Dataset Card for librivox_spanish

Dataset Summary

Librivox is a non-commercial, non-profit and ad-free project that is dedicated to make all books in the public domain available, for free, in audio format on the internet. According to this, we downloaded 300 titles in Spanish to create the LIBRIVOX SPANISH CORPUS.

The LIBRIVOX SPANISH CORPUS has a duration of 73 hours and it is constituted by audio files between 3 and 10 seconds long, manually segmented. Transcription are also manually made by Spanish native speakers. The recordings are divided between male/female and native/non-native speakers.

Example Usage

The LIBRIVOX SPANISH CORPUS contains only the train split:

from datasets import load_dataset
librivox_spanish = load_dataset("ciempiess/librivox_spanish")

It is also valid to do:

from datasets import load_dataset
librivox_spanish = load_dataset("ciempiess/librivox_spanish",split="train")

Supported Tasks

automatic-speech-recognition: The dataset can be used to test a model for Automatic Speech Recognition (ASR). The model is presented with an audio file and asked to transcribe the audio file to written text. The most common evaluation metric is the word error rate (WER).

Languages

The language of the corpus is Spanish.

Dataset Structure

Data Instances

{
  'audio_id': 'LBVX_F_69_NNT_0035', 
  'audio': {
    'path': '/home/carlos/.cache/HuggingFace/datasets/downloads/extracted/a506b24788064c4a89c858f829b408b0d2445c9cc30e52087e38ceee60fa03d7/non_native/female/F_69/LBVX_F_69_NNT_0035.flac', 
    'array': array([ 2.4414062e-04, -6.1035156e-05, -2.1362305e-04, ...,
       -6.1035156e-04, -4.8828125e-04, -7.6293945e-04], dtype=float32), 'sampling_rate': 16000
  }, 
  'speaker_id': 'F_69', 
  'speaker_group': 'non_native', 
  'gender': 'female', 
  'duration': 9.975000381469727, 
  'normalized_text': 'del pequeño dormido en la mejilla que con timido afán su madre besa y se refleja alegre en la fajilla'
}

Data Fields

• audio_id (string) - id of audio segment
• audio (datasets.Audio) - a dictionary containing the path to the audio, the decoded audio array, and the sampling rate. In non-streaming mode (default), the path points to the locally extracted audio. In streaming mode, the path is the relative path of an audio inside its archive (as files are not downloaded and extracted locally).
• speaker_id (string) - id of speaker
• speaker_group (string) - native or non native
• gender (string) - gender of speaker (male or female)
• duration (float32) - duration of the audio file in seconds.
• normalized_text (string) - normalized audio segment transcription

Data Splits

The corpus counts just with the train split which has a total of 36338 speech files from 77 female speakers and 77 male speakers with a total duration of 73 hours and 1 minute.

Dataset Creation

Curation Rationale

The LIBRIVOX SPANISH CORPUS (LSC) has the following characteristics:

• The LSC has an exact duration of 73 hours and 1 minute. It has 36338 audio files.

• The LSC counts with 154 different speakers: 77 men and 77 women.

• Every audio file in the LSC has a duration between 3 and 10 seconds approximately.

• Data in LSC is classified by speaker. It means, all the recordings of one single speaker are stored in one single directory.

• Data is also classified according to the gender (male/female) of the speakers and according to the way they speak (native/non-native).

• Audio and transcriptions in the LSC are segmented and transcribed by native speakers of the Spanish language

• Audio files in the LSC are distributed in a 16khz@16bit mono format.

• Every audio file has an ID that is compatible with ASR engines such as Kaldi and CMU-Sphinx.

Source Data

Initial Data Collection and Normalization

The LIBRIVOX SPANISH CORPUS is a speech corpus designed to train acoustic models for automatic speech recognition and it is made out of 300 audio books taken from Librivox.

Annotations

Annotation process

The annotation process is at follows:

• 1. A whole podcast is manually segmented keeping just the portions containing good quality speech.
• 2. A second pass os segmentation is performed; this time to separate speakers and put them in different folders.
• 3. The resulting speech files between 5 and 10 seconds are transcribed by students from different departments (computing, engineering, linguistics). Most of them are native speakers but not with a particular training as transcribers.

Who are the annotators?

The LIBRIVOX SPANISH CORPUS was created under the umbrella of the social service program "Desarrollo de Tecnologías del Habla" of the "Facultad de Ingeniería" (FI) in the "Universidad Nacional Autónoma de México" (UNAM) between 2016 and 2019 by Carlos Daniel Hernández Mena, head of the program.

Personal and Sensitive Information

The dataset could contain names revealing the identity of some speakers; on the other side, the recordings come from publicly available podcasts, so, there is not a real intent of the participants to be anonymized. Anyway, you agree to not attempt to determine the identity of speakers in this dataset.

Considerations for Using the Data

Social Impact of Dataset

This dataset is valuable because it contains well pronounced speech with low noise.

Discussion of Biases

The dataset is gender balanced. It is comprised of 77 female speakers and 77 male speakers.

Other Known Limitations

LIBRIVOX SPANISH CORPUS by Carlos Daniel Hernández Mena is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License CC-BY-SA-4.0 and it utilizes material from Librivox. This work was done with the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Dataset Curators

The dataset was collected by students belonging to the social service program "Desarrollo de Tecnologías del Habla". It was curated by Carlos Daniel Hernández Mena in 2019.

Licensing Information

CC-BY-SA-4.0

Citation Information

@misc{carlosmena2020librivoxspanish,
      title={LIBRIVOX SPANISH CORPUS: Audio and Transcriptions taken from Librivox.org},
      ldc_catalog_no={LDC2020S01},
      DOI={https://doi.org/10.35111/a44z-6x49},
      author={Hernandez Mena, Carlos Daniel},
      journal={Linguistic Data Consortium, Philadelphia},
      year={2020},
      url={https://catalog.ldc.upenn.edu/LDC2020S01},
}

Contributions

The author would like to thank to Alejandro V. Mena, Elena Vera and Angélica Gutiérrez for their support to the social service program: "Desarrollo de Tecnologías del Habla." He also thanks to the social service students for all the hard work.

Special thanks to the Librivox team for publishing all the recordings that constitute the LIBRIVOX SPANISH CORPUS.

This dataset card was created as part of the objectives of the 16th edition of the Severo Ochoa Mobility Program (PN039300 - Severo Ochoa 2021 - E&T).