ASVspoof 2021: Towards Spoofed and Deepfake Speech Detection in the Wild

[Md Sahidullah]

*Apologies for cross-posting*

Dear colleagues,

we are pleased to announce release of the keys and detailed metadata for the ASVspoof 2021 logical access (LA), physical access (PA) and deepfake (DF) databases used for the ASVspoof 2021 challenge (website: https://www.asvspoof.org/index2021.html). 

Each database consists of real bonafide (human) and spoofed speech generated with diverse spoofing attack algorithms. The databases are available under an Open Data Commons Attribution Licence and can be downloaded from the Zenodo repository.

SPEECH DATA:

Link to the LA subset: https://zenodo.org/record/4837263 

Link to the PA subset: https://zenodo.org/record/4834716 

Link to the DF subset: https://zenodo.org/record/4835108

EVALUATION KEYS AND METADATA:

Link to the LA keys and metadata: https://www.asvspoof.org/asvspoof2021/LA-keys-full.tar.gz

Link to the PA keys and metadata: https://www.asvspoof.org/asvspoof2021/PA-keys-full.tar.gz

Link to the DF keys and metadata: https://www.asvspoof.org/asvspoof2021/DF-keys-full.tar.gz  

CHALLENGE BASELINES & SCRIPTS TO COMPUTE EER and min t-DCF:

https://github.com/asvspoof-challenge/2021 

POST-CHALLENGE ANALYSIS PAPER:

Finally, an overview article on the ASVspoof 2021 challenge has been submitted to IEEE/ACM-T-ASLP.  It contains a detailed analysis of the results, limitations and potential future directions and is available from arXiv: https://arxiv.org/abs/2210.02437 

We hope the databases and additional resources will be of use for benchmarking and future research.

Best regards,

Xuechen Liu, Xin Wang, Md Sahidullah, Jose Patino, Héctor Delgado, Tomi Kinnunen, Massimiliano Todisco, Junichi Yamagishi, Nicholas Evans, Andreas Nautsch and Kong Aik Lee

Contact: in...@asvspoof.org