Automatic Speech Recognition Quality Prediction based on Large Language Models
Main Article Content
Abstract
Despite the rapid development in the field of automatic speech recognition (ASR) systems, the recognition quality remains poor for audio recordings with acoustic degradation (music, crowd shouts, sounds of machinery, etc.). This becomes especially important when implementing models in critical areas that are characterized by increased attention to control and reliability of the results (aviation, medicine, autopilots, etc.). Error in such areas is very expensive and the use of models becomes impossible, even if the error occurs rarely. To reduce risks, it is advisable to evaluate the expected recognition quality in advance.
This article proposes an approach to predicting the Word Error Rate (WER) based on the acoustic characteristics of the signal and calculating the perplexity of language models. The proposed method involves the creation of diverse sets of audio data by applying various types of acoustic observations to pure speech samples at various levels of quality and intelligibility. Unlike previous studies, a complete set of speech features is extracted and analyzed: prediction of the value of the signal-to-noise ratio (SNR), neural network sound quality metrics (NISQA, etc.) as well as the perplexity of the text of the ASR hypothesis using the language model as an additional feature for training a unified model.
Experiments are being conducted using modern speech recognition architectures to demonstrate the effectiveness of the proposed method in predicting WER in various acoustic conditions. It is shown that the inclusion of perplexity significantly improves the quality of the WER prediction, in particular for data where acoustic features are weakly correlated with recognition errors. The results are applicable for automatic evaluation of the expected quality of speech recognition and filtering of audio inputs.
Article Details
References
2. Nishida T. et al. Description and Discussion on DCASE 2025 Challenge Task 2: First-shot Unsupervised Anomalous Sound Detection for Machine Condition Monitoring // arXiv preprint arXiv:2506.10097. 2025. https://doi.org/10.48550/arXiv.2506.10097
3. Valladares-Poncela A., Fraga-Lamas P., Fernández-Caramés T.M. On-Device Automatic Speech Recognition for IIoT and Extended Reality Industrial Metaverse Applications // Engineering Proceedings. 2024. Vol. 82. Article 3. https://doi.org/10.3390/ecsa-11-20466
4. Polevoj A.V., Koruhova Yu.S. Ob odnom podhode k formal'noj verifikacii nejrosetevyh modelej // Nauchnyj servis v seti Internet: trudy XXVI Vserossijskoj nauchnoj konferencii (23–25 sentyabrya 2024 g., onlajn). M.: IPM im. M.V. Keldysha, 2024. S. 223–235. https://doi.org/10.20948/abrau-2024-11
5. Jakhar N., Srivastava S., Baby A. A Unified Approach to Multilingual Automatic Speech Recognition with Improved Language Identification for Indic Languages // Proc. Interspeech. 2024. С. 3949–3953.
6. Shah M.A., Noguero D.S., Heikkila M.A., Raj B., Kourtellis N. Speech robust bench: A robustness benchmark for speech recognition // arXiv preprint arXiv:2403.07937. 2024. https://doi.org/10.48550/arXiv.2403.07937
7. Baevski A., Zhou Y., Mohamed A., Auli M. wav2vec 2.0: A framework for self-supervised learning of speech representations // Advances in Neural Information Processing Systems. 2020. Vol. 33. P. 12449–12460.
8. Hsu W.-N., Bolte B., Tsai Y.-H.H., Lakhotia K., Salakhutdinov R., Mohamed A. HuBERT: Self-supervised speech representation learning by masked prediction of hidden units // IEEE/ACM Transactions on Audio, Speech, and Language Processing. 2021. Vol. 29. P. 3451–3460.
9. Tsilfidis A., Mporas I., Mourjopoulos J., Fakotakis N. Automatic speech recognition performance in different room acoustic environments with and without dereverberation preprocessing // Computer Speech & Language. 2013. Vol. 27. No. 1. P. 380–395.
