accentdb

  • Description:

AccentDB is a multi-pairwise parallel corpus of structured and labelled accented speech. It contains speech samples from speakers of 4 non-native accents of English (8 speakers, 4 Indian languages); and also has a compilation of 4 native accents of English (4 countries, 13 speakers) and a metropolitan Indian accent (2 speakers). The dataset available here corresponds to release titled accentdb_extended on https://accentdb.github.io/#dataset

Split Examples
'train' 17,313
  • Feature structure:
FeaturesDict({
    'audio': Audio(shape=(None,), dtype=int64),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=9),
    'speaker_id': string,
})
  • Feature documentation:
Feature Class Shape Dtype Description
FeaturesDict
audio Audio (None,) int64
label ClassLabel int64
speaker_id Tensor string
  • Citation:
@InProceedings{ahamad-anand-bhargava:2020:LREC,
  author    = {Ahamad, Afroz  and  Anand, Ankit  and  Bhargava, Pranesh},
  title     = {AccentDB: A Database of Non-Native English Accents to Assist Neural Speech Recognition},
  booktitle      = {Proceedings of The 12th Language Resources and Evaluation Conference},
  month          = {May},
  year           = {2020},
  address        = {Marseille, France},
  publisher      = {European Language Resources Association},
  pages     = {5353--5360},
  url       = {https://www.aclweb.org/anthology/2020.lrec-1.659}
}