znhoughton/TeleSpoof

Important note

This dataset includes most of the data presented in Houghton, Pluth, Hosier, & Gurbani (2025). However, due to privacy considerations, part of the personal data that was collected is not released as part of this data. This data is referred to as "clipwise" in the paper. Other than the exclusion of clipwise, this dataset includes all the other data mentioned in the paper.

In the future, we hope to update this with additional data that we are not prohibited from releasing.

Usage:

The audio files are included inside .tar.gz files. These can be accessed by either following the method used in the dataset-usage-example.ipynb where the files are accessed from the hf_path which lists which .tar.gz file the target file is.

Alternatively, the files can be accessed by extracting all of the .tar.gz files and then using the path value which includes the path of each file.

If you find the dataset useful, please cite the paper as follows:

@INPROCEEDINGS{dataset-telephony-2025,
author={Zachary Nicholas Houghton and Dan Pluth and Jordan Hosier and Vijay K. Gurbani},
booktitle={2025 IEEE International Conference and Expo on Real Time Communications at Illinois Institute of Technology (RTC)},
title={A Novel Dataset for Testing Anti-spoofing Models in a Telephony Environment},
month={October},
year={2025},
volume={},
number={},
pages={},
organization={IEEE}
}

In addition to citing the above paper, please cite the following papers as well:

Call Home dataset:

@article{canavan1997callhome,
title={Callhome american english speech},
author={Canavan, Alexandra and Graff, David and Zipperlen, George},
journal={Linguistic Data Consortium},
year={1997}
}

MLAAD dataset:

@inproceedings{muller2024mlaad,
title={Mlaad: The multi-language audio anti-spoofing dataset},
author={M{"u}ller, Nicolas M and Kawa, Piotr and others},
booktitle={2024 International Joint Conference on Neural Networks (IJCNN)},
pages={1--7},
year={2024},
organization={IEEE}
}

M-AILABS dataset:

@misc{dataset2024m,
title={The m-ailabs speech dataset},
author={Dataset, TMAS},
year={2024}
}

ASVSpoof2019 dataset:

@article{wang2020asvspoof,
title={ASVspoof 2019: A large-scale public database of synthesized, converted and replayed speech},
author={Wang, Xin and Yamagishi, Junichi and others},
journal={Computer Speech & Language},
volume={64},
pages={101114},
year={2020},
publisher={Elsevier}
}