Recent progress in speech processing has highlighted that high-quality performance across languages requires substantial training data for each individual language. While existing multilingual datasets cover many languages, they often contain insufficient data for most languages. Thus, trained models perform poorly on the majority of the supported languages. Our work addresses this challenge by introducing a scalable pipeline for constructing speech datasets from parliamentary recordings. The proposed pipeline includes robust components for media retrieval and a two-stage alignment algorithm designed to handle non-verbatim transcripts and long-form audio. Applying this pipeline to recordings from 22 European parliaments, we extract over 61k hours of aligned speech segments, achieving substantial per-language coverage with 19 languages exceeding 1k hours and 22 languages exceeding 500 hours of high-quality speech data. We obtain an average 41.8\% reduction in word error rates over baselines when finetuning an existing ASR model on our dataset, demonstrating the usefulness of our approach.
@article{arxiv.2510.00514,
title = {EuroSpeech: A Multilingual Speech Corpus},
author = {Samuel Pfisterer and Florian Grötschla and Luca A. Lanzendörfer and Florian Yan and Roger Wattenhofer},
journal= {arXiv preprint arXiv:2510.00514},
year = {2025}
}
Comments
Published in the 39th Conference on Neural Information Processing Systems (NeurIPS 2025) Track on Datasets and Benchmark