Elamur Research⌕
☰
home / papers / 2026 / Developing Zila: A Spoken Language Resource for the Endangered Slovenian Gail Valley Dialect

Developing Zila: A Spoken Language Resource for the Endangered Slovenian Gail Valley Dialect

Summary

Slovenian, a less-resourced South Slavic language, has speech resources that almost exclusively cover the standard variety. Most of its roughly 51 dialects and subdialects remain undocumented. The authors address this gap for the Gail Valley dialect, an archaic and severely endangered variety used by the Slovenian minority in the Austrian Carinthian area surrounding Feistritz an der Gail (Slovenian: Bistrica na Zilji).

They describe building Zila, a corpus targeting 100 hours of dialect speech with manual transcriptions. The local community and Carinthian cultural organisations were closely involved in collecting the recordings. The dialect diverges sharply from standard Slovenian through strong diphthongization, "akanje", palatalisations and various consonant shifts. To capture these features, the team wrote dedicated transcription guidelines. As a first diagnostic, they applied a recognizer built for standard Slovenian to the dialect material to estimate how far the two varieties differ.

Zila will be freely released through CLARIN.SI. It is meant to serve as a foundation for dialect-specific ASR and related tools, such as dialect learning or translation into the standard language. It also contributes to preserving an endangered linguistic heritage.

Key points

  • Resource: Zila, a speech database of 100 hours of manually transcribed Gail Valley dialect speech, planned for free distribution via the CLARIN.SI repository.
  • Endangered community: Slovenian speakers make up only about 5–10% of the population in the relevant municipalities. Fewer than 1,000 people are estimated to retain a well-preserved form of the dialect.
  • Transcription: A new, dialect-specific rule set covers features absent from the standard language, such as l > v before back vowels, k > č before front vowels and h > š.
  • Context: Zila is the first Slovenian corpus with a substantial amount of dialect speech. Earlier resources, including the ~1,000-hour Artur database from the RSDO project, cover mostly standard speech.
  • Funding and setting: The work is part of the LINGUA project (EU Interreg Slovenia–Austria 2021–2027), with partners including the Hermagoras Association in Klagenfurt.

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 10.63317/3fhhk948chhm

@inproceedings{zgank2026developing,
  title = {{Developing Zila: A Spoken Language Resource for the Endangered Slovenian Gail Valley Dialect}},
  author = {Zgank, Andrej and Donaj, Gregor and Kolaric, Urh and Sereinig, Usi and Koren-Zwitter, Tatjana and Boto, Sanja and Zwitter-Grilc, Sabina and Vidinic, Jasna and Verdonik, Darinka},
  booktitle = {Proceedings of the Fifteenth Language Resources and Evaluation Conference},
  year = {2026},
  doi = {10.63317/3fhhk948chhm},
  url = {https://doi.org/10.63317/3fhhk948chhm}
}

