--- title: "AusKidTalk: Developing Transcription — Szalay 2026" description: "Transcribing spontaneous child speech is expensive." authors: ["Tuende Szalay", "Zheng Nan", "Renata Huang", "Mostafa Shahin", "Sirojan Tharmakulasingam", "Kirrie Ballard", "Beena Ahmed"] year: 2026 venue: "Proceedings of the Fifteenth Language Resources and Evaluation Conference" source: "https://aclanthology.org/2026.lrec-1.458/" doi: "10.63317/4j6otnjq8c3n" licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)" licence_url: "https://creativecommons.org/licenses/by-nc/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/papers/szalay-2026-auskidtalk-developing-transcription-guidelines-continuous/" --- [Elamur Research](https://md.research.elamur.ai/) · [this page in HTML](https://research.elamur.ai/free/papers/szalay-2026-auskidtalk-developing-transcription-guidelines-continuous/) [home](https://md.research.elamur.ai/) / [papers](https://md.research.elamur.ai/free/papers/) / [2026](https://md.research.elamur.ai/free/papers/2026/) / AusKidTalk: Developing Transcription Guidelines for Continuous Australian English Child Speech # AusKidTalk: Developing Transcription Guidelines for Continuous Australian English Child Speech **Authors:** Tuende Szalay, Zheng Nan, Renata Huang, Mostafa Shahin, Sirojan Tharmakulasingam, Kirrie Ballard, Beena Ahmed · **Year:** 2026 · **Venue:** Proceedings of the Fifteenth Language Resources and Evaluation Conference · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2026.lrec-1.458/) · [PDF](https://aclanthology.org/2026.lrec-1.458.pdf) · [DOI 10.63317/4j6otnjq8c3n](https://doi.org/10.63317/4j6otnjq8c3n) · [OpenAlex](https://openalex.org/W7160081276) **Citations:** 0 citations (openalex-snapshot · 2026-10-14) ## Summary Transcribing spontaneous child speech is expensive. ASR systems trained on adult or American English speech perform poorly on young Australian English (AusE) speakers, so their output has to be checked by people. Yet published corpora seldom document how this manual correction is carried out, which undermines reliability and replicability. The authors set out the transcription workflow for the semi-spontaneous story-telling task in AusKidTalk, the first AusE child speech corpus. Orthographic transcripts were first produced with the project's existing ASR toolkit and then corrected by annotators under a new protocol. The protocol combines written guidelines, corpus-specific training materials, a custom correction interface, and ground truth tasks scored with consistency metrics. Nine annotators corrected narratives from 261 children. Correction took roughly eleven times the audio length, far below figures reported for fully manual work. The authors attribute this to the fairly accurate ASR output (23% WER), and annotators agreed on most ground truth checks. The workflow offers a reusable template for ASR-assisted transcription of spontaneous or under-resourced speech, although stories from younger children may need further correction. ## Key points - AusKidTalk currently holds recordings of 620 AusE-speaking children aged 3–12, covering three scripted and two semi-spontaneous tasks. The story-telling task averages about 3.5 minutes per child. - Initial transcripts came from the