Datasets:

kyutai
/

Audio-NTREX-4L

license: cc
configs:
  - config_name: default
    data_files:
      - split: valid
        path: data/valid-*
      - split: test
        path: data/test-*
dataset_info:
  features:
    - name: id
      dtype: string
    - name: source_language
      dtype: string
    - name: target_language
      dtype: string
    - name: source_ntrex_file
      dtype: string
    - name: target_ntrex_file
      dtype: string
    - name: ntrex_lines
      list: int32
    - name: tts
      dtype: string
    - name: source_audio
      dtype: audio
    - name: source_text
      dtype: string
    - name: source_aligned_transcript
      struct:
        - name: text
          list: string
        - name: timestamp
          list:
            list: float64
    - name: target_text
      dtype: string
  splits:
    - name: valid
      num_bytes: 7218314006
      num_examples: 1800
    - name: test
      num_bytes: 7170348264
      num_examples: 1800
  download_size: 14054511442
  dataset_size: 14388662270
task_categories:
  - translation
language:
  - fr
  - es
  - pt
  - de
  - en
pretty_name: Audio-NTREX-4L
size_categories:
  - 1K<n<10K

Audio-NTREX-4L

logo

Dataset Description

Audio-NTREX-4L is a long-form multilingual speech translation dataset from 🇫🇷 French, 🇪🇸 Spanish, 🇵🇹 Portuguese and 🇩🇪 German to 🇬🇧 English designed to evaluate speech translation models on multi-sentence utterances. It is built from the text translation dataset NTREX by aggregating multiple sentences from a same context to create new source texts and their reference translation. We then use 3 different state-of-the-art commercial Text-To-Speech systems from ElevenLabs, Cartesia and Gradium to synthesize the source texts into speech. We condition audio generations using voices from the multilingual CML-TTS dataset.

Dataset Summary

Original data: NTREX
Source modalities: Audio, Text
Target modality: Text
Source languages: French, Spanish, Portuguese, German
Target language: English
Total number of source/target pairs: 3600
Number of unique source texts per language: 300
Average source sample duration: 45 seconds

Dataset Construction

We use the following files containing text translation data from the NTREX-128 corpus:

🇬🇧 English: newstest2019-ref.eng-US.txt
🇫🇷 French: newstest2019-ref.fra.txt
🇪🇸 Spanish: newstest2019-ref.spa.txt
🇵🇹 Portuguese: newstest2019-ref.por.txt
🇩🇪 German: newstest2019-ref.deu.txt

Using the English file, we select 300 groups of consecutive lines belonging to a same original document to form our multi-sentences source texts and obtain the target text translations accordingly. We define an id for each source-target pair as a hash of the ordered NTREX line indexes it comes from.

We clean the source and target texts by removing elements in parentheses to make them better suited for natural speech. Each source text is then synthesized into 3 audio versions, each using a different TTS system and a different voice conditioning.

We transcribe the synthesized audio using the openai/whisper-large-v3 Speech-To-Text model and check Word Error Rate with respect to the source texts to ensure that audio versions were correctly synthesized.

We split the 3600 source/target pairs into balanced valid and test sets such that all pairs with the same target text stay in the same set i.e. we keep 150 different id for each language in each set.

Citations

If you use this dataset, please cite:

@unpublished{hibikizero2026,
  title={Simultaneous Speech-to-Speech Translation Without Aligned Data},
  author={Tom Labiausse and Romain Fabre and Yannick Estève and Alexandre Défossez and Neil Zeghidour},
  note={Preprint},
  year={2026},
  url={https://arxiv.org/abs/2602.11072v1}
}

License: CC BY-NC-SA 4.0