Audio Description is becoming a significant part of television and film production.
What is it, why does it matter to you, and how are you supposed to deal with it?

Imagine watching a film with your eyes closed. Who just entered the room? What was in the letter? Why did everyone suddenly go quiet? For people who are blind or visually impaired, Audio Description can make the difference between simply hearing a programme and actually experiencing the story alongside everyone else.

So why does this matter to you? European legislation requires audiovisual media services to become progressively more accessible, with Audio Description explicitly part of that framework. And the requirements are already finding their way into production. 

At a SOUND DISCOVERY industry day at TV 2 – initiated and facilitated by Danish Sound Cluster – users, Audio Description professionals, broadcasters and technology developers came together to look at what this means in practice.

Nick Dunkerley sums up what we learned: why Audio Description matters, how it is produced today, why the current process is slow and expensive, and how the industry can make it work at scale. And is there a perhaps there is a silver lining.

Expanding experiences with multisensory stimuli

Live music brings people together, yet concerts often rely on hearing and familiar ways of participating. This workshop explores how haptic feedback, visualizations, spatial audio, and audience interaction can offer more ways to experience and shape a performance.

Working with performers and participants with diverse sensory abilities, we will discuss, create and evaluate inclusive concert experiences where accessibility becomes part of the artistic design.

Who Are My Conversational Partners?

Turn-Taking-Based Identification in Multi-Talker Scenarios

by Leila Eslami

In multi-talker environments, hearing aids face a challenge beyond noise reduction: identifying which talker the user intends to engage with. This presentation focuses on a turn-taking-based approach that uses voice-activity patterns to identify conversational partners. I will present the Conversational Alignment Score (CAS), its generalization to conversations with different numbers of talkers, and recent results showing how identification accuracy depends on the amount of observed conversational activity. I will also briefly discuss head movement as an additional behavioral cue that may support faster and more robust conversational partner identification.

by Asjid Tanveer

This study shows how portable headcaps (mobile EEG) can track which speaker a person is listening to in multi-talker environments. Using a contrastive learning framework, we use a deep learning three model pipeline where the framework matches EEG brain signals directly to speech patterns across three listening scenarios. The model outperforms traditional decoding methods and works in a subject-independent manner, offering a practical path forward for smart hearing aids.

Elaine NG – Director of Research at Oticon

Elaine Ng is Director of Research at Oticon A/S, Denmark, and Adjunct Associate Professor at Linköping University, Sweden. A trained clinical audiologist, she brings more than 20 years of experience spanning audiological practice and hearing research. Since joining Oticon in 2017, she has led BrainHearing research, exploring the interplay between hearing, cognition, and communication, and how hearing aids can support better communication and cognitive outcomes for people with hearing loss.

Martin Bergmann – Co-Founder and CTO at Augmented Hearing

Martin has 28 years of experience in digital audio, AI, and software development. He holds an MSc in Engineering and has a track record of spotting technology opportunities that meet real user needs, and of taking them from first concept to market-ready product. Along the way he has led numerous initiatives across research, engineering, and product.

As CTO of Augmented Hearing, Martin sets the technical direction of the company and its Sharpi product: real-time, causal speech enhancement that deploys across a range of hardware platforms, from servers to embedded devices. Sharpi is built for demanding environments such as contact centers, emergency dispatch, and critical communications, where speech clarity matters. His long background in hearing care keeps the work anchored in how people actually listen and where communication breaks down, so the solutions are both technically advanced and practical to deploy.

Eirini Liapikou – Senior Audio & Embedded Software QA Engineer at Bang & Olufsen.

Eirini holds an MEng degree in Civil Engineering, MSc in Acoustic Engineering from DTU, and studied Sound and Music Computing at Aalborg University. Before B&O she worked on hearing and headset technology (GN,Oticon and more). She has published in JASA on lowering headphone listening levels through touch, and is building a real-time audio feedback system for judo falls. Outside engineering, she is as a performer/music producer, with airplay on Greek National radio.

How does the increasing use and potential of audio and earbuds affect our culture, society, and the way we interact with each other? And how does it affect our hearing health?

What would it take for Denmark and the EU to follow the French and make an ISO standard for acoustic quality in restaurants and bars? 

We’ve talked and discussed acoustics in restaurants and bars for what seems like forever. In France, they’re trying a different approach and have introduced a standard. Is it time for us to follow the French and introduce a standard for acoustic quality here in Denmark?