AusKidTalk/UNSW ASR toolkit, which reached a 23% word error rate on this task. - Nine annotators submitted corrections for 261 children's stories plus 25 ground truth tasks. Correcting one 3.5-minute story took about 40 minutes, roughly 11× real time. For comparison, earlier literature cites about 50× for manual transcription and 15× for ASR-assisted transcription. - Annotators were consistent on 15 of the 25 ground truth submissions, and most inconsistent results came from a single difficult ground truth task. - The authors argue that commercial cloud ASR systems (IBM-Watson, Google, Amazon) and OCTRA can be unsuitable for this kind of data. These tools require uploading speech to external servers, which conflicts with ethical data-storage constraints. *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](https://md.research.elamur.ai/free/shelves/linguistics/) ## Cite The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{szalay2026auskidtalk, title = {{AusKidTalk: Developing Transcription Guidelines for Continuous Australian English Child Speech}}, author = {Szalay, Tuende and Nan, Zheng and Huang, Renata and Shahin, Mostafa and Tharmakulasingam, Sirojan and Ballard, Kirrie and Ahmed, Beena}, booktitle = {Proceedings of the Fifteenth Language Resources and Evaluation Conference}, year = {2026}, doi = {10.63317/4j6otnjq8c3n}, url = {https://doi.org/10.63317/4j6otnjq8c3n} } ``` - [← previous Augmenting LLM Reasoning with Dynamic Notes Writing for Complex MultiHop QA](https://md.research.elamur.ai/free/papers/maheshwary-2026-augmenting-llm-reasoning-dynamic-notes/) - [next → AuthorMix: Modular Authorship Style Transfer via Layer-wise Adapter Mixing](https://md.research.elamur.ai/free/papers/thillainathan-2026-authormix-modular-authorship-style-transfer/) - [↑ back to papers of 2026](https://md.research.elamur.ai/free/papers/2026/) ## References 36 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Beena Ahmed, Kirrie Ballard, Denis Burnham, Tharmakulasingam Sirojan, Hadi Mehmood, Dominique Estival, Elise Baker, Felicity Cox, Joanne Arciuli, Titia Benders, et al. 2021. AusKidTalk: an auditory-visual corpus of 3-to 12-year-old australian children’s speech. In Annual Conference of the International Speech Communication Association (22nd: 2021), pages 3680–3684. International Speech Communication Association. [find](https://scholar.google.com/scholar?q=Beena+Ahmed%2C+Kirrie+Ballard%2C+Denis+Burnham%2C+Tharmakulasingam+Sirojan%2C+Hadi+Mehmood%2C+Dominique+Estival%2C+Elise+Baker%2C+Felicity+Cox%2C+Joanne+Arciuli%2C+Titia+Benders%2C+et+al%2E+2021%2E+AusKidTalk%3A+an+auditory%2Dvi) 2. Ahmed Adel Attia, Jing Liu, Wei Ai, Dorottya Demszky, and Carol Espy-Wilson. 2024. Kid-whisper: Towards bridging the performance gap in automatic speech recognition for children vs. adults. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society, volume 7, pages 74–80. [find](https://scholar.google.com/scholar?q=Ahmed+Adel+Attia%2C+Jing+Liu%2C+Wei+Ai%2C+Dorottya+Demszky%2C+and+Carol+Espy%2DWilson%2E+2024%2E+Kid%2Dwhisper%3A+Towards+bridging+the+performance+gap+in+automatic+speech+recognition+for+children+vs%2E+adults%2E+In+Proceed) 3. Thierry Bazillon, Yannick Esteve, and Daniel Luzzati. 