Elamur Research⌕
☰
home / papers / 2026 / Is Semi-Automatic Transcription Useful in Corpus Creation? Preliminary Considerations on the KIParla Corpus

Is Semi-Automatic Transcription Useful in Corpus Creation? Preliminary Considerations on the KIParla Corpus

Summary

Transcribing spontaneous conversation by hand is slow and costly. This paper asks whether automatic speech recognition (ASR) could speed up the creation of KIParla, a growing corpus of spoken Italian of about 228 hours and roughly 2M tokens, without lowering quality. KIParla is transcribed in ELAN using Jefferson-inspired conventions and relies mostly on student interns, who turn over often. The authors ran a two-phase experiment with 11 volunteer interns of differing experience. Each intern transcribed the same 10-minute audio extracts twice: once from scratch and once by correcting ASR output. The extracts covered free conversation, semi-structured interviews and interviews with L2 speakers. The outputs were compared using word-level alignment, annotation-based counts (overlaps, prolongations, transcription units, non-verbal behaviour) and statistical modelling. ASR assistance made transcription faster, but it did not reliably improve accuracy. The effects depended on workflow setup, conversation type and transcriber expertise. The study matters for two reasons. It is one of the few same-audio comparisons of manual and ASR-assisted workflows, and it offers a monitoring framework that corpus projects can reuse.

Key points

  • Participants: 11 interns took part, 4 "experts" and 7 "novices." Interns counted as experts if their prior transcription work amounted to 45 minutes of speech or more.
  • Design: Manual and ASR-assisted transcriptions were produced for identical segments. This allowed a direct comparison, which earlier work such as Gorisch and Schmidt (2024) could not make.
  • Data: Extracts came from KIParla modules, including KIPasti (free conversation) and L2 interview material such as StraParlaBO.
  • Modelling: A Generalized Additive Mixed Model (GAMM) was used to estimate the effects of transcriber, expertise, data type and phase. Transcriber identity and conversation type were often significant, while expertise rarely was.
  • Availability: Data and code are released at github.com/KIParla/asr-assisted-transcription. The authors suggest that task-specific fine-tuning of ASR could make it viable in the KIParla workflow.

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

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/3gw6i7cbs5rc

@inproceedings{simonotti2026is,
  title = {{Is Semi-Automatic Transcription Useful in Corpus Creation? Preliminary Considerations on the KIParla Corpus}},
  author = {Simonotti, Martina and Pannitto, Ludovica and Zucchini, Eleonora and Ballarè, Silvia and Mauri, Caterina},
  booktitle = {Proceedings of the Fifteenth Language Resources and Evaluation Conference},
  year = {2026},
  doi = {10.63317/3gw6i7cbs5rc},
  url = {https://doi.org/10.63317/3gw6i7cbs5rc}
}

