Elamur Research⌕
☰
home / papers / 2020 / Arabic Speech Rhythm Corpus: Read and Spontaneous Speaking Styles

Arabic Speech Rhythm Corpus: Read and Spontaneous Speaking Styles

Summary

The authors address the absence of a forensically usable Arabic speech database for studying rhythm and for forensic voice comparison (FVC), where between- and within-speaker variability must be known. They recorded 10 Egyptian speakers, balanced for gender, in two styles. The read part reuses the BonnTempo corpus (BTC) design. A phonetically rich, fully diacritized short story was written originally in Arabic rather than translated. In the spontaneous part, speakers talked about their daily life or studies and then used a map to explain how to get to the university from a well-known nearby place. The recordings carry both manual and automatic time labels, which supports quantitative rhythm analysis of Modern Standard Arabic (MSA) and the Egyptian dialect. The collection follows the Morrison, Rose and Zhang (2012) FVC protocol. The resource is presented as work in progress and as an Arabic extension of BTC. It allows direct rhythm comparison with English, German, French and Italian, and could support speaker-recognition and speaking-style research.

Key points

  • 10 gender-balanced Egyptian Arabic speakers, recorded in read and spontaneous styles.
  • The read passage has roughly 8 sentences and 178 phonological syllables, about 4 minutes per speaker. A professional linguist added full diacritics.
  • Spontaneous speech has two parts: free talk lasting over a minute, then a map-based route description.
  • Collection follows the FVC protocol's three criteria: sessions recorded at different times, varied speaking styles, and usability under channel mismatch.
  • The corpus extends BonnTempo (5 speaking rates; English, German, French, Italian) and has manual plus automatic time labeling covering MSA and Egyptian Arabic.

Interesting ideas

  • I1. The authors avoid translated standard texts such as "The North Wind and the Sun" because translations may read unnaturally and lose phonetic balance, so they wrote native material instead.
  • I2. Copying an existing corpus's collection protocol exactly makes the new data directly comparable with the older languages.
  • I3. The same speakers produce both a controlled read task and free spontaneous speech, which separates style effects from speaker effects.
  • I4. Full diacritization removes pronunciation ambiguity before recording, which fixes the intended reading in the stimulus itself.
  • I5. A map serves as a shared visual stimulus, so spontaneous route descriptions stay grounded and comparable across speakers.

For Mova

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

  • I1 (answer): Build school-test and fable QA items from texts originally written in English rather than translated ones, so unnatural phrasing does not distort what the reader is measured on.
  • I2 (learn): When adding new evaluation sets, mirror the format and procedure of an established benchmark, so score differences reflect the system rather than the setup.
  • I3 (learn): Pair edited and conversational texts on the same content to check whether parser or frame accuracy drops because of register rather than topic.
  • I4 (tokens): Keep a small gold set where lemma and POS ambiguities are explicitly resolved in the input, to isolate tagger errors from genuine ambiguity.
  • I5 (world): Use direction-giving texts with a known underlying map as test material for tracking spatial entities and event order, since the correct world state can be checked.

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

Fields: Linguistics & languages

Topics Semitic languages

 Institutions

🎓University of ZurichZurich · Switzerland44 works · 10,252 citations

Also on the byline: Alzahra University.

Cite

Cite as Omnia Ibrahim, Homa Asadi, Eman Kassem and Volker Dellwo. 2020. Arabic Speech Rhythm Corpus: Read and Spontaneous Speaking Styles. Proceedings of the Twelfth Language Resources and Evaluation Conference.

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{ibrahim2020arabic,
  title = {{Arabic Speech Rhythm Corpus: Read and Spontaneous Speaking Styles}},
  author = {Omnia Ibrahim, Homa Asadi, Eman Kassem and Volker Dellwo},
  booktitle = {Proceedings of the Twelfth Language Resources and Evaluation Conference},
  year = {2020},
  url = {https://aclanthology.org/2020.lrec-1.657/}
}

