SpeechFake: A Large-Scale Multilingual Speech Deepfake Dataset Incorporating Cutting-Edge Generation Methods

Resource type

Authors/contributors

Huang, Wen (Author)
Gu, Yanmei (Author)
Wang, Zhiming (Author)
Zhu, Huijia (Author)
Qian, Yanmin (Author)

Title

Abstract

SpeechFake is a large-scale multilingual dataset for speech deepfake detection, featuring over 3 million fake samples across 46 languages. Generated using 30 diverse open-source models* spanning text-to-speech (TTS), voice conversion or clone (VC), and neural vocoder (NV) methods, it offers rich metadata and strong coverage of modern generation techniques, enabling robust and generalizable detection research.

Citation Key

_bd

URL

https://github.com/YMLLG/SpeechFake

Citation

Huang, W., Gu, Y., Wang, Z., Zhu, H., & Qian, Y. (n.d.). SpeechFake: A Large-Scale Multilingual Speech Deepfake Dataset Incorporating Cutting-Edge Generation Methods [Dataset]. Retrieved https://github.com/YMLLG/SpeechFake

Audio Data

Synthetic Speech