Hearing voices at the National Library -- a speech corpus and acoustic model for the Swedish language

TL;DR

This paper presents new Swedish acoustic models for speech recognition, leveraging large, balanced speech corpora and wav2vec 2.0, resulting in a superior model called VoxRex that benefits cultural heritage institutions.

Contribution

Developed VoxRex, a Swedish speech recognition model trained on diverse, large-scale corpora, and demonstrated its effectiveness with improved performance over existing models.

Findings

VoxRex outperforms existing Swedish ASR models.

Combining VoxRex with pretrained language models enhances accuracy.

Models are openly available for research and cultural heritage applications.

Abstract

This paper explains our work in developing new acoustic models for automated speech recognition (ASR) at KBLab, the infrastructure for data-driven research at the National Library of Sweden (KB). We evaluate different approaches for a viable speech-to-text pipeline for audiovisual resources in Swedish, using the wav2vec 2.0 architecture in combination with speech corpuses created from KB's collections. These approaches include pretraining an acoustic model for Swedish from the ground up, and fine-tuning existing monolingual and multilingual models. The collections-based corpuses we use have been sampled from millions of hours of speech, with a conscious attempt to balance regional dialects to produce a more representative, and thus more democratic, model. The acoustic model this enabled, "VoxRex", outperforms existing models for Swedish ASR. We also evaluate combining this model with various pretrained language models, which further enhanced performance. We conclude by highlighting the potential of such technology for cultural heritage institutions with vast collections of previously unlabelled audiovisual data. Our models are released for further exploration and research here: https://huggingface.co/KBLab.