References

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

  1. Alghamdi M., Alhamid A., and Aldasuqi M.,(2003). Database of Arabic Sounds: Sentences, Technical Report, Saudi Arabia. find
  2. Al-Sobh, M., Abu-Melhim, A., &Bani-Hani, N. (2015). Diglossia as a result of language variation in Arabic: Possible solutions in light of language planning. Journal of Language Teaching and Research, 6(2),274- .27 find
  3. Al-Suwaiyan L. A. (2018). Diglossia in the Arabic Language, International Journal of Language and Linguistics,Vol. 5, No. 3, September 2018, doi:10.30845/ijll.v5n3p22 ↗
  4. Altuwaim, Y. A. Alotaibi and S. Selouani (2014). Investigation into the speech rhythm of two Saudi dialects using the SAAVB corpus. 2014 6th International Symposium on Communications, Control and Signal Processing (ISCCSP), Athens, 2014, pp. 632-635. doi: 10.1109/ISCCSP.2014.6877954 ↗
  5. Bateson, M. C. (2003). Arabic Language Handbook. Washington, Georgetown University Press. find
  6. Boersma, Paul & Weenink, David (2019). Praat: doing phonetics by computer [Computer program]. Version 6.1.06, retrieved 8 November 2019 from http://www.praat.org/ ↗
  7. Dellwo, V., & Fourcin, A. (2013). Rhythmic characteristics of voice between and within languages. TRANEL - Travaux neuchâtelois de linguistique, 59, 87–107. find
  8. Dellwo, V., Leemann, A., and Kolly, M.-J. (2015). Rhythmic variability between speakers: articulatory, prosodic, and linguistic factors, Journal of the Acoustical Society of America 137: 1513-1528. find
  9. Dellwo, V., Steiner, I., Aschenberner, B., Dankovičová, J., and Wagner, P. (2004). The BonnTempo-Corpus and BonnTempo-Tools: A database for the study of speech rhythm and rate. in Proceedings of the 8th ICSLP, Jeju Island, Korea. find
  10. Droua-Hamdani, S.-A. Selouani, M. Boudraa, W. Cichocki, (2010). Algerian arabic rhythm classification., in: ExLing, pp. 37–40. find
  11. Enzinger E. and Morrison G. S. (2012). The importance of using between-session test data in evaluating the performance of forensic-voice-comparison systems. in the 14th Australasian International Conference on Speech Science and Technology, Sydney, Australia, Proceedings , pp. 137-140. Ghazali, S./R. Hamdi, R./M. find
  12. Barkat, M. (2002). Speech rhythm variation in Arabic dialects. – In: Bernard Bel/I. Marlin (eds.), Proceedings of the Speech Prosody 2002 Conference, 11-13 April 2002, Aix-en-Provence: Laboratoire Parole et Langage, 127-132. find
  13. Grabe, E., & Low, E. L. (2002). Durational variability in speech and the rhythm class hypothesis. In Papers in Laboratory Phonology VII (Eds. Gussenhoven, E. & Low, E. L.), Berlin: Mouton de Gruyter, 515–546. find
  14. Grabe, E., Low, E.L. (2003). Durational variability in speech and the rhythm class hypothesis. Papers in laboratory phonology 7, 515-546. find
  15. Hetzron, R. (1997). Classical Arabic.The Semitic languages. New York, NY: Routledge. find
  16. Kinoshita, K. (2001). Testing realistic forensic speaker identification in Japanese: A likelihood ratio based approach using formants . Ph.D. dissertation, Australian National University. find
  17. Kirchhoff, K. and Vergyri, D. (2005). Cross-dialectal data sharing for acoustic modeling in arabic speech recognition. Speech Communication , 46(1):37–51. find
  18. Hamdi, R., Ghazali, S., & Barkat-Defradas, M. (2005). Syllable Structure in Spoken Arabic: A comparative investigation. Interspeech, 2245– 2248. find
  19. He, L. and Dellwo, V. (2016). The role of syllable intensity in between-speaker rhythmic variability. International Journal of Speech, Language and the Law 2 3(2): 243–273. https://doi.org/10.1558/ijsll.v23i2.30345 ↗
  20. Jang, T-Y. (2009). Automatic assessment of non-native prosody using rhythm metrics: Focusing on Korean speakers’ English pronunciation. In SFU Working Papers in Linguistics Vol. 2 . Simon Fraser University, Vancouver, Canada. find
  21. Leemann, M.-J. Kolly, and V. Dellwo (2014). Speaker-individuality in suprasegmental temporal features: Implications for forensic voice comparison. Forensic Sci. Int. 238, 59–67. find
  22. Marcus, S. M. (1981). Acoustic determinants of perceptual center (P-center) location. Perception & Psychophysics, 30, 247–256. find
  23. Morrison, G. S., Rose, P., and Zhang, C. (2012). “Protocol for the collection of databases of recordings for forensic-voice-comparison research and practice,” Aus. J. of Forensic Sci. doi:10.1080/00450618.2011.630412. ↗
  24. Nolan, Francis and Eva Liina Asu. (2009). The Pairwise Variability Index and Coexisting Rhythms in Language. Phonetica 66 (1-2),pp. 64-77. find
  25. O’Rourk, E. (2008). Speech rhythm variation in dialects of Spanish: Applying the Pairewise Variability Index and variation Coefficients to Peruvian Spanish. Speech Prosody, 6-9 May, Brazil. find
  26. Ramus, F., Nespor, M., Mehler, J. (1999). Correlates of linguistic rhythm in the speech signal. Cognition 72,1-28. find
  27. Rose, P. (2002) Forensic Speaker Identification. New York: Taylor & Francis. https://doi.org/10.1201/9780203166369 ↗
  28. Watson, JCE (2011). Word stress in Arabic. In: The Blackwell companion to phonology. Wiley-Blackwell, Oxford, 2990-3019 (p. 2991) find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/ibrahim-2020-arabic-speech-rhythm-corpus-read/