2008. Manual vs assisted transcription of prepared and spontaneous speech. In LREC. [find](https://scholar.google.com/scholar?q=Thierry+Bazillon%2C+Yannick+Esteve%2C+and+Daniel+Luzzati%2E+2008%2E+Manual+vs+assisted+transcription+of+prepared+and+spontaneous+speech%2E+In+LREC%2E) 4. Denis Burnham, Dominique Estival, Steven Fazio, Jette Viethen, Felicity Cox, Robert Dale, Steve Cassidy, Julien Epps, Roberto Togneri, Michael Wagner, et al. 2011. Building an audio-visual corpus of Australian English: large corpus collection with an economical portable and replicable black box. ISCA. [find](https://scholar.google.com/scholar?q=Denis+Burnham%2C+Dominique+Estival%2C+Steven+Fazio%2C+Jette+Viethen%2C+Felicity+Cox%2C+Robert+Dale%2C+Steve+Cassidy%2C+Julien+Epps%2C+Roberto+Togneri%2C+Michael+Wagner%2C+et+al%2E+2011%2E+Building+an+audio%2Dvisual+corpus+of+A) 5. Strassel S. M. Cieri, C. 2009. Closer still to a robust, all digital, empirical, reproducible sociolinguistic methodology. In NWAV 38: New Ways of Analyzing Variation. University of Ottawa. Doggy Dog Cartoons. Doggy dog. [find](https://scholar.google.com/scholar?q=Strassel+S%2E+M%2E+Cieri%2C+C%2E+2009%2E+Closer+still+to+a+robust%2C+all+digital%2C+empirical%2C+reproducible+sociolinguistic+methodology%2E+In+NWAV+38%3A+New+Ways+of+Analyzing+Variation%2E+University+of+Ottawa%2E+Doggy+Dog+) 6. Christoph Draxler and Julian Pömp. 2022. OCTRA - an innovative approach to orthographic transcription. In INTERSPEECH, pages 5217–5218. [find](https://scholar.google.com/scholar?q=Christoph+Draxler+and+Julian+P%C3%B6mp%2E+2022%2E+OCTRA+%2D+an+innovative+approach+to+orthographic+transcription%2E+In+INTERSPEECH%2C+pages+5217%E2%80%935218%2E) 7. Yannick Esteve, Thierry Bazillon, Jean-Yves Antoine, Frédéric Béchet, and Jérôme Farinas. 2010. The EPAC corpus: Manual and automatic annotations of conversational speech in french broadcast news. In LREC. [find](https://scholar.google.com/scholar?q=Yannick+Esteve%2C+Thierry+Bazillon%2C+Jean%2DYves+Antoine%2C+Fr%C3%A9d%C3%A9ric+B%C3%A9chet%2C+and+J%C3%A9r%C3%B4me+Farinas%2E+2010%2E+The+EPAC+corpus%3A+Manual+and+automatic+annotations+of+conversational+speech+in+french+broadcast+news%2E+In+) 8. Jun Gao, Aijun Li, and Ziyu Xiong. 2012. Mandarin multimedia child speech corpus: Cass_child. In 2012 International Conference on Speech Database and Assessments, pages 7–12. IEEE. [find](https://scholar.google.com/scholar?q=Jun+Gao%2C+Aijun+Li%2C+and+Ziyu+Xiong%2E+2012%2E+Mandarin+multimedia+child+speech+corpus%3A+Cass%5Fchild%2E+In+2012+International+Conference+on+Speech+Database+and+Assessments%2C+pages+7%E2%80%9312%2E+IEEE%2E) 9. John S Garofolo, Lori F Lamel, William M Fisher, Jonathan G Fiscus, and David S Pallett. 1993. DARPA TIMIT acoustic-phonetic continous speech corpus CD-ROM. nist speech disc 1-1.1. NASA STI/Recon technical report n, 93:27403. [find](https://scholar.google.com/scholar?q=John+S+Garofolo%2C+Lori+F+Lamel%2C+William+M+Fisher%2C+Jonathan+G+Fiscus%2C+and+David+S+Pallett%2E+1993%2E+DARPA+TIMIT+acoustic%2Dphonetic+continous+speech+corpus+CD%2DROM%2E+nist+speech+disc+1%2D1%2E1%2E+NASA+STI%2FRecon+tech) 10. Meghan Lammie Glenn, Stephanie M Strassel, Haejoong Lee, Kazuaki Maeda, Ramez Zakhary, and Xuansong Li. 2010. Transcription methods for consistency, volume and efficiency. In LREC. [find](https://scholar.google.com/scholar?q=Meghan+Lammie+Glenn%2C+Stephanie+M+Strassel%2C+Haejoong+Lee%2C+Kazuaki+Maeda%2C+Ramez+Zakhary%2C+and+Xuansong+Li%2E+2010%2E+Transcription+methods+for+consistency%2C+volume+and+efficiency%2E+In+LREC%2E) 11. Simon Gonzalez, James Grama, and Catherine E Travis. 