PhD Studentship - Modelling Dialogue Understandability in Media Using Listener Types and AI

Academic supervisor: Dr. Ben Shirley, b.g.shirley@salford.ac.uk

Academic co-supervisor: Dr. Rebecca Vos

Industrial supervisors: Dr. Martin Walsh and Ted Laverty (Xperi)

The studentship is fully funded and includes:

• A fee waiver
• A stipend of £22,852 per annum for three and a half years
• All bench fees and consumable costs
• Funds specifically allocated for conference travel

Final date for applications: 30^th September 2026

Interviews will be held on 21^st October 2026

The candidate must be in a position to register by January 2027

Project overview

This fully funded PhD studentship, supported by DTS (Xperi), will investigate why dialogue in film, television and streaming media can be difficult to understand even when it is audible. The project will move beyond traditional speech intelligibility measures by focusing on dialogue understandability: how well listeners comprehend narrative content without excessive effort.

Research aim and objectives

The project will develop a listener-aware framework for predicting dialogue understanding across different audiences, devices and listening environments. It aims to produce a quantitative Dialogue Understandability Rating that reflects real-world viewing experiences more accurately than existing audio quality metrics.

Key objectives include:

• Defining and measuring Dialogue Understandability for quantitative rating
• Modelling listener types, including hearing profiles, language proficiency and accent familiarity
• Exploring how AI, speech models and large language models can support prediction of comprehension and listening effort
• Developing validated models and evaluation protocols for media content.

Approach and outcomes

The research will combine speech and audio analysis with AI techniques, using existing datasets and newly collected listener data. It will examine how listening context, device, room acoustics and cognitive effort affect understanding. Expected outcomes include a validated computational model, new datasets and experimental protocols, and practical insights for streaming, broadcast, accessibility and media production.

Candidate profile

Applicants should have a background in machine learning, audio engineering, speech processing or a related discipline, with an interest in human-centred media technologies.

Funding eligibility: This studentship is only available to students with Home fee status in the UK.

Enquiries: Informal enquiries may be made to Dr. Ben Shirley by email: b.g.shirley@salford.ac.uk

Curriculum vitae and supporting statement explaining their interest should be sent to: m.watts@salford.ac.uk

Note to applicant: In addition to applying for this role the successful candidate will also be required to complete the University application process which applies to all students wishing to study at The University of Salford. How to apply for research studies can be found here: www.salford.ac.uk/study/postgraduate/applying/applying-for-research

Please note that the deadline for student admissions is Monday 16^th November 2026 for a January 2027 intake.

Advert information