10. Loweimi E., Ahadi S.M., Drugman T., Loveymi S. On the importance of pre-emphasis and window shape in phase-based speech recognition // Advances in Nonlinear Speech Processing: 6th International Conference, NOLISP 2013. 2013. P. 160–167.
11. Yin S., Liu C., Zhang Z., Lin Y., Wang D., Tejedor J., Zheng T.F., Li Y. Noisy training for deep neural networks in speech recognition // EURASIP Journal on Audio, Speech, and Music Processing. 2015. Vol. 2015. P. 1–14.
12. Kim C., Stern R.M. Power-normalized cepstral coefficients (PNCC) for robust speech recognition // IEEE/ACM Transactions on Audio, Speech, and Language Processing. 2016. Vol. 24. No. 7. P. 1315–1329.
13. Li J., Deng L., Gong Y., Haeb-Umbach R. An overview of noise-robust automatic speech recognition // IEEE/ACM Transactions on Audio, Speech, and Language Processing. 2014. Vol. 22. No. 4. P. 745–777.
14. Stern R.M., Morgan N. Hearing is believing: Biologically inspired methods for robust automatic speech recognition // IEEE Signal Processing Magazine. 2012. Vol. 29. No. 6. P. 34–43.
15. Radford A., Kim J.W., Xu T., Brockman G., McLeavey C., Sutskever I. Robust speech recognition via large-scale weak supervision // International Conference on Machine Learning. 2023. P. 28492–28518.
16. Bouchakour L., Debyeche M. Noise-robust speech recognition in mobile network based on convolution neural networks // International Journal of Speech Technology. 2022. Vol. 25. P. 1–9.
17. Duarte J., Colcher S. Noise-Robust Automatic Speech Recognition: A Case Study for Communication Interference // Journal on Interactive Systems. 2024. Vol. 15. P. 670–681.
18. Chen G., O’Shaughnessy D., Tolba H. A performance investigation of noisy voice recognition over IP telephony networks // Proc. Interspeech. 2005. P. 2681–2684.
19. Ali A., Renals S. Word error rate estimation for speech recognition: e-WER // Proc. of the 56th Annual Meeting of the Association for Computational Linguistics. 2018. Vol. 2. P. 20–24. https://doi.org/10.18653/v1/P18-2004
20. Ali A., Renals S. Word error rate estimation without ASR output: e-WER2 // arXiv preprint arXiv:2008.03403. 2020.
21. Chowdhury S.A., Ali A. Multilingual word error rate estimation: e-WER3 // ICASSP 2023–2023 IEEE International Conference on Acoustics, Speech and Signal Processing. 2023. P. 1–5.
22. Litman D., Hirschberg J., Swerts M. Predicting automatic speech recognition performance using prosodic cues // 1st Meeting of the North American Chapter of the Association for Computational Linguistics. 2000. P. 1–8.
23. Gallardo L.F., Möller S., Beerends J. Predicting automatic speech recognition performance over communication channels from instrumental speech quality and intelligibility scores // Proc. Interspeech 2017. 2017. P. 2939–2943.
24. Beerends J.G., Schmidhuber C., Berger J., Obermann M., Ullmann R., Pomy J., Kevhl M. Perceptual Objective Listening Quality Assessment (POLQA), the Third Generation ITU-T Standard for End-to-End Speech Quality Measurement. Part I: Temporal Alignment // AES: Journal of the Audio Engineering Society. 2013. Vol. 61. No. 6. P. 366–384.
25. Polevoi A., Kragin A., Loukachevitch N. Ground Truth-Free WER Prediction for ASR via Audio Quality and Model Confidence Features // International Conference on Speech and Computer. Cham: Springer Nature Switzerland, 2025. P. 29–44.
