New football sound control techology draws encouraging initial response
UK football fans have responded well to tests of a new technology that would give them control over the balance between commentary and crowd noise, as well as improving the sound quality of online broadcasts. This is the conclusion of a BBC White Paper detailing results of an experimental transmission using IP-streaming and object-based processing as part of Radio 5 Live’s coverage of the Championship League play-off final in May.
The game between Crystal Palace and Watford, held at Wembley Stadium, was streamed live on the internet as well as broadcast on medium wave analogue and digital radio. A controlled number of listeners was given access to the test coverage, which allowed them to not only choose whether they had more commentary over crowd noise or louder effects compared to the commentators’ descriptions of the action but also which end of the stadium they listened to. This meant they could hear the reactions of either the Crystal Palace fans or Watford’s supporters – or a mixture of both.
The technical set-up and test results are set out in White Paper 272, Object-based Audio Applied to Football Broadcasts – also known as the BBC Radio 5 Live football experiment – published this month by BBC R&D. One pair of stereo mics was positioned at the Crystal Palace end of Wembley and another at the Watford end. These made up three live IP streams with a mono feed from the commentary box; all were controlled using the HTML5 web audio Javascript API, which allowed the listener at home to set a balance between the crowd and commentator feeds and select which ‘side’ of the stadium was dominant in the mix.
An A-D converter was used to feed the mic signals to an USB port on a PC, where a multi-channel mapper and three 128kbps AAC codecs designed for the project delivered the streams to BBC R&D’s facility. The feeds were then transcoded into MP3 and Ogg Vorgis formats so that the broadcast could be distributed, through an Icecast server, to computers using the main operating systems.
On the technical side of the broadcast the White Paper, written by Mark Mann, Andrew Bonney, Anthony Churnside and Frank Melchior of BBC R&D, states that the test was an improvement in audio quality and clarity over 5 Live’s “typical” online broadcasts using the 56kps G.722 mono codec. The authors acknowledge there were problems, including the Icecast server having to be restarted 20 minutes into the broadcast because new connections were being made; the Ogg Vorbis system used for the commentary had to be rebooted at the beginning of the second half, causing a ‘spike’ in activity as listeners changed the crowd/commentary balance and end selection.
Initially the test was restricted to 1000 listeners but extended to 2800 during the game because of high demand. Of those who responded to the post-test survey, 57 percent said being able to control the commentary/crowd mix made for a “much better” listening experience, 22 percent said it was “slightly better”, while only seven percent found it “much worse”. A similar number, 55 percent, thought the test “much better” than traditional radio coverage; again 22 percent found it only “slightly better” and eight percent said it was “much worse”. For the impact of being able to choose ends, the “much better” figure fell to just 19 percent; 53 percent said it was “slightly better”; and seven percent found it “much worse”.
The BBC previously gave listeners the opportunity to set their own balance between commentary and crowd noise during the Wimbledon tennis championships in 2011. This was based on the Dialogue Enhancement system developed by Fraunhofer IIS; the 5 Live football experiment used object-based technology developed by BBC R&D that is part of a wider research programme into future audio formats, which began in 2012.
Technologist Anthony Churnside told SVG Europe that there were still “some challenges” to deal with before the commentary/crowd/ends technology could be “taken forward” but that, theoretically, it could be applied to both radio and television broadcasts.