You're currently viewing an old version of this dataset. To see the current version, click here.

ASVspoof2019

The ASVspoof2019 LA subset consists of three parts, training, development, and evaluation. Each partition has a disjoint set of speakers. The average duration of the utterances is 2 to 3 seconds. The spoofed utterances are generated using 19 VC and TTS algorithms (6 types of attacks in the training and the development sets, 13 in the evaluation set).

Data and Resources

Cite this as

Galina Lavrentyeva, Sergey Novoselov, Andzhukaev Tseren, Marina Volkova, Artem Gorlanov, Alexandr Kozlov (2024). Dataset: ASVspoof2019. https://doi.org/10.57702/xe01keru

DOI retrieved: December 16, 2024

Additional Info

Field Value
Created December 16, 2024
Last update December 16, 2024
Defined In https://doi.org/10.48550/arXiv.2304.03323
Citation
  • https://doi.org/10.48550/arXiv.2407.04034
  • https://doi.org/10.48550/arXiv.1904.05576
Author Galina Lavrentyeva
More Authors
Sergey Novoselov
Andzhukaev Tseren
Marina Volkova
Artem Gorlanov
Alexandr Kozlov
Homepage http://www.asvspoof.org/asvspoof2019/