26. Mittag G., Naderi B., Chehadi A., Möller S. NISQA: A deep CNN–self-attention model for multidimensional speech quality prediction with crowdsourced datasets // Proc. Interspeech 2021. 2021. P. 1–5. https://doi.org/10.21437/Interspeech.2021-299
27. Kim C., Stern R. Robust signal-to-noise ratio estimation based on waveform amplitude distribution analysis // Proc. Interspeech 2008. 2008. P. 2598–2601. https://doi.org/10.21437/Interspeech.2008-644
28. Subakan C., Ravanelli M., Cornell S., Grondin F. REAL-M: Towards Speech Separation on Real Mixtures // arXiv preprint arXiv:2110.10812. 2021.
29. Radford A., Wu J., Child R., Luan D., Amodei D., Sutskever I. Language models are unsupervised multitask learners // OpenAI Blog. 2019.
30. Yang A. et al. Qwen2 technical report // arXiv preprint arXiv:2407.10671. 2024.
31. ai-sage GigaChat3-10B-A1.8B // Hugging Face. URL: https://huggingface.co/ai-sage/GigaChat3-10B-A1.8B
32. Yandex YandexGPT-5-Lite-8B-instruct // Hugging Face. URL: https://huggingface.co/yandex/YandexGPT-5-Lite-8B-instruct
33. RefalMachine RuadaptQwen3-8B-Hybrid // Hugging Face. URL: https://huggingface.co/RefalMachine/RuadaptQwen3-8B-Hybrid
34. Prokhorenkova L., Gusev G., Vorobev A., Dorogush A.V., Gulin A. CatBoost: unbiased boosting with categorical features // arXiv preprint. 2019.
https://doi.org/10.48550/arXiv.1706.09516
35. Conneau A. et al. FLEURS: Few-shot Learning Evaluation of Universal Representations of Speech // arXiv preprint arXiv:2205.12446. 2022.
36. Panayotov V., Chen G., Povey D., Khudanpur S. Librispeech: An ASR corpus based on public domain audio books // ICASSP 2015 IEEE International Conference on Acoustics, Speech and Signal Processing. 2015. P. 5206–5210.
This work is licensed under a Creative Commons Attribution 4.0 International License.
Presenting an article for publication in the Russian Digital Libraries Journal (RDLJ), the authors automatically give consent to grant a limited license to use the materials of the Kazan (Volga) Federal University (KFU) (of course, only if the article is accepted for publication). This means that KFU has the right to publish an article in the next issue of the journal (on the website or in printed form), as well as to reprint this article in the archives of RDLJ CDs or to include in a particular information system or database, produced by KFU.
All copyrighted materials are placed in RDLJ with the consent of the authors. In the event that any of the authors have objected to its publication of materials on this site, the material can be removed, subject to notification to the Editor in writing.
Documents published in RDLJ are protected by copyright and all rights are reserved by the authors. Authors independently monitor compliance with their rights to reproduce or translate their papers published in the journal. If the material is published in RDLJ, reprinted with permission by another publisher or translated into another language, a reference to the original publication.
By submitting an article for publication in RDLJ, authors should take into account that the publication on the Internet, on the one hand, provide unique opportunities for access to their content, but on the other hand, are a new form of information exchange in the global information society where authors and publishers is not always provided with protection against unauthorized copying or other use of materials protected by copyright.
RDLJ is copyrighted. When using materials from the log must indicate the URL: index.phtml page = elbib / rus / journal?. Any change, addition or editing of the author's text are not allowed. Copying individual fragments of articles from the journal is allowed for distribute, remix, adapt, and build upon article, even commercially, as long as they credit that article for the original creation.
Request for the right to reproduce or use any of the materials published in RDLJ should be addressed to the Editor-in-Chief A.M. Elizarov at the following address: amelizarov@gmail.com.
The publishers of RDLJ is not responsible for the view, set out in the published opinion articles.
We suggest the authors of articles downloaded from this page, sign it and send it to the journal publisher's address by e-mail scan copyright agreements on the transfer of non-exclusive rights to use the work.