of that difference had not been tested. The authors synthesized twelve /b/-initial monosyllables with Amazon Polly neural voices, two US and two UK. The words used six non-high vowels, each with an oral (/d/) or nasal (/n/) coda. They masked the stimuli with noise and had US and UK listeners pick the word from options that differed in vowel and coda. Speaker dialect mattered most: UK oral-coda words were recognized best, and nasal-coda words were hardest in both dialects. Listeners missed nasal codas more often for UK voices, which fits weaker British anticipatory nasalization. In both varieties, vowel quality was harder to recover before nasals. Listener dialect had little effect. The findings bear on TTS intelligibility and on theories of perceptually driven sound change.

Key points

  • Stimuli: bad/ban, bed/ben, bade/bane, bod/bon, bud/bun, bode/bone. Some are pseudo-words, included to complete the minimal pairs.
  • Voices came from Amazon Polly Neural TTS: "Salli" and "Joanna" (US) and "Amy" and "Emma" (UK), all female.
  • Audio was normalized to 70 dB, mixed with white noise at 0 dB SNR, and padded with 250 ms of noise on each side.
  • Accuracy was 83% for UK oral codas, 67% for US oral codas, 61% for UK nasal codas and 60% for US nasal codas.
  • Missed nasal codas were more frequent for UK voices. Vowel-quality errors rose before nasals in both dialects, and the listener's own dialect barely mattered. Analysis was Bayesian.

Interesting ideas

  • I1. Errors are split by which dimension failed (vowel quality versus coda identity) rather than counted only as right or wrong, which shows what kind of information was lost.
  • I2. One acoustic cue (vowel nasality) helps one decision (detecting the coda) while hurting another (identifying the vowel), so a cue has to be judged against each decision it touches.
  • I3. A crossed design (speaker dialect × listener dialect) separates the variety of the input from the variety of the perceiver.
  • I4. Adding controlled noise pushes performance off the ceiling, which exposes differences that clean conditions hide.
  • I5. Answer options vary along exactly the dimensions under study, so each wrong choice is diagnostic.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (tokens): the tagger and morphology evaluation could report confusions separately by feature (POS, number, tense), so fixes target the specific attribute that breaks.
  • I2 (frames): when a lexical or syntactic cue is added to role or speech-act inference, check its effect on every downstream judgment, not just the one it was designed for.
  • I3 (tree): cross treebank source dialect with parser training dialect to see whether errors come from the input text or from the parser's expectations.
  • I4 (learn): add graded perturbations (typos, dropped function words) to evaluation sets as stress tests that reveal which parser witnesses degrade first.
  • I5 (answer): build multiple-choice probes on fables in which distractors differ from the answer in one controlled aspect (agent, time, quantity), so a wrong answer identifies the faulty reasoning step.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓University College LondonLondon · United Kingdom88 works · 19,432 citations🎓University of California, DavisDavis · United States28 works · 41,916 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksPsychology · 1,552 worksExperimental and Cognitive Psychology · 870 works

🧭 Topics 1Research subjects and languages the work is about.English · 2,263 works

Cite

Cite as Jakub Gwizdzinski, Santiago Barreda, Christopher Carignan and Georgia Zellou. 2023. Perceptual identification of oral and nasalized vowels across American English and British English listeners and TTS voices. Frontiers in Communication 8.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.3389/fcomm.2023.1307547

@inproceedings{gwizdzinski2023perceptual,
  title = {{Perceptual identification of oral and nasalized vowels across American English and British English listeners and TTS voices}},
  author = {Jakub Gwizdzinski, Santiago Barreda, Christopher Carignan and Georgia Zellou},
  booktitle = {Frontiers in Communication 8},
  year = {2023},
  doi = {10.3389/fcomm.2023.1307547},
  url = {https://doi.org/10.3389/fcomm.2023.1307547}
}

