IBC2022: Technical Papers Programme to address real world issues

This year’s papers cover all industry sectors across the media, entertainment and technology industry and will be delivered by a mix of industry professionals, academics and R&D experts. Issues being addressed include 5G, video coding, AI and VR/XR.

“We build the technical paper sessions around the very best submissions the industry wants to talk about,” explained Paul Entwistle, Chair of the IBC Technical Papers Committee. “Video coding is a cornerstone technology and we continue to showcase advances in this important field. Artificial intelligence and machine learnt assistance in production continues to grow, impress and raise challenging questions. Importantly, we also look beyond entertainment and explore sustainability and the broader impact our technologies have on society.”

The Technical Papers programme will begin on Friday 9 September at 11:00 in Room E102 with Energy-efficiency for a More Sustainable Future. This informative and entertaining session examines three very distinct approaches towards this objective. The first, by a major broadcaster, began as a means of exploring the economies of cloud-based working but soon became overtaken by the pandemic and home-based working. Now, post-pandemic, the organisation has been pervaded by a new culture of adventure and it has embraced this new way of working, with its very substantial energy savings.

Another development recognises that high dynamic range mobile displays, increasingly used for streamed video, consume considerable energy. The team describes the creation of an app that can predict the battery capacity required to watch HDR content; this informs users how to choose energy-saving viewing options.

Finally, we examine the novel principles employed in the design of a pro-active smart home which takes account of the lifestyles and media usage habits of its inhabitants.

Stephan Heimbecher, Head of CC Production and Infrastructure, Technology and Production Directorate & Chairman, EBU SP Digital Media Production, SWR, chairs the session, which also features Tim Guilder, Head of Production Technology – UK, ITV Studios; Joe Newcombe, Media Technology Strategist, Microsoft UK; and Regina Bernhaupt, Director User Experience Research, ruwido Austria.

This is followed by 5G – Tools, Trials and Media Production at 13:30. Featuring Daniel Silhavy, Project Manager, Fraunhofer FOKUS, and Ian Wagdin, Senior Technology Transfer Manager, BBC, this session will cover two papers. The first describes open-source tools (5G-MAG) using three relevant use-cases: TV and radio services over LTE-based 5G terrestrial broadcast, broadband-broadcast seamless switching and broadcast-on-demand.

The second paper investigates live, multi-camera production built on a 5G standalone non-public network. It discusses the interoperability challenges with existing equipment and presents solutions, as well as highlighting the additional event coverage a wireless IP-based network approach could bring.

This session is supported by two further papers: RAI explores mobile live event capture in its paper, using a hybrid 4G/5G and traditional radio link approach. In addition, Anri Consulting presents trial results of using FeMBMS (further evolved multimedia broadcast multicast service) for live, multi-angle, high-definition streaming in sports stadiums.

At 16:00, the topic is The Immersive Olympics. In this two-paper session, attendees will learn how 5.1.4 immersive audio was added and used across all sports, as well as the opening and closing ceremonies of the Tokyo 2020 and Beijing Winter 2022 games. The second paper adds VR. With over 100 hours streamed of fully produced Olympic events from Beijing 2022, with graphics and commentary including live coverage of six sports, all in 8K VR, this is set to be a masterclass in VR production. Simon Gauntlett, Director of Imaging Standards and Technology – Dolby Europe, and Nuno Duarte, Senior Audio Manager – Olympic Broadcast Services, will be guiding attendees though the detail.

Friday’s sessions conclude with New Approaches to Social Media. Among today’s diverse media technologies, social applications play a large role, providing new means for people to communicate and to share opinions and experiences with others. This session explores two developments that aim to do exactly this.

The first is a new concept in social media which has been pioneered by a 14-organisation European collaboration. Unlike established social media it operates with a peer-to-peer approach and involves novel messaging ideas and a blockchain-based rewards system. Ville Ollikainen, Research Team Leader at VTT Technical Research Centre of Finland will talk attendees though the paper.

Second, is a sign language education tool from a Korean consortium which operates on the Android TV platform. It specifically aims to assist deaf children who are born to hearing parents by teaching sign language to the whole family. The app has been on trial with 20 families since November 2021. Kevin (Inkoo) Lee, CEO at EQ4ALL, will share how successful the trial has been.

Saturday’s Technical Paper programme kicks off with a double header exploring XR and Avatars, including a look at the Best Technical Paper Award-winning 5G Edge-XR project.

“Beyond faster broadband, 5G’s capabilities are beginning to be felt in our industry. The future of enhanced entertainment with VR and XR continues to push technical boundaries with compelling user experiences,” said Entwistle.

The first paper describes a system for encoding and synthesising facial movements, which are then superimposed on either a realistic or chosen ‘face’ (it can also synthesise those movements from audio speech). Everyone has become very familiar with videoconferencing over the last two years, both professionally and socially – but the metaverse is going to be much more than a grid of talking heads. The potential for the technology, as described by Jun Xu, a Doctor at Shanghai Jiao Tong University, is fascinating.

The second paper explores 5G with edge computing, the promise that 5G can deliver more than just faster, denser broadband. A collaborative project investigating a set of genuinely compelling XR enhanced broadcast experiences, specifically in sport (football, boxing, MotoGP and rugby), this is key premium content. The paper, presented by Andrew Gower, Head of Immersive Content Research at BT, highlights the advantages gained by an edge-enhanced approach over traditional client-side processing, while acknowledging the business challenges such an infrastructure presents.

In a three-paper session starting at 12:15, both speciality as well as mainstream video encoding applications will be covered in Advances in Video Coding.

Kevin Mockford, Product Manager, V-Nova, shares the company’s approach and impressive results to encoding 6 degree-of-freedom point clouds with an emphasis on achieving real-time decode on high-end consumer devices. Julien Le Tanou, Principal Engineer, Video Technology at MediaKind, explores encoding optimisations for live streamed game content, such as esport – identifying and exploiting differences between game and natural content, the latter an assumption upon which much compression design has been built. Finally, Xavier Ducloux, Senior Innovation Manager, Harmonic, puts forward the case for Dynamic Range Encoding, a technique with legacy challenges, but which with AI assistance could now be used in live encoding situations.

This session is also supported by two further worthy papers: Bitmovin describes methods with good results that can reduce the computational cost of multi-encoding for adaptive streaming; and RAI continues to demonstrate how AI is impacting video compression design, with details of a machine learnt intra-prediction scheme and the use of super-resolution in post-loop filtering.

The 2022 Technical Papers programme ends with a session at 14:15 entitled AI Is Advancing Media Production. This will look at three thought-provoking studies. First, work supported by a consortium of national Asian cricket teams looks at how the broadcast content of cricket matches is being analysed to derive detailed performance statistics of the teams. Then, from Japan, we take a look at a remarkable new approach to text-to-speech conversion which is able to relate textual symbols directly to phonemes. The result is genuinely human-like speech which is being used to replace a human news presenter. Finally, we look at where international intellectual property law currently stands on the protection and ownership of AI-generated ‘creative’ work, such as a musical symphony. Speakers are Kiyoshi Kurihara, Research Engineer, NHK (Japan Broadcasting Corporation); Masoumeh Izadi, Founder, TVCONAL; and Lisa Logan, General Counsel, UK and EMEA, International Literary Properties. IBC