ADD 2022: the first Audio Deep Synthesis Detection Challenge

Yi, Jiangyan; Fu, Ruibo; Tao, Jianhua; Shuai, Nie,; Ma, Huili; Wang, Chenglong; Wang, Tao; Tian, Zhengkun; Bai, Ye; Fan, Cunhang; Shan, Liang; Wang, Shiming; Zhang, Shuai; Yan, Xinrui; Xu, Lei; Wen, Zhengqi; Li, Haizhou; Lian, Zheng; Liu, Bin

doi:10.1109/icassp43922.2022.9746939

Cited by 66 publications

(30 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Aim to accelerate and foster research on detecting deep synthesis and manipulated audios, Audio Deep Synthesis Detection Challenge [18] is held as Signal Processing Grand Challenge on ICASSP 2022. The challenge contains four tracks, among which the Track 3 is an audio fake game (FG) which includes two sub tasks: Track 3.1 Generation task (FG-G) is a generation task aims to generate fake audios that can fool the fake detection model.…”

Section: Introductionmentioning

confidence: 99%

Time Domain Adversarial Voice Conversion for ADD 2022

Cheng

Guo

Tan

et al. 2022

ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

In this paper, we describe our speech generation system for the first Audio Deep Synthesis Detection Challenge (ADD 2022). Firstly, we build an any-to-many voice conversion (VC) system to convert source speech with arbitrary language content into target speaker's fake speech. Then the converted speech generated from VC is post-processed in time-domain to improve the deception ability. The experimental results show that our system has adversarial ability against anti-spoofing detectors with a little compromise in audio quality and speaker similarity. This system ranks top in Track 3.1 in the ADD 2022, showing that our method could also gain good generalization ability against different detectors.

show abstract

Section: Introductionmentioning

confidence: 99%

Time Domain Adversarial Voice Conversion for ADD 2022

Cheng

Guo

Tan

et al. 2022

ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…Resorting to the power of deep neural networks (DNNs) in representation learning, speaker embedding-based ASV approaches have achieved state-of-the-art performances on several benchmark datasets [1][2][3][4][5][6][7][8][9][10][11][12]. Despite the impressive performance, current ASV models are vulnerable to spoofing attacks, which include audio replay, text-to-speech (TTS) and voice conversion (VC) [13][14][15][16][17][18], back-door attacks [19] and related emerged adversarial attacks [20,21]. Various methods have been proposed to tackle adversarial attacks [22][23][24][25][26][27][28][29].…”

Section: Introductionmentioning

confidence: 99%

Tackling Spoofing-Aware Speaker Verification with Multi-Model Fusion

Wu¹,

Kang²,

Meng³

et al. 2022

The Speaker and Language Recognition Workshop (Odyssey 2022)

View full text Add to dashboard Cite

Recent years have witnessed the extraordinary development of automatic speaker verification (ASV). However, previous works show that state-of-the-art ASV models are seriously vulnerable to voice spoofing attacks, and the recently proposed highperformance spoofing countermeasure (CM) models only focus solely on the standalone anti-spoofing tasks, and ignore the subsequent speaker verification process. How to integrate the CM and ASV together remains an open question. A spoofing aware speaker verification (SASV) challenge has recently taken place with the argument that better performance can be delivered when both CM and ASV subsystems are optimized jointly. Under the challenge's scenario, the integrated systems proposed by the participants are required to reject both impostor speakers and spoofing attacks from target speakers, which intuitively and effectively matches the expectation of a reliable, spoofingrobust ASV system. This work focuses on fusion-based SASV solutions and propose a multi-model fusion framework to leverage the power of multiple state-of-the-art ASV and CM models. The proposed framework vastly improves the SASV-EER from 8.75% to be 1.17%, which is 86% relative improvement compared to the best baseline system in the SASV challenge.

show abstract

“…To increase the security of the ASV system and reduce the risk of spoofing attacks, the ASVspoof challenge has been successfully held in 2015 [32], 2017 [10] , 2019 [28] and 2021 [34]. And then the first Audio Deep Synthesis Detection Challenge (ADD 2022) [37] was also successfully held. To improve the performance of the audio deepfake detection system, some researchers divided the whole frequency band into different subbands and explored the impact of different frequency bands on the audio deepfake detection task.…”

Section: Introductionmentioning

confidence: 99%

Audio Deepfake Detection Based on a Combination of F0 Information and Real Plus Imaginary Spectrogram Features

Xue¹,

Fan²,

Zhao³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Recently, pioneer research works have proposed a large number of acoustic features (log power spectrogram, linear frequency cepstral coefficients, constant Q cepstral coefficients, etc.) for audio deepfake detection, obtaining good performance, and showing that different subbands have different contributions to audio deepfake detection. However, this lacks an explanation of the specific information in the subband, and these features also lose information such as phase. Inspired by the mechanism of synthetic speech, the fundamental frequency (F0) information is used to improve the quality of synthetic speech, while the F0 of synthetic speech is still too average, which differs significantly from that of real speech. It is expected that F0 can be used as important information to discriminate between bonafide and fake speech, while this information cannot be used directly due to the irregular distribution of F0. Insteadly, the frequency band containing most of F0 is selected as the input feature. Meanwhile, to make full use of the phase and full-band information, we also propose to use real and imaginary spectrogram features as complementary input features and model

show abstract

ADD 2022: the first Audio Deep Synthesis Detection Challenge

Cited by 66 publications

References 16 publications

Time Domain Adversarial Voice Conversion for ADD 2022

Time Domain Adversarial Voice Conversion for ADD 2022

Tackling Spoofing-Aware Speaker Verification with Multi-Model Fusion

Audio Deepfake Detection Based on a Combination of F0 Information and Real Plus Imaginary Spectrogram Features

Contact Info

Product

Resources

About