References

58 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Amazon (2022). Amazon Web Services. Available online: http://aws.amazon.com/ es/ec2/ (accessed November 12, 2021). ↗
  2. Carignan, C., Chen, J., Harvey, M., Stockigt, C., Simpson, J., and Strangways, S. (2023). An investigation of the dynamics of vowel nasalization in Arabana using machine learning of acoustic features. Lab. Phonol. 14, 1. doi: 10.16995/labphon.9152 ↗
  3. Aoki, N. B., Cohn, M., and Zellou, G. (2022). The clear speech intelligibility benefit for text-to-speech voices: Effects of speaking style and visual guise. JASA Express Lett. 2, 4. doi: 10.1121/10.0010274 ↗
  4. Chen, M. Y. (1997). Acoustic correlates of English and French nasalized vowels. J. Acoust. Soc. Am. 102, 2360–2370. doi: 10.1121/1.419620 ↗
  5. Barreda, S., and Silbert, N. (2023). Bayesian Multilevel Models for Repeated Measures Data: A Conceptual and Practical Introduction in R. Oxfordshire: Taylor and Francis. find
  6. Clopper, C. G. (2014). Sound change in the individual: Effects of exposure on cross-dialect speech processing. Lab. Phonol. 5, 69–90. doi: 10.1515/lp-2014-0004 ↗
  7. Beddor, P. S. (1993). “The perception of nasal vowels,” in Nasals, Nasalization, and the Velum, Phonetics and Phonology vol. 5, eds M. K. Huffman and R. A. Krakow (Cambridge, MA: Academic Press). ↗
  8. Cohn, M., Segedin, B. F., and Zellou, G. (2022). Acoustic-phonetic properties of Siri-and human-directed speech. J. Phonet. 90, 101123. doi: 10.1016/j.wocn.2021.101123 ↗
  9. Beddor, P. S. (2007). “Nasals and nasalization: The relation between segmental and coarticulatory timing,” in Proceedings of the 16th International Congress of Phonetic Sciences, Saarbrücken, 249–254. find
  10. Cohn, M., and Zellou, G. (2020). “Perception of concatenative vs. neural text-to-speech (TTS): Differences in intelligibility in noise and language attitudes,” in Proceedings of Interspeech. find
  11. Beddor, P. S. (2009). A coarticulatory path to sound change. Language 165, 785–821. doi: 10.1353/lan.0.0165 ↗
  12. Cox, F., and Palethorpe, S. (2014). “Phonologisation of vowel duration and nasalised/æ/in Australian English,” in Proceedings of the 15th Australasian International Conference on Speech Science and Technology, 33–36. find
  13. Beddor, P. S., and Krakow, R. A. (1999). Perception of coarticulatory nasalization by speakers of English and Thai: Evidence for partial compensation. J. Acoust. Soc. Am. 106, 2868–2887. doi: 10.1121/1.428111 ↗
  14. De Decker, P. M., and Nycz, J. R. (2012). Are tense [æ]s really tense? The mapping between articulation and acoustics. Lingua 122, 810–821. doi: 10.1016/j.lingua.2012.01.003 ↗
  15. Beddor, P. S., Krakow, R. A., and Goldstein, L. M. (1986). Perceptual constraints and phonological change: a study of nasal vowel height. Phonology 3, 197–217. doi: 10.1017/S0952675700000646 ↗
  16. Delvaux, V. (2009). Perception du contraste de nasalité vocalique en français. J. French Lang. Stud. 19, 25–59. doi: 10.1017/S0959269508003566 ↗
  17. Beddor, P. S., McGowan, K. B., Boland, J. E., Coetzee, A. W., and Brasher, A. (2013). The time course of perception of coarticulation. J. Acoust. Soc. Am. 133, 2350–2366. doi: 10.1121/1.4794366 ↗
  18. Diakoumakou, E. (2004). Coarticulatory Vowel Nasalization in Modern Greek. Ann Arbor: University of Michigan. find
  19. Dodd, N., Cohn, M., and Zellou, G. (2023). Comparing alignment toward American, British, and Indian English text-to-speech (TTS) voices: Influence of social attitudes and talker guise. Front. Comp. Sci. 5, 1204211. doi: 10.3389/fcomp.2023.1204211 ↗
