Report of the meeting of the SC-03-12 Working Group on Forensic Audio, held online, 2023-10-11.
SC-03-12 chair Marisa Déry called the meeting to order.
The report of the previous meeting, held online 2023-05-02, was approved as written.
The agenda was approved as written.
In attendence were M T Déry, G P Jalandoni Elviña, T Birkeland, S Beck, G Reid, A Zlogar, A Rosen, J K McElveen, M Noisternig, D Notowitz, D Marks, H Malik, J Broyle, D Andrews, C Grigoras, J Hansen.
AES-X259 Criteria for the authentication of digital audio recordings
AES-X259 will likely be a progression of AES43. We need to revisit Criteria for authentication, this time in the digital medium. We went through the table of content of AES43 and will divide sections of new papers amongst ourselves. There was a discussion on document editing and exporting history. We need to briefly describe what a "lab" is now due to ISO-17025. Method validation is more important than tools, as Daubert is more about method than tools. When mentioning equipment we should state" equipment capable of" instead of brand name. Critical Listening will be an important aspect of the document. The group discussed several current docs such as ENFSE and SWGDE and the importance of reading all docs before tackling our document.
The group discussed the Biometric Privacy Act (BIPA), Best Practices guideline.
Liaisons
Andy Rosen gave a quick update on Deepfake Audio which is a concern for the entertainment industry. He also updated us on the work of CP2A.
Andy suggested we keep an eye on Franhofer in December regarding spotting Deepfakes with a spectrogram and Machine Learning where Language independence is a factor.
SMPTE will draft a paper on Ethical AI.
New projects
We will not produce a position position paper on Deepfake Audio. The Topic will be discussed at AES NY AI Town Hall.
New business
No new business was raised.
The next meeting will be held online in May 2024, scheduled in conjunction with the AES 156th Convention.
