Participants will be introduced to the basics of state-of-the-art automatic speaker recognition systems and how a system can be used to conduct research. On the first day, we will cover pre-processing, feature extraction, speaker modelling, scoring and calibration. On the second day, we will focus on developing research questions and running tests that are relevant to your own research. The course will offer a mixture of mini-lectures and practical sessions, where you will get to use the VOCALISE automatic speaker recognition system, produced by Oxford Wave Research.

The course will take place in-person at the University of York on Thursday 16 and Friday 17 January 2025. The course will run from 9:30am to 4:30pm BST (GMT+1) on the Thursday and from 9:30am to 12:30pm on the Friday.

You must be a current research student to be eligible to attend the course.

Course Description

In-person at the University of York.

Course Leader: York Forensic Speech Science Team, Oxford Wave Research

This day-and-a-half-long workshop will focus on understanding the basics of automatic speaker recognition and how you can use a system for research purposes. The course is intended for research students working in any area who are interested in (i) learning more about how automatic speaker recognition systems work, (ii) understanding how systems can be used in forensic casework, and (iii) expanding their methodological toolkit to use automatic speaker recognition systems in research. 

 

The course will mix theory with hands-on practical sessions, where you will get to use the VOCALISE automatic speaker recognition system. You will also have the opportunity to discuss your research plans with members of the York Forensic Speech Science team and representatives from Oxford Wave Research.

 

The course will enable participants to:

• Learn the basic workflow of state-of-the-art automatic speaker recognition systems

• Conduct both individual comparisons and larger-scale tests with an automatic speaker recognition system

• Explore the potential of automatic systems, whilst understanding the challenges associated with their use both in forensic casework and in research

• Develop well-motivated research questions that can be addressed with automatic speaker recognition systems

 

Tutors

Vincent Hughes

Amelia Gully

Jessica Wormald

(Philip Harrison)

Finnian Kelly

Linda Gerlach