Join us for an interactive session where we investigate potential solutions and look for ways to work actively on this matter on the other side of the session.

Carsten Olesen – President, HARMAN Consumer Audio

Carsten Olesen is a valued public speaker for Industry specific events with an extensive knowledge of the Consumer Audio market. He is based at the HARMAN EMEA Headquarters located in Amsterdam, in The Netherlands, and holds both engineering and business administration degrees.

Carsten Olesen started his career with HARMAN in 1999. He started as a Brand Director for HARMAN International, then moved to the Sales organization, where he was responsible for Commercial Strategy, Execution of Business Plan, Strategic Account Development, and Commercial and Distribution Policy.

In January 2009 he became the Commercial Director for HARMAN International in the EMEA Region. In 2012 he moved to the position of GM & VP of the EMEA Region. Under his leadership, the EMEA organization has undergone massive growth, as has the Consumer Audio Market.  

In June, 2019, Mr. Olesen was promoted to the position of President of the Consumer Audio Division. In this role, Carsten leads Global Consumer Audio business including the Engineering and Design functions. In 2020, he won the prestigious International TCG Leadership Award from the Grand Jury, comprised of the top executives of the leading tech retailers across Europe.

LOCATION: STUDIO HALL

Designing Sound Around Human Needs

Mengqiu Zhang – Distinguished Engineer at HARMAN

Mengqiu Zhang leads the Denmark-based innovation team focusing on environmental noise control, spatial audio, and biosensing technologies for hearables. She holds a PhD from The Australian National University and previously worked at Ericsson on MPEG-I immersive audio standardization. Her expertise spans audio signal processing, immersive audio, intelligent listening systems, and next-generation audio experiences.

Tanveer Bhuiyan – Senior Engineer at HARMAN

Tanveer Bhuiyan is experienced in deep learning and signal processing, with a focus on DSP, hearing aid applications, and objective Sound Quality models. His expertise also includes sensor technology, audiology, and the analysis, interpretation, and feature extraction of biosignals such as ECG, PPG, and EEG. In addition, he has experience with embedded systems and programming in Python, MATLAB, and C++.

About Aria 2 research glasses and the research behind Conversation Focus and Hearing Enhancement

Dr. Lunner has more than 30 years of experience spanning hearing technology, audio, cognition and human–machine interaction. In 1995, together with Oticon, he presented the DigiFocus, the world’s first hearing aid based on digital signal processing. His work on low-power digital filter banks became fundamental to several generations of hearing aids, used by millions worldwide, and contributed to two EU IST Grand Prizes.

Dr. Lunner subsequently pioneered Cognitive Hearing Science, demonstrating the importance of cognitive ability and working memory for speech understanding in challenging listening environments and for optimizing hearing-aid signal processing.

He has authored 200+ scientific publications, with approximately 12,000 citations (H-index 54), holds 40+ patents, and has reviewed research for leading journals and international funding agencies. He has also led 40 employees at Eriksholm Research Centre and built extensive collaborations with leading research groups across Europe, North America and Australia.

His contributions have been recognized with, among others, Linköping University Alumnus of the Year (2016) and the William Demant Fellowship (2017) for his role in shaping the future of audiology.

How do we benefit from technology without losing the core values of being creative, thinking human beings?

An audience interactive panel conversation about use and impact of AI in engineering, education and society.

Peter Gerstoft is a Danish-American scientist and engineer specializing in ocean acoustics, seismology, and signal processing. He is a Distinguished Data Scientist at the Scripps Institution of Oceanography at the University of California, San Diego and an adjunct professor in the Department of Electrical and Computer Engineering at UC San Diego and DTU, Copenhagen.

Gerstoft began his career in acoustics and vibrations at Odegaard & Danneskiold-Samsøe (1987–1992). He then served as a Senior Scientist at the NATO SACLANT Undersea Research Centre in La Spezia, Italy, from 1992 to 1997. Between 1999 and 2000, Gerstoft worked as a Senior Seismic Acoustic Officer with the Comprehensive Nuclear-Test-Ban Treaty Organization. He has been a Data Scientist at the Scripps Institution of Oceanography since 1997.

Since 2013, he has held an adjunct faculty position in Electrical and Computer Engineering at UC San Diego, where he teaches courses on seismology, data assimilation, and machine learning for physical systems.

LOCATION: STUDIO HALL