Benchmark-ready packaging of the DEEP-VOICE real-vs-AI-generated speech dataset
(arXiv 2308.12734), for speech anti-spoofing and
synthetic / deepfake voice detection.
DEEP-VOICE is a binary-classification benchmark: bonafide (genuine human speech)
vs. spoof (AI voice-converted speech). The spoof side is generated with
Retrieval-based Voice Conversion (RVC), converting one real speaker's recording
into the voice of another; the bonafide side is… See the full description on the dataset page:
https://huggingface.co/datasets/SpeechAntiSpoofingBenchmarks/DeepVoice.