  20. Bilal, D., and Barfield, J. K. (2021). Hey there! what do you look like? user voice switching and interface mirroring in voice-enabled digital assistants (VDAs). Proc. Assoc. Inform. Sci. Technol. 58, 1–12. doi: 10.1002/ pra2.431 ↗
  21. Gessinger, I., Cohn, M., Zellou, G., and Möbius, B. (2022). Cross-cultural comparison of gradient emotion perception: human vs. Alexa TTS voices. Proc. Interspeech 2022, 4970–4974. doi: 10.21437/Interspeech.2022-146 ↗
  22. Bongiovanni, S. (2021). Acoustic investigation of anticipatory vowel nasalization in a Caribbean and a non-Caribbean dialect of Spanish. Linguist. Vangu. 7, 20200008. doi: 10.1515/lingvan-2020-0008 ↗
  23. Hajek, J. (2013). “Vowel nasalization,” in The World Atlas of Language Structures Online, eds. M. S. Dryer and M. Haspelmath. Leipzig: Max Planck Institute for Evolutionary Anthropology. find
  24. Bürkner, P.-C. (2017). brms: An R package for Bayesian multilevel models using Stan. J. Statist. Softw. 80, 1–28. doi: 10.18637/jss.v080.i01 ↗
  25. Burridge, K., and Kortmann, B. (2008). Varieties of English. The Pacific and Australasia. Berlin: De Gruyter Mouton. doi: 10.1515/9783110208412.0.23 ↗
  26. Hartley, L. C., and Preston, D. R. (2002). “The names of US English: Valley girl, cowboy, yankee, normal, nasal and ignorant,” in Standard English. London: Routledge, 207–238. find
  27. Carignan, C. (2014). An acoustic and articulatory examination of the oral in nasal: the oral articulations of French nasal vowels are not arbitrary. J. Phonet. 46, 23–33. doi: 10.1016/j.wocn.2014.05.001 ↗
  28. Hosseinzadeh, N. M., Kambuziya, A. K. Z., and Shariati, M. (2015). British and American phonetic varieties. J. Lang. Teach. Res. 6, 647–655. doi: 10.17507/jltr.0603.23 ↗
  29. Carignan, C. (2018a). Using ultrasound and nasalance to separate oral and nasal contributions to formant frequencies of nasalized vowels. J. Acoust. Soc. Am. 143, 2588–2601. doi: 10.1121/1.5034760 ↗
  30. Krakow, R. A., Beddor, P. S., Goldstein, L. M., and Fowler, C. A. (1988). Coarticulatory influences on the perceived height of nasal vowels. J. Acoust. Soc. Am. 83, 1146–1158. doi: 10.1121/1.396059 ↗
  31. Carignan, C. (2018b). An Examination of Oral Articulation of Vowel Nasality in the Light of the Independent Effects of Nasalization on Vowel Quality. Padua: Associazione Italiana Scienze della Voce. find
  32. Maddieson, I. (2013). “Absence of common consonants,” in The World Atlas of Language Structures Online, eds. M. S. Dryer, and M. Haspelmath. Leipzig: Max Planck Institute for Evolutionary Anthropology. find
  33. Carignan, C. (2018c). Using naïve listener imitations of native speaker productions to investigate mechanisms of listener-based sound change. Lab. Phonol. 9, 1. doi: 10.5334/labphon.136 ↗
  34. Mielke, J., Carignan, C., and Thomas, E. R. (2017). The articulatory dynamics of pre-velar and pre-nasal/æ/-raising in English: an ultrasound study. J. Acou. Soc. Am. 142, 332–349. doi: 10.1121/1.4991348 ↗
  35. Miller, G. A., and Nicely, P. E. (1955). An analysis of perceptual confusions among some English consonants. J. Acou. Soc. Am. 27, 338–352. doi: 10.1121/1.1907526 ↗
  36. Styler, W. (2017). On the acoustical features of vowel nasality in English and French. J. Acou. Soc. Am. 142, 2469–2482. doi: 10.1121/1.5008854 ↗
  37. Ohala, J. J. (1993). Coarticulation and phonology. Lang. Speech 36, 155–170. doi: 10.1177/002383099303600303 ↗
  38. Tamminga, M., and Zellou, G. (2015). “Cross-dialectal differences in nasal coarticulation in American English,” in ICPhS. find
  39. Ohala, J. J., and Ohala, M. (1995). Speech perception and lexical representation: the role of vowel nasalization in Hindi and English. Phonology and phonetic evidence. Papers in Lab. Phonol. IV, 41–60. doi: 10.1017/CBO9780511554315.004 ↗
