Augmenting Librispeech with French Translations: A Multimodal Corpus for Direct Speech Translation Evaluation

TL;DR

This paper introduces a new 236-hour parallel speech translation corpus by aligning LibriSpeech with French translations, facilitating end-to-end speech translation research.

Contribution

It creates and releases a large, open-source parallel speech-text corpus for French-English translation based on LibriSpeech and French e-books, filling a significant resource gap.

Findings

Automatic alignment scores correlate with human judgments.

The corpus is suitable for direct speech translation experiments.

Manual evaluation confirms alignment quality.

Abstract

Recent works in spoken language translation (SLT) have attempted to build end-to-end speech-to-text translation without using source language transcription during learning or decoding. However, while large quantities of parallel texts (such as Europarl, OpenSubtitles) are available for training machine translation systems, there are no large (100h) and open source parallel corpora that include speech in a source language aligned to text in a target language. This paper tries to fill this gap by augmenting an existing (monolingual) corpus: LibriSpeech. This corpus, used for automatic speech recognition, is derived from read audiobooks from the LibriVox project, and has been carefully segmented and aligned. After gathering French e-books corresponding to the English audio-books from LibriSpeech, we align speech segments at the sentence level with their respective translations and obtain 236h of usable parallel data. This paper presents the details of the processing as well as a manual evaluation conducted on a small subset of the corpus. This evaluation shows that the automatic alignments scores are reasonably correlated with the human judgments of the bilingual alignment quality. We believe that this corpus (which is made available online) is useful for replicable experiments in direct speech translation or more general spoken language translation experiments.