Datasets:

SpeechAntiSpoofingBenchmarks
/

CFAD

license: cc-by-4.0
language:
  - zh
pretty_name: CFAD
task_categories:
  - audio-classification
size_categories:
  - 10K<n<100K
configs:
  - config_name: default
    data_files:
      - split: test
        path: data/test-*.parquet
tags:
  - anti-spoofing
  - audio-deepfake-detection
  - speech
  - benchmark
  - arena-ready
arxiv:
  - '2207.12308'

CFAD

Benchmark-ready packaging of the CFAD (Chinese Fake Audio Detection) clean test set (arXiv 2207.12308), for speech anti-spoofing and synthetic / deepfake voice detection on Mandarin Chinese speech.

Overview

CFAD is a large-scale Chinese fake-audio detection corpus. This repo packages the clean version's two test partitions:

test_seen — spoof systems and real corpora also present in the train/dev splits.
test_unseen — spoof systems and real corpora held out of training, to measure generalisation to unseen attacks.

The label is binary: bonafide (genuine human speech) vs. spoof (synthesized / vocoded speech). It is derived from the source directory layout — clips under fake_clean/ are spoof; clips under real_clean/ are bonafide. There is no separate protocol file.

Partition	Side	Sources	n
test_seen	bonafide	aishell1, aishell3, magicread, thchs30	14,000
test_seen	spoof	gl, hifigan, lpcnet, mbmelgan, pwg, straight, stylegan, wavenet	28,000
test_unseen	bonafide	magicconversa, selfrecording	7,000
test_unseen	spoof	fasthifigan, partiallyfake, tacohifigan, world	14,000

One aishell1 outlier (BAC009S0764W0123.wav, a 178 MB / ~97 min concatenation artifact) is dropped, so the bonafide count is 20,999 rather than 21,000.

License & redistribution

CFAD is released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license, which permits redistribution with attribution — see LICENSE.txt. Audio is re-encoded to 16 kHz mono FLAC for a uniform schema; labels are unmodified.

Schema

Column	Type	Description
`path`	`string`	source-relative path (e.g. `test_seen_clean/fake_clean/gl/SSB07800001_gl.wav`), unique
`audio`	`Audio(16000)`	16 kHz mono FLAC (re-encoded from the 16 kHz PCM source wav)
`label`	`ClassLabel`	`"bonafide"` (0) / `"spoof"` (1)
`notes`	`string`	JSON: `utterance_id`, `partition`, `source`

notes example:

{"utterance_id": "CFAD_test_seen_clean_fake_clean_gl_SSB07800001_gl", "partition": "test_seen", "source": "gl"}

Quick Start

from datasets import load_dataset

ds = load_dataset("SpeechAntiSpoofingBenchmarks/CFAD", split="test")
print(ds[0])

Stats

Stat	Value
Total trials	62,999
Bonafide	20,999
Spoof	42,000

Source provenance

Paper: https://arxiv.org/abs/2207.12308
Scope: the clean version's test_seen_clean + test_unseen_clean partitions (the codec/noisy robustness variants and train/dev splits are not included).
Labels derived from the source directory layout (fake_clean/ = spoof, real_clean/ = bonafide).

Evaluation

For evaluation instructions and submission format, see submissions/README.md.

Citation

Paper: Haoxin Ma et al., CFAD: A Chinese Dataset for Fake Audio Detection, arXiv:2207.12308 (2022).

@article{ma2022cfad,
  title   = {CFAD: A Chinese Dataset for Fake Audio Detection},
  author  = {Ma, Haoxin and Yi, Jiangyan and Wang, Chenglong and Yan, Xinrui and
             Tao, Jianhua and Wang, Tao and Wang, Shiming and Fu, Ruibo},
  journal = {arXiv preprint arXiv:2207.12308},
  year    = {2022}
}

Maintainer

Contact: k.n.borodin@mtuci.ru