Datasets:

twangodev
/

radiotalk-voices-2k

license: cc-by-4.0
task_categories:
  - text-to-speech
language:
  - en
source_datasets:
  - mythicinfinity/libritts_r
tags:
  - voice-cloning
  - tts
  - reference-voices
  - libritts
size_categories:
  - 1K<n<10K
pretty_name: radiotalk-voices-2k
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/shard-*.parquet

radiotalk-voices-2k

2,000 English reference voices — one 12–30s clip per speaker, selected as the longest qualifying utterance per speaker from LibriTTS-R. Built for zero-shot TTS voice cloning in the radiotalk pipeline.

Stats

2,000 voices · 12.03 hours total
Duration: min 12.0s · median 21.8s · mean 21.7s · max 30.0s
24 kHz, mono, FLAC-encoded

Schema

Column	Type	Description
`voice_id`	string	Stable 12-hex-char id, derived from `(source, source_speaker_id)`
`audio`	Audio	FLAC bytes at 24 kHz (Hub viewer plays inline)
`text`	string	Normalized transcript from LibriTTS-R (`text_normalized`)
`source_clip_id`	string	Utterance id in LibriTTS-R, format `{speaker}_{chapter}_{section}_{utterance}`

Loading

from datasets import load_dataset

ds = load_dataset("twangodev/radiotalk-voices-2k", split="train")
row = ds[0]
print(row["voice_id"], row["text"][:60])
audio = row["audio"]  # {"array": np.ndarray, "sampling_rate": 24000, "path": ...}

Citation

If you use this dataset, please cite both this repository and the upstream LibriTTS-R corpus.

@misc{ding2026radiotalk,
  author       = {Ding, James},
  title        = {{radiotalk}: Data pipeline and reference voice pool for TTS},
  year         = {2026},
  howpublished = {\url{https://github.com/twangodev/radiotalk}},
}

@inproceedings{koizumi2023libritts,
  title     = {{LibriTTS-R}: A Restored Multi-Speaker Text-to-Speech Corpus},
  author    = {Koizumi, Yuma and Zen, Heiga and Karita, Shigeki and Ding, Yifan and Yatabe, Kohei and Morioka, Nobuyuki and Bacchiani, Michiel and Zhang, Yu and Han, Wei and Bapna, Ankur},
  booktitle = {Proc. Interspeech},
  year      = {2023},
}

License

All clips are derived from LibriTTS-R, licensed CC BY 4.0. This dataset inherits that license. See NOTICE.md for the full source attribution.