Hugging Face's logo

Datasets:
parler-tts
/
libritts_r_filtered

Tasks:
Text-to-Speech
Automatic Speech Recognition
Modalities:
Audio
Text
Formats:
parquet
Languages:
English
Size:
100K - 1M
ArXiv:
Libraries:
Datasets
Dask
License:
Dataset card Viewer Files Community
libritts_r_filtered / README.md
ylacombe's picture
ylacombe HF staff
Update README.md
8980d65 verified
preview code
|
raw
history blame contribute delete
7.08 kB
metadata
license: cc-by-4.0
task_categories:
  - text-to-speech
  - automatic-speech-recognition
language:
  - en
size_categories:
  - 10K<n<100K
dataset_info:
  - config_name: clean
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 24000
      - name: text_normalized
        dtype: string
      - name: text_original
        dtype: string
      - name: speaker_id
        dtype: string
      - name: path
        dtype: string
      - name: chapter_id
        dtype: string
      - name: id
        dtype: string
    splits:
      - name: dev.clean
        num_bytes: 1506311977.8882804
        num_examples: 5589
      - name: test.clean
        num_bytes: 1432099582.6705585
        num_examples: 4689
      - name: train.clean.100
        num_bytes: 8985618654.720787
        num_examples: 32215
      - name: train.clean.360
        num_bytes: 31794257100.91056
        num_examples: 112326
    download_size: 44461321972
    dataset_size: 43718287316.190186
  - config_name: other
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 24000
      - name: text_normalized
        dtype: string
      - name: text_original
        dtype: string
      - name: speaker_id
        dtype: string
      - name: path
        dtype: string
      - name: chapter_id
        dtype: string
      - name: id
        dtype: string
    splits:
      - name: dev.other
        num_bytes: 1042714063.4789225
        num_examples: 4342
      - name: test.other
        num_bytes: 1061489621.2561874
        num_examples: 4716
      - name: train.other.500
        num_bytes: 50718457351.73659
        num_examples: 194626
    download_size: 54153699917
    dataset_size: 52822661036.471695
configs:
  - config_name: clean
    data_files:
      - split: dev.clean
        path: clean/dev.clean-*
      - split: test.clean
        path: clean/test.clean-*
      - split: train.clean.100
        path: clean/train.clean.100-*
      - split: train.clean.360
        path: clean/train.clean.360-*
  - config_name: other
    data_files:
      - split: dev.other
        path: other/dev.other-*
      - split: test.other
        path: other/test.other-*
      - split: train.other.500
        path: other/train.other.500-*
pretty_name: Filtered LibriTTS-R

Dataset Card for Filtered LibriTTS-R

This is a filtered version of LibriTTS-R. It has been filtered based on two sources:

  1. LibriTTS-R paper [1], which lists samples for which speech restoration have failed
  2. LibriTTS-P [2] list of excluded speakers for which multiple speakers have been detected.

LibriTTS-R [1] is a sound quality improved version of the LibriTTS corpus which is a multi-speaker English corpus of approximately 585 hours of read English speech at 24kHz sampling rate, published in 2019.

Usage

Example

Loading the clean config with only the train.clean.360 split.

from datasets import load_dataset

load_dataset("blabble-io/libritts_r", "clean", split="train.clean.100")

Streaming is also supported.

from datasets import load_dataset

load_dataset("blabble-io/libritts_r", streaming=True)

Splits

There are 7 splits (dots replace dashes from the original dataset, to comply with hf naming requirements):

  • dev.clean
  • dev.other
  • test.clean
  • test.other
  • train.clean.100
  • train.clean.360
  • train.other.500

Configurations

There are 3 configurations, each which limits the splits the load_dataset() function will download.

The default configuration is "all".

  • "dev": only the "dev.clean" split (good for testing the dataset quickly)
  • "clean": contains only "clean" splits
  • "other": contains only "other" splits
  • "all": contains only "all" splits

Columns 

{
    "audio": datasets.Audio(sampling_rate=24_000),
    "text_normalized": datasets.Value("string"),
    "text_original": datasets.Value("string"),
    "speaker_id": datasets.Value("string"),
    "path": datasets.Value("string"),
    "chapter_id": datasets.Value("string"),
    "id": datasets.Value("string"),
}

Example Row 

{
  'audio': {
    'path': '/home/user/.cache/huggingface/datasets/downloads/extracted/5551a515e85b9e463062524539c2e1cb52ba32affe128dffd866db0205248bdd/LibriTTS_R/dev-clean/3081/166546/3081_166546_000028_000002.wav', 
    'array': ..., 
    'sampling_rate': 24000
  }, 
  'text_normalized': 'How quickly he disappeared!"',
  'text_original': 'How quickly he disappeared!"',
  'speaker_id': '3081', 
  'path': '/home/user/.cache/huggingface/datasets/downloads/extracted/5551a515e85b9e463062524539c2e1cb52ba32affe128dffd866db0205248bdd/LibriTTS_R/dev-clean/3081/166546/3081_166546_000028_000002.wav', 
  'chapter_id': '166546', 
  'id': '3081_166546_000028_000002'
}

Dataset Details

Dataset Description

  • License: CC BY 4.0

Dataset Sources [optional]

Citation 

@ARTICLE{Koizumi2023-hs,
  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",
  abstract      = "This paper introduces a new speech dataset called
                   ``LibriTTS-R'' designed for text-to-speech (TTS) use. It is
                   derived by applying speech restoration to the LibriTTS
                   corpus, which consists of 585 hours of speech data at 24 kHz
                   sampling rate from 2,456 speakers and the corresponding
                   texts. The constituent samples of LibriTTS-R are identical
                   to those of LibriTTS, with only the sound quality improved.
                   Experimental results show that the LibriTTS-R ground-truth
                   samples showed significantly improved sound quality compared
                   to those in LibriTTS. In addition, neural end-to-end TTS
                   trained with LibriTTS-R achieved speech naturalness on par
                   with that of the ground-truth samples. The corpus is freely
                   available for download from
                   \textbackslashurl\{http://www.openslr.org/141/\}.",
  month         =  may,
  year          =  2023,
  copyright     = "http://creativecommons.org/licenses/by-nc-nd/4.0/",
  archivePrefix = "arXiv",
  primaryClass  = "eess.AS",
  eprint        = "2305.18802"
}

@misc{kawamura2024librittspcorpusspeakingstyle,
      title={LibriTTS-P: A Corpus with Speaking Style and Speaker Identity Prompts for Text-to-Speech and Style Captioning}, 
      author={Masaya Kawamura and Ryuichi Yamamoto and Yuma Shirahata and Takuya Hasumi and Kentaro Tachibana},
      year={2024},
      eprint={2406.07969},
      archivePrefix={arXiv},
      primaryClass={eess.AS},
      url={https://arxiv.org/abs/2406.07969}, 
}