References

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

  1. Austrian Census 2001 statistics (2025). Available: https://www.statistik.at/en/statistics/population-and-society/population/population-stock/historic-censuses ↗
  2. Barras, C., Geoffrois, E., Wu, Z., & Liberman, M. (2001). Transcriber: development and use of a tool for assisting speech corpora production. Speech Communication, 33(1-2), 5-22. find
  3. Bartoloth M. (2021). Mijalca majalca: Zilska basən / Ziljska basen / Ein Gailtaler Märchen. Klagenfurt, Vienna and Ljubljana: Mohorjeva 1 Slovenian CLARIN.SI repository: https://www.clarin.si/repository/xmlui/ ↗
  4. Bhangale, K. B., & Mohanaprasad, K. (2021). A review on speech processing using machine learning paradigm. International Journal of Speech Technology, 24(2), 367-388. find
  5. Dobrišek, S., Gros, J., Mihelič, F. and Pavešić, N. (1998). Recording and labelling of the GOPOLIS Slovenian speech database. First International Conference on language resources & evaluation: Granada, Spain, 28-30 May 1998 (1089—1096). European Language Resources Association. find
  6. Dobrišek, S., Žganec Gros, J., Žibert, J., Mihelič, F. and Pavešić, N. (2017). Speech Database of Spoken Flight Information Enquiries SOFES 1.0. find
  7. Kaiser, J. and Kačič, Z. (1997). SpeechDat (II) Slovenian Database for the Fixed Telephone Network. Maribor, Slovenia: University of Maribor. find
  8. Slo-Bench Automatic Speech Recognition leaderboard (2025). Available: https://slobench.cjvt.si/leaderboard/view/10. ↗
  9. Škofic, J. et al., (2011). Slovenski lingvistični atlas. 1, Človek (telo, bolezni, družina): 2. Komentarji (Vol. 1). Založba ZRC. find
  10. Verdonik, D., Kosem, I., Vitez, A., Krek, S. and Stabej, M. (2013). Compilation, transcription and usage of a reference speech corpus: The case of the Slovene corpus GOS. Language resources and evaluation, 47(4), 1031—1048. find
  11. Verdonik., D. (2018). Korpus in baza Gos Videolectures. In Zbornik 11. konference Jezikovne tehnologije. find
  12. Verdonik, D., Bizjak, A., Žgank, A., Maučec, M. S., Trojar, M., Gros, J. Ž., ... & Dobrišek, S. (2025). Strategies for managing time and costs in speech corpus creation: insights from the Slovenian ARTUR corpus. Language Resources and Evaluation, 59(3), 1899-1924. find
  13. Yao, Z., Di Wu, Wang, X., Zhang, B., Yu, F., Yang, C., ... & Lei, X. (2021). find
  14. WeNet: Production Oriented Streaming and Non-Streaming End-to-End Speech Recognition Toolkit. In Interspeech (Vol. 2021, pp. 4054-4058). find
  15. Žgank, A., Kačič, Z. and Horvat, B. (2002). Preliminary evaluation of Slovenian mobile database PoliDat. Proceedings of the Third International Conference on Language Resources and Evaluation (LREC’02). find
  16. Žgank, A., Rotovnik, T., Sepesy Maučec, M., Verdonik, D., Kitak, J., Vlaj, D., Hozjan, V., Kačič, Z. and Horvat, B. (2004). Acquisition and annotation of Slovenian broadcast news database. Fourth international conference on language resources and evaluation, LREC 2004, Lisbon, Portugal. find
  17. Žgank, A., Rotovnik, T., Grašič, M., Kos, M., Vlaj, D., and Kačič, Z. (2006). Sloparl-Slovenian parliamentary speech and text corpus for large vocabulary continuous speech recognition. Ninth International Conference on Spoken Language Processing. find
  18. Žgank, A., Sepesy Maučec, M. and Verdonik, D. (2016). The SI TEDx-UM speech database: A new Slovenian spoken language resource. Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC'16) (4670—4673). find
  19. Žibert, J., Martinčić-Ipšić, S., Ipšić, I. and Mihelič, F. (2003). Bilingual speech recognition of Slovenian and Croatian weather forecasts. Proceedings EC-VIP-MC 2003. 4th EURASIP Conference focused on Video/Image Processing and Multimedia Communications (IEEE Cat. No. 03EX667) (Vol. 2, pp. 637— 64… find
  20. Žibert, J. and Mihelič, F. (2004). Development of Slovenian broadcast news speech database. Proceedings of the Fourth International Conference on Language Resources and Evaluation (LREC’04). 10. find
  21. Language Resource References Kaiser, Janez; et al., 2014, Slovenian SpeechDat(II) FDB-1000, ISLRN: 009-566-242-903-2, http://catalog.elra.info/product_info.php?produ cts_id=188 ↗
  22. Verdonik, Darinka; et al., 2021, Spoken corpus Gos VideoLectures 4.2 (transcription), Slovenian language resource repository CLARIN.SI, ISSN 2820-4042, http://hdl.handle.net/11356/1444. ↗
  23. Verdonik, Darinka; et al., 2023a, ASR database ARTUR 1.0 (audio), Slovenian language resource repository CLARIN.SI, ISSN 2820-4042, http://hdl.handle.net/11356/1776. ↗
  24. Verdonik, Darinka; et al., 2023b, ASR database ARTUR 1.0 (transcriptions), Slovenian language resource repository CLARIN.SI, ISSN 2820-4042, http://hdl.handle.net/11356/1772. ↗
  25. Verdonik, Darinka; et al., 2023c, Spoken corpus Gos 2.1 (transcriptions), Slovenian language resource repository CLARIN.SI, ISSN 2820-4042, http://hdl.handle.net/11356/1863. ↗
  26. Žgank, Andrej; et al., 2014, Slovenian BNSI Broadcast News Speech Corpus, ISLRN: 502-280-144-938-4, http://catalog.elra.info/product_info.php?produ cts_id=1064 ↗

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zgank-2026-developing-zila-spoken-language-resource/