Most modern radio receiving devices but also cars, smartphones include a colour screen and integrate digital, hybrid or web radio. With these features, pictures can be transmitted alongside the audio.
EBU/Eurovision has added visualisation for the live sports events it covers.
The content provided consists of an international signal with live sport results, photos from the events and quotes that broadcasters can use on their radio or web platforms.
Many broadcasters have distributed the signal either on DAB/DAB+1 or RadioDNS2 or embedded the slideshow in their website.
In this paper, we describe briefly how the project for sports was set in place, the workflow and how it has been distributed and used.
A platform was developed that automatically ingest, process the content and produce the slides.
We explore also the future possibilities for this new kind of media and how analytics could be used.
Sport coverage on radio is popular, but has specific requirements: people do other things while listening to radio, so any enhancements to the audio experience must respect this.
Visual Radio allows to transmit pictures alongside the audio. However, there’s a clear distinction from video content: this visualisation is not meant for watching but for occasional glancing and generally the slides stays for 10-20 seconds and can be repeated.
Producing images is new for radio but we can already see radio stations providing CD covers, illustrations of the show that is on air or sometimes non program associated information such as weather, general road traffic status, etc.
Eurovision is already covering many sport events for television thanks to the partnership they have with many sport federations. So came the idea to explore possibilities for radio.
After a first test on Swimming championship in Barcelona with some EBU broadcasters in 2013, a real world live coverage was provided on European Championships in Zurich in 2014 and also for Skiing Championships in Vail in 2015.
The challenge is to produce content in an easy and low cost way because broadcasters cannot afford growing their production team for these visual elements.
Figure 1 shows the big picture of the workflow and infrastructure put in place.
Here’s a brief description of the functional elements of the workflow:
Data Gateway: This is the content ingest, receiving the information sources and storing the data in a database.
Data Processor: This block is in charge of processing ingested content to filter and extract the data and prepare them for production
Content Manager: this is the actual slides production platform running on the workstation of the person controlling the visual production. It receives pre-processed data that it applies in slides templates. The output of the content manager is images in various format.
Data Dispatch: delivery of the produced content on the different platforms of the participating broadcasters. DAB slideshow, RadioDNS/RadioