"In this contribution we show the influence of Voice over IP (VoIP) transmission for the task of forensic speaker verification if packet loss occurs. We particularly tested the influence of packet loss up to 20% on two modern automatic speaker verification systems with different feature sets. One of the systems was our own development to control features and the classifier parameters. The other one was a commercial system used in the forensic field. Furthermore, a phonetic acoustic analysis of the formant deviation and its influence on speaker verification is given. The results indicate that packet loss up to 20% has only minor influence on the performance of modern computer-based speaker verification tools and that the changes of the formant frequencies are not significant. However, some dynamic formant transitions get lost at high packet loss rates."
Authors:
Bitzer, Joerg; Rollwage, Christian; Neumann, Martin
Affiliations:
Project Group Hearing, Speech and Audiotechnology, Fraunhofer Institute for Digital Media Technology, Oldenburg, Germany; University of Applied Science Wilhelmshaven/Oldenburg, Oldenburg, Germany(See document for exact affiliation information.)
AES Conference:
54th International Conference: Audio Forensics (June 2014)
Paper Number:
2-3
Publication Date:
June 12, 2014
Subject:
Forensic Speaker Recognition
Click to purchase paper as a non-member or you can login as an AES member to see more options.
No AES members have commented on this paper yet.
To be notified of new comments on this paper you can subscribe to this RSS feed. Forum users should login to see additional options.
If you are not yet an AES member and have something important to say about this paper then we urge you to join the AES today and make your voice heard. You can join online today by clicking here.