Machine Learning—Based Splicing Detection in Digital Audio Recordings for Audio Forensics
×
Cite This
Citation & Abstract
R. Patole, and PR. P.. Rege, "Machine Learning--Based Splicing Detection in Digital Audio Recordings for Audio Forensics," J. Audio Eng. Soc., vol. 69, no. 11, pp. 793-804, (2021 November.). doi: https://doi.org/10.17743/jaes.2021.0044
R. Patole, and PR. P.. Rege, "Machine Learning--Based Splicing Detection in Digital Audio Recordings for Audio Forensics," J. Audio Eng. Soc., vol. 69 Issue 11 pp. 793-804, (2021 November.). doi: https://doi.org/10.17743/jaes.2021.0044
Abstract: Authentication of audio recordings is an important task in the field of audio forensics. Splicing is the practice of manipulating recorded audio to replace or insert an external sound into the original audio track. Due to the ease with which digital audio recordings can be spliced, forgery and tampering of audio recordings with a criminal intent or intent to destroy their integrity are common practices. This paper describes a methodology for splicing detection in digital audio recordings with a comparative analysis of the effectiveness of different feature sets and classifiers. Different feature sets including conventional, chroma, and reverberation-based features are evaluated, compared, and combined to produce better classification accuracy. Exhaustive experimentation has been done to take into account factors such as the duration of the attack, effect of noise, and effect of compression. The Analytic Hierarchy Process is used to evaluate different performance parameters and identify the most suitable machine learning classifier for splicing detection based on priority weights assigned to the different performance parameters. Results indicate that Long Short-Term Memory with a feature set containing Mel-Frequency Cepstral Coefficients and Decay Rate Distribution features has the best performance compared with other classifiers and feature sets.
@article{patole2021machine,
author={patole, rashmika and rege, priti p.},
journal={journal of the audio engineering society},
title={machine learning--based splicing detection in digital audio recordings for audio forensics},
year={2021},
volume={69},
number={11},
pages={793-804},
doi={https://doi.org/10.17743/jaes.2021.0044},
month={november},}
@article{patole2021machine,
author={patole, rashmika and rege, priti p.},
journal={journal of the audio engineering society},
title={machine learning--based splicing detection in digital audio recordings for audio forensics},
year={2021},
volume={69},
number={11},
pages={793-804},
doi={https://doi.org/10.17743/jaes.2021.0044},
month={november},
abstract={authentication of audio recordings is an important task in the field of audio forensics. splicing is the practice of manipulating recorded audio to replace or insert an external sound into the original audio track. due to the ease with which digital audio recordings can be spliced, forgery and tampering of audio recordings with a criminal intent or intent to destroy their integrity are common practices. this paper describes a methodology for splicing detection in digital audio recordings with a comparative analysis of the effectiveness of different feature sets and classifiers. different feature sets including conventional, chroma, and reverberation-based features are evaluated, compared, and combined to produce better classification accuracy. exhaustive experimentation has been done to take into account factors such as the duration of the attack, effect of noise, and effect of compression. the analytic hierarchy process is used to evaluate different performance parameters and identify the most suitable machine learning classifier for splicing detection based on priority weights assigned to the different performance parameters. results indicate that long short-term memory with a feature set containing mel-frequency cepstral coefficients and decay rate distribution features has the best performance compared with other classifiers and feature sets.},}
TY - paper
TI - Machine Learning--Based Splicing Detection in Digital Audio Recordings for Audio Forensics
SP - 793
EP - 804
AU - Patole, Rashmika
AU - Rege, Priti P.
PY - 2021
JO - Journal of the Audio Engineering Society
IS - 11
VO - 69
VL - 69
Y1 - November 2021
TY - paper
TI - Machine Learning--Based Splicing Detection in Digital Audio Recordings for Audio Forensics
SP - 793
EP - 804
AU - Patole, Rashmika
AU - Rege, Priti P.
PY - 2021
JO - Journal of the Audio Engineering Society
IS - 11
VO - 69
VL - 69
Y1 - November 2021
AB - Authentication of audio recordings is an important task in the field of audio forensics. Splicing is the practice of manipulating recorded audio to replace or insert an external sound into the original audio track. Due to the ease with which digital audio recordings can be spliced, forgery and tampering of audio recordings with a criminal intent or intent to destroy their integrity are common practices. This paper describes a methodology for splicing detection in digital audio recordings with a comparative analysis of the effectiveness of different feature sets and classifiers. Different feature sets including conventional, chroma, and reverberation-based features are evaluated, compared, and combined to produce better classification accuracy. Exhaustive experimentation has been done to take into account factors such as the duration of the attack, effect of noise, and effect of compression. The Analytic Hierarchy Process is used to evaluate different performance parameters and identify the most suitable machine learning classifier for splicing detection based on priority weights assigned to the different performance parameters. Results indicate that Long Short-Term Memory with a feature set containing Mel-Frequency Cepstral Coefficients and Decay Rate Distribution features has the best performance compared with other classifiers and feature sets.
Authentication of audio recordings is an important task in the field of audio forensics. Splicing is the practice of manipulating recorded audio to replace or insert an external sound into the original audio track. Due to the ease with which digital audio recordings can be spliced, forgery and tampering of audio recordings with a criminal intent or intent to destroy their integrity are common practices. This paper describes a methodology for splicing detection in digital audio recordings with a comparative analysis of the effectiveness of different feature sets and classifiers. Different feature sets including conventional, chroma, and reverberation-based features are evaluated, compared, and combined to produce better classification accuracy. Exhaustive experimentation has been done to take into account factors such as the duration of the attack, effect of noise, and effect of compression. The Analytic Hierarchy Process is used to evaluate different performance parameters and identify the most suitable machine learning classifier for splicing detection based on priority weights assigned to the different performance parameters. Results indicate that Long Short-Term Memory with a feature set containing Mel-Frequency Cepstral Coefficients and Decay Rate Distribution features has the best performance compared with other classifiers and feature sets.
Authors:
Patole, Rashmika; Rege, Priti P.
Affiliations:
Department of Electronics and Telecommunication, College of Engineering, Pune, India; Department of Electronics and Telecommunication, College of Engineering, Pune, India(See document for exact affiliation information.) JAES Volume 69 Issue 11 pp. 793-804; November 2021
Publication Date:
November 8, 2021Import into BibTeX
Permalink:
http://www.aes.org/e-lib/browse.cfm?elib=21530