  40. Wright, J. T. (1975). “Effects of vowel nasalization on the perception of vowel height,” in Nasálfest: Papers from a Symposium on Nasals and Nasalization, eds. C. A. Ferguson, L. M. Hyman, and J. J. Ohala. Palo Alto, CA: Stanford University Language Universals Project, 373–388. find
  41. Onsuwan, C. (2005). Temporal Relations Between Consonants and Vowels in Thai Syllables. Ann Arbor: University of Michigan. find
  42. Wright, J. T. (1986). “The behavior of nasalized vowels in perceptual vowel space,” in Experimental Phonology, J. J. Ohala, and J. J. Jaeger. New York: Academic Press, 45–67. find
  43. Pittam, J. (1987). Listeners’ evaluations of voice quality in Australian English speakers. Lang. Speech 30, 99–113. doi: 10.1177/002383098703000201 ↗
  44. Zellou, G. (2017). Individual differences in the production of nasal coarticulation and perceptual compensation. J. Phonet. 61, 13–29. doi: 10.1016/j.wocn.2016.12.002 ↗
  45. R Core Team (2021). R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing. Available online at: https://www.R-project.org/ (accessed November 1, 2022). ↗
  46. Zellou, G. (2022). Coarticulation in Phonology. Cambridge: Cambridge University Press. doi: 10.1017/9781009082488 ↗
  47. Scarborough, R. (2012). Lexical similarity and speech production: neighborhoods for nonwords. Lingua 122, 164–176. doi: 10.1016/j.lingua.2011.06.006 ↗
  48. Zellou, G., and Brotherton, C. (2021). Phonetic imitation of multidimensional acoustic variation of the nasal split short-a system. Speech Commun. 135, 54–65. doi: 10.1016/j.specom.2021.10.005 ↗
  49. Scarborough, R. (2013). Neighborhood-conditioned patterns in phonetic detail: relating coarticulation and hyperarticulation. J. Phonet. 41, 491–508. doi: 10.1016/j.wocn.2013.09.004 ↗
  50. Zellou, G., Cohn, M., and Block, A. (2021). Partial compensation for coarticulatory vowel nasalization across concatenative and neural text-to-speech. J. Acou. Soc. Am. 149, 3424–3436. doi: 10.1121/10.0004989 ↗
  51. Scarborough, R., and Zellou, G. (2012). Acoustic and perceptual similarity in coarticulatorily nasalized vowels. Interspeech 2012, 1408–1411. doi: 10.21437/Interspeech.2012-669 ↗
  52. Zellou, G., and Dahan, D. (2019). Listeners maintain phonological uncertainty over time and across words: The case of vowel nasality in English. J. Phonet. 76, 100910. doi: 10.1016/j.wocn.2019.06.001 ↗
  53. Scarborough, R., and Zellou, G. (2013). Clarity in communication: “Clear” speech authenticity and lexical neighborhood density effects in speech production and perception. J. Acou. Soc. Am. 134, 3793–3807. doi: 10.1121/1.48 24120 ↗
  54. Zellou, G., Scarborough, R., and Kemp, R. (2020). Secondary phonetic cues in the production of the nasal short-a system in California English. Interspeech 2020, 25–29. doi: 10.21437/Interspeech.2020-1322 ↗
  55. Stan Development Team (2023). Stan Modeling Language Users Guide and Reference Manual, Version. Available online at: https://mc-stan.org (accessed November 1, 2022). ↗
  56. Zellou, G., Scarborough, R., and Nielsen, K. (2016). Phonetic imitation of coarticulatory vowel nasalization. J. Acou. Soc. Am. 140, 3560–3575. doi: 10.1121/1.4966232 ↗
  57. Stoakes, H. M., Fletcher, J. M., and Butcher, A. R. (2020). Nasal coarticulation in Bininj Kunwok: an aerodynamic analysis. J. Int. Phonetic Assoc. 50, 305–332. doi: 10.1017/S0025100318000282 ↗
  58. Zellou, G., and Tamminga, M. (2014). Nasal coarticulation changes over time in Philadelphia English. J. Phonet. 47, 18–35. doi: 10.1016/j.wocn.2014.09.002 ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/gwizdzinski-2023-perceptual-identification-oral-nasalized-vowels/