References

26 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. 2024. ELAN (version 6.9) [computer software]. https://archive.mpi.nl/tla/elan. Retrieved from https://archive.mpi.nl/tla/elan. ↗
  2. Silvia Ballarè, Caterina Mauri, et al. 2020. La creazione del corpus kiparla: criteri metodologici e prospettive future. RID, RIVISTA ITALIANA DI DIALETTOLOGIA, 44:53–69. find
  3. Ronald Cumbal, Birger Moell, Jose Lopes, and Olof Engwall. 2024. You don’t understand me!: Comparing asr results for l1 and l2 speakers of swedish. find
  4. Yashesh Gaur, Walter S. Lasecki, Florian Metze, and Jeffrey P. Bigham. 2016. The effects of automatic speech recognition quality on human transcription latency. In Proceedings of the 13th International Web for All Conference, W4A ’16, New York, NY, USA. Association for Computing Machinery. find
  5. Anne Sophie Ghyselen, Anne Breitbarth, Melissa Farasyn, Jacques Van Keymeulen, and Arjan van Hessen. 2020. Clearing the transcription hurdle in dialect corpus building: The corpus of southern dutch dialects as case study. Frontiers in Artificial Intelligence, 3. find
  6. Jan Gorisch and Thomas Schmidt. 2024. Evaluating workflows for creating orthographic transcripts for oral corpora by transcribing from scratch or correcting ASR-output. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COL… find
  7. Nikolaus P. Himmelmann. 2018. Meeting the transcription challenge. In Bradley McDonnell, Andrea L. Berez-Kroeker, and Gary Holton, editors, Reflections on Language Documentation 20 Years after Himmelmann 1998, volume 15 of Language Documentation & Conservation Special Publication, pages 33–40. Unive… find
  8. Gail Jefferson. 2004. Glossary of transcript symbols with an introduction. In Gene H. Lerner, editor, Pragmatics & Beyond New Series, volume 125, pages 13–31. John Benjamins Publishing Company, Amsterdam. find
  9. Dagmar Jung and Nikolaus P Himmelmann. 2011. Retelling data: Working on transcription. Documenting endangered languages: Achievements and perspectives, pages 201–220. find
  10. Dietrich Klakow and Jochen Peters. 2002. Testing the correlation of word error rate and perplexity. Speech Communication, 38(1):19–28. find
  11. Christian Lehmann. 2004. Data in linguistics. find
  12. Andreas Liesenfeld, Alianda Lopez, and Mark Dingemanse. 2023. The timing bottleneck: Why timing and overlap are mission-critical for conversational user interfaces, speech recognition and dialogue systems. find
  13. Alianda Lopez, Andreas Liesenfeld, and Mark Dingemanse. 2022. Evaluation of automatic speech recognition for conversational speech in Dutch, English and German: What goes missing? In Proceedings of the 18th Conference on Natural Language Processing (KONVENS 2022), pages 135–143, Potsdam, Germany. find
  14. KONVENS 2022 Organizers. find
  15. Caterina Mauri, Silvia Ballarè, Eugenio Goria, Massimo Cerruti, and Francesco Suriano. 2019. Kiparla corpus: A new resource for spoken italian. In Proceedings of the 6th Italian Conference on Computational Linguistics CLiC-it. find
  16. Caterina Mauri, Eugenio Goria, Eleonora Zucchini, and Beatrice Bernasconi. 2025. Il progetto Diver-SIta. documentare la diversità nell’italiano parlato. In Silvia Ballarè and Caterina Mauri, editors, CLUB Working Papers in Linguistics, volume 9, pages 147–156. CLUB – Circolo Linguistico dell’Univers… find
  17. Saul B. Needleman and Christian D. Wunsch. 1970. A general method applicable to the search for similarities in the amino acid sequence of two proteins. Journal of Molecular Biology, 48(3):443– 453. find
  18. Alec Radford, Jong Wook Kim, Tao Xu, Greg Brockman, Christine McLeavey, and Ilya Sutskever. 2023. Robust speech recognition via large-scale weak supervision. In International conference on machine learning, pages 28492–28518. PMLR. find our card
  19. Muhammad Umair, Julia Mertens, Saul Albert, and J. Ruiter. 2022. Gailbot: An automatic transcription system for conversation analysis. Dialogue Discourse, 13:63–95. find
  20. Vincenzo Norman Vitale, Loredana Schettino, and Francesco Cutugno. 2024. Modelling filled particles and prolongation using end-to-end automatic speech recognition systems: A quantitative and qualitative analysis. In Proceedings of the 10th Italian Conference on Computational Linguistics (CLiC-it 202… find
  21. Yu Wang, Leyi Lao, Langchu Huang, Gabriel Skantze, Yang Xu, and Hendrik Buschmeier. 2025. Investigating the representation of backchannels and fillers in fine-tuned language models. find
  22. Hiroyoshi Yamasaki, Jérôme Louradour, Julie Hunter, and Laurent Prévot. 2023. Transcribing and aligning conversational speech: A hybrid pipeline applied to french conversations. find
  23. Vicky Zayats, Trang Tran, Richard Wright, Courtney Mansfield, and Mari Ostendorf. 2019. Disfluencies and human speech transcription errors. In Interspeech 2019, pages 3088–3092. find
  24. Eleonora Zucchini, Silvia Ballarè, and Caterina Mauri. accepted. Parlare italiano in contesto mi-gratorio: il corpus stra-parlabo. Études romanes de Brno. 12. Language Resource References Mauri, Caterina and Ballarè, Silvia and Zucchini, Eleonora. 2024a. KIPasti. distributed via ILC-CNR for CLARIN-I… find
  25. Mauri, Caterina and Ballarè, Silvia and Zucchini, Eleonora. 2024b. ParlaBO. distributed via ILC-CNR for CLARIN-IT repository hosted at Institute for Computational Linguistics “A. Zampolli”, 1.1.0. PID http://hdl.handle.net/20.500.11752/OPEN-2126. ↗
  26. Zucchini, Eleonora, and Ballarè, Silvia and Mauri, Caterina. 2026. StraParlaBO. Alma Mater Studiorum - Università di Bologna, 0.1.0. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/simonotti-2026-is-semi-automatic-transcription-useful/