BBC R&D goes to college - PSNEurope

BBC R&D goes to college

In these highly competitive, technology reliant broadcasting times, research and development has never been more important. But as these are equally commercially tough times for broadcasters, R&D could be seen as a luxury. Kevin Hilton looks at BBC R&D's vision for future audio technologies and how academia fits in.
Publish date:

A common misconception about technological research is that it is all about boffins sitting around waiting for inspiration to strike and then spending years bringing their concept to fruition. The reality is a little more prosaic. There is still the Eureka moment - or at least the Big Idea - but the driving imperative is as much about getting developments on to the market. For broadcasters this means developing something that benefits both the programme makers and viewers at home, as well as making savings on the cost of production. R&D does not come cheap and is even more expensive when done in isolation, so collaborations between institutions are more common. In 2009 Matthew Postgate, Controller of BBC Research and Development, said he was keen to establish collaborative projects with other broadcasters, industry and academia. The first fruits of that aim can be seen in the Audio Research Partnership, a five-year collaboration between BBC R&D and five UK universities that are international leaders in sound. The primary partners with BBC R&D are the Universities of Salford and Surrey, which bring expertise in acoustics and audio-visual applications respectively. The BBC will also work closely with the Universities of York and Southampton and Queen Mary, University of London. Graham Thomas, Section Leader of the Production Magic team within BBC R&D, says this is the core of the Partnership but adds that bringing in other academic institutions and possibly research bodies in industry is still a possibility. "We have collaborated with academia in the past but on a smaller scale," he comments. "We're taking the idea further by identifying universities that we thought would be really good to work with because they are world experts in their field." In recent years the BBC has come under pressure from government and detractors to reduce expenditure and concentrate on making "quality" programmes. Regardless of its merit, research is an area that is often cut back so any new R&D partnerships could be seen as a way to scale down the BBC's own operations in this area. Thomas states the Audio Research Partnership and any similar projects are not associated with plans to downsize BBC R&D and farm out the work. "We want to expand what we do in an encouraging and viable way," he comments, pointing to the recent expansion of the in-house audio research team. In the last ten years attention has been on the visual aspects of TV broadcasting; high definition, interactivity and, most recently, stereoscopic 3D. Expanding the audio R&D section and forming the Research Partnership, says Thomas, shows that the BBC now wants to get sound back on an equal footing with its pictorial counterpart. "Audio had become a little over-looked and eclipsed," explains Thomas. "My colleagues in Audio & Music were looking at things like Ultra High Definition TV and asked me 'What's sound got?' So that's why we're re-focusing on audio, to allow it to catch up." Thomas' background is in video so he acknowledges this is an opportunity for him to say audio has not been forgotten and "do something exciting" with it. The main areas the Partnership will look at to begin with are: Source separation (analysing individual components in a signal for "audio un-mixing"); audio semantics (automated metadata); spatial audio, involving expanding BBC R&D's existing research into Ambisonics and Periphony (essentially 3D sound); and room acoustics to optimise rooms for spatial sound in both broadcast centres and the home. There is also pure research into speech recognition for archives and audio coding, involving low delay and IP distribution. A lot of this, says Thomas, is applicable to both radio and TV, with online and mobile delivery also a major consideration. But with the BBC making further moves into 3D TV during recent Wimbledon coverage, there are calls for an audio accompaniment to stereoscopy that is more than current discrete surround sound. Which is where spatial sound could come in. "It goes beyond 5.1 by having channels above and below as well as front and back," Thomas explains. This research takes place at BBC R&D North's new facilities at MediaCityUK in Salford, which have a dedicated listening room (pictured, with an Ambisonics decode unit). The room has numerous loudspeakers and Thomas says there is the general view that spatial sound can only be reproduced in this way. "We're looking at different ways of doing this, including using headphones and not having specific configurations of loudspeakers," he comments. "In general it opens up the possibilities." Part of the research is to determine whether Ambisonics is the best method for achieving Periphony. The University of Salford is contributing its experience in room acoustics to look at how spaces can be improved to compensate for their shortcomings and achieve 3D sound without turning homes into labs. The University and BBC R&D are also involved, along with Technicolor and other media groups and institutions, in the pan-European FascinatE project on high resolution interactive and immersive television. The BBC R&D lab at Salford is now in full operation. And while researchers do not like to be held too much to a timetable, Thomas says the intention is to deliver first results within a year or two. Of course, any practical or commercial implementation is a little further off than that.


BBC evaluates surround Carols

Carols from King's College, Cambridge are not only a traditional feature of UK Christmas broadcasting, they are a test-bed for audio recording experiments and at the end of last year the BBC put different versions online for evaluation, writes Kevin Hilton.