CodecFake+

CodecFake+ is a large-scale dataset designed to evaluate the detection of deepfake speech generated by neural audio codecs, containing a diverse range of codec architectures and data from both re-synthesis and advanced CoSG models.

Papers using CodecFake+ (6)

Towards Generalized Source Tracing for Codec-Based Deepfake Speech2025Codecfake+: A Large-scale Neural Audio Codec-based Deepfake Speech Dataset2025Codec-Based Deepfake Source Tracing via Neural Audio Codec Taxonomy2025CodecFake+: Codec-Based Resynthesized Data as a Proxy for Detecting CodecFake Speech2025CodecFake: Enhancing Anti-Spoofing Models Against Deepfake Audios from Codec-Based Speech Synthesis Systems2024 · 2 citesCodecfake: An Initial Dataset for Detecting LLM-based Deepfake Audio2024