2020. Comparing the performance of forced aligners used in sociophonetic research. Linguistics Vanguard, 6(1):20190058. [↗](https://doi.org/10.1515/lingvan-2019-0058) 12. Jan Gorisch, Michael Gref, and Thomas Schmidt. 2020. Using automatic speech recognition in spoken corpus curation. In Proceedings of the Twelfth Language Resources and Evaluation Conference, pages 6423–6428. [find](https://scholar.google.com/scholar?q=Jan+Gorisch%2C+Michael+Gref%2C+and+Thomas+Schmidt%2E+2020%2E+Using+automatic+speech+recognition+in+spoken+corpus+curation%2E+In+Proceedings+of+the+Twelfth+Language+Resources+and+Evaluation+Conference%2C+pages+642) 13. Nina Grønnum. 2009. A Danish phonetically annotated spontaneous speech corpus (DanPASS). Speech Communication, 51(7):594–603. [↗](https://doi.org/10.1016/j.specom.2008.11.002) 14. Vera Kempe, Patricia J Brooks, and Steven Gillis. 2024. Four decades of open language science: The CHILDES project. Language Teaching Research Quarterly, 44:15–30. [↗](https://doi.org/10.32038/ltrq.2024.44.04) 15. Abigail Anne Kressner, Kirsten Maria Jensen-Rico, Johannes Kizach, Brian Kai Loong Man, Anja Kofoed Pedersen, Lars Bramsløw, Lise Bruun Hansen, Laura Winther Balling, Brent Kirkwood, and Tobias May. 2024. A corpus of audio-visual recordings of linguistically balanced, Danish sentences for speech-in-noise experiments. Speech Communication, 165:103141. [↗](https://doi.org/10.1016/j.specom.2024.103141) 16. Chia-ying Lee and James Glass. 2011. A transcription task for crowdsourcing with automatic quality control. In Twelfth Annual Conference of the International Speech Communication Association. Citeseer. [find](https://scholar.google.com/scholar?q=Chia%2Dying+Lee+and+James+Glass%2E+2011%2E+A+transcription+task+for+crowdsourcing+with+automatic+quality+control%2E+In+Twelfth+Annual+Conference+of+the+International+Speech+Communication+Association%2E+Citeseer) 17. Laura L Lee and Susan M Canter. 1971. Developmental sentence scoring: A clinical procedure for estimating syntactic development in children’s spontaneous speech. Journal of Speech and Hearing Disorders, 36(3):315–340. [↗](https://doi.org/10.1044/jshd.3603.315) 18. Mark Y Liberman. 2019. Corpus phonetics. Annual Review of Linguistics, 5(1):91–107. [find](https://scholar.google.com/scholar?q=Mark+Y+Liberman%2E+2019%2E+Corpus+phonetics%2E+Annual+Review+of+Linguistics%2C+5%281%29%3A91%E2%80%93107%2E) 19. Houjun Liu, Brian MacWhinney, Davida Fromm, and Alyssa Lanzi. 2023. Automation of language sample analysis. Journal of Speech, Language, and Hearing Research, 66(7):2421–2433. [↗](https://doi.org/10.1044/2023_jslhr-22-00642) 20. Daniela Mereu and Alessandro Vietti. 2021. Dialogic ItAlian: the creation of a corpus of Italian spontaneous speech. Speech Communication, 130:1–14. [↗](https://doi.org/10.1016/j.specom.2021.03.002) 21. Vassil Panayotov, Guoguo Chen, Daniel Povey, and Sanjeev Khudanpur. 2015. Librispeech: an ASR corpus based on public domain audio books. In 2015 IEEE international conference on acoustics, speech and signal processing (ICASSP), pages 5206–5210. IEEE. [find](https://scholar.google.com/scholar?q=Vassil+Panayotov%2C+Guoguo+Chen%2C+Daniel+Povey%2C+and+Sanjeev+Khudanpur%2E+2015%2E+Librispeech%3A+an+ASR+corpus+based+on+public+domain+audio+books%2E+In+2015+IEEE+international+conference+on+acoustics%2C+speech+and+) 22. Sérgio Paulo and Luís C Oliveira. 2004. Automatic phonetic alignment and its confidence measures. In Advances in Natural Language Processing: 4th International Conference, EsTAL 2004, Alicante, Spain, October 20-22, 2004. Proceedings 4, pages 36–44. Springer. [find](https://scholar.google.com/scholar?q=S%C3%A9rgio+Paulo+and+Lu%C3%ADs+C+Oliveira%2E+2004%2E+Automatic+phonetic+alignment+and+its+confidence+measures%2E+In+Advances+in+Natural+Language+Processing%3A+4th+International+Conference%2C+EsTAL+2004%2C+Alicante%2C+Spain%2C) 23. Humberto Pérez-Espinosa, Juan Martínez-Miranda, Ismael Espinosa-Curiel, Josefina Rodríguez-Jacobo, Luis Villaseñor-Pineda, and Himer Avila-George. 2020. IESC-child: an interactive emotional children’s speech corpus. Computer Speech & Language, 59:55–74. [↗](https://doi.org/10.1016/j.csl.2019.06.006) 24. Mark A Pitt, Keith Johnson, Elizabeth Hume, Scott Kiesling, and William Raymond. 2005. The Buckeye corpus of conversational speech: Labeling conventions and a test of transcriber reliability. Speech Communication, 45(1):89–95. [↗](https://doi.org/10.1016/j.specom.2004.09.001) 25. Lars Rumberg, Christopher Gebauer, Hanna Ehlert, Maren Wallbaum, Lena Bornholt, Jörn Ostermann, and Ulrike Lüdtke. 2022. kidsTALC: A corpus of 3-to 11-year-old German children’s connected natural speech. In INTERSPEECH, pages 5160–5164. [find](https://scholar.google.com/scholar?q=Lars+Rumberg%2C+Christopher+Gebauer%2C+Hanna+Ehlert%2C+Maren+Wallbaum%2C+Lena+Bornholt%2C+J%C3%B6rn+Ostermann%2C+and+Ulrike+L%C3%BCdtke%2E+2022%2E+kidsTALC%3A+A+corpus+of+3%2Dto+11%2Dyear%2Dold+German+children%E2%80%99s+connected+natural+spee) 26. Barbara Schuppler, Martin Hagmüller, and Alexander Zahrer. 2017. A corpus of read and conversational Austrian German. Speech Communication, 94:62–74. [↗](https://doi.org/10.1016/j.specom.2017.09.003) 27. Mostafa Ali Shahin, Renée Lu, Julien Epps, and Beena Ahmed. 2020. UNSW system description for the shared task on automatic speech recognition for non-native children’s speech. In Proc. Interspeech, pages 265–268. [find](https://scholar.google.com/scholar?q=Mostafa+Ali+Shahin%2C+Ren%C3%A9e+Lu%2C+Julien+Epps%2C+and+Beena+Ahmed%2E+2020%2E+UNSW+system+description+for+the+shared+task+on+automatic+speech+recognition+for+non%2Dnative+children%E2%80%99s+speech%2E+In+Proc%2E+Interspeech%2C+pa) 28. Prashanth Gurunath Shivakumar and Panayiotis Georgiou. 2020. Transfer learning from adult to children for speech recognition: Evaluation, analysis and recommendations. Computer speech & language, 63:101077. [↗](https://doi.org/10.1016/j.csl.2020.101077) 29. Rajni Sobti, Kalpna Guleria, and Virender Kadyan. 2024. Comprehensive literature review on children automatic speech recognition system, acoustic linguistic mismatch approaches and challenges. Multimedia Tools and Applications, pages 1–63. [↗](https://doi.org/10.1007/s11042-024-18753-4) 30. Rosy Southwell, Wayne Ward, Viet Anh Trinh, Charis Clevenger, Clay Clevenger, Emily Watts, Jason Reitman, Sidney D’Mello, and Jacob Whitehill. 2024. Automatic speech recognition tuned for child speech in the classroom. In Icassp 2024-2024 ieee international conference on acoustics, speech and signal processing (icassp), pages 12291–12295. IEEE. [find](https://scholar.google.com/scholar?q=Rosy+Southwell%2C+Wayne+Ward%2C+Viet+Anh+Trinh%2C+Charis+Clevenger%2C+Clay+Clevenger%2C+Emily+Watts%2C+Jason+Reitman%2C+Sidney+D%E2%80%99Mello%2C+and+Jacob+Whitehill%2E+2024%2E+Automatic+speech+recognition+tuned+for+child+speech) 31. Rachele Sprugnoli, Giovanni Moretti, Luisa Bentivogli, and Diego Giuliani. 2017. Creating a ground truth multilingual dataset of news and talk show transcriptions through crowdsourcing. Language Resources and Evaluation, 51:283– 317. [↗](https://doi.org/10.1007/s10579-016-9372-5) 32. Tünde Szalay, Louise Ratko, Mostafa Shahin, Tharmakulasingam Sirojan, Kirrie Ballard, Felicity Cox, and Beena Ahmed. 2022a. A semi-automatic workflow for orthographic transcription of a novel speech corpus: A case study of AusKidTalk. In Proceedings of the 18th Australasian International Conference on Speech Science and Technology, Canberra. Australasian Speech Science and Technology Association. [find](https://scholar.google.com/scholar?q=T%C3%BCnde+Szalay%2C+Louise+Ratko%2C+Mostafa+Shahin%2C+Tharmakulasingam+Sirojan%2C+Kirrie+Ballard%2C+Felicity+Cox%2C+and+Beena+Ahmed%2E+2022a%2E+A+semi%2Dautomatic+workflow+for+orthographic+transcription+of+a+novel+speech+c) 33. Tünde Szalay, Mostafa Shahin, Beena Ahmed, and Kirrie Ballard. 2022b. Knowledge of accent differences can be used to predict speech recognition. Proc. Interspeech 2022, pages 1372–1376. [find](https://scholar.google.com/scholar?q=T%C3%BCnde+Szalay%2C+Mostafa+Shahin%2C+Beena+Ahmed%2C+and+Kirrie+Ballard%2E+2022b%2E+Knowledge+of+accent+differences+can+be+used+to+predict+speech+recognition%2E+Proc%2E+Interspeech+2022%2C+pages+1372%E2%80%931376%2E) 34. Tünde Szalay, Mostafa Shahin, Tharmakulasingam Sirojan, Zheng Nan, Renata Huang, Kirrie Ballard, and Beena Ahmed. 2025. AusKidTalk: Using strategic data collection and out-of-domain tools to semi-automate novel corpora annotation. In Proc. Interspeech, pages 4268–4272. [find](https://scholar.google.com/scholar?q=T%C3%BCnde+Szalay%2C+Mostafa+Shahin%2C+Tharmakulasingam+Sirojan%2C+Zheng+Nan%2C+Renata+Huang%2C+Kirrie+Ballard%2C+and+Beena+Ahmed%2E+2025%2E+AusKidTalk%3A+Using+strategic+data+collection+and+out%2Dof%2Ddomain+tools+to+semi%2Dauto) 35. Rachael Tatman and Conner Kasten. 2017. Effects of talker dialect, gender & race on accuracy of Bing Speech and YouTube automatic captions. In Interspeech, pages 934–938. [find](https://scholar.google.com/scholar?q=Rachael+Tatman+and+Conner+Kasten%2E+2017%2E+Effects+of+talker+dialect%2C+gender+%26+race+on+accuracy+of+Bing+Speech+and+YouTube+automatic+captions%2E+In+Interspeech%2C+pages+934%E2%80%93938%2E) 36. Alicia Beckford Wassink, Cady Gansen, and Isabel Bartholomew. 2022. Uneven success: automatic speech recognition and ethnicity-related dialects. Speech Communication, 140:50–70. [↗](https://doi.org/10.1016/j.specom.2022.03.009)