Source recording device recognition based on device noise estimation

作者：Zou, Ling; He, Qian-Hua^*; Kuang, Xi-Chao; Li, Hong-Tao; Cai, Zi-Wen

来源：Journal of Jilin University (Engineering and Technology Edition), 2017, 47(1): 274-280.

DOI：10.13229/j.cnki.jdxbgxb201701040

• Classification technique
• Device noise
• Digital audio
• Noise estimation
• Noise estimation algorithm
• Noise estimators
• Recording devices
• Speech recording

To resolve this problem, first, the procedure of the speed recording is analyzed and a recording device fingerprint based on device noise is proposed. Then, a noise estimator including two noise estimation algorithms is applied to acquire sufficient device noise. Finally, to evaluate the effectiveness of the proposed device fingerprint, five classification techniques are utilized to identify the source recording device. Experiments are carried out on two corpora. One corpus comprises speech recordings obtained by 22 recording devices coming from four types of common recording devices, and the other corpus consists of speech recordings from 21 cell phones. Experiment results verify the effectiveness of the proposed method.

• 出版日期2017-1-1
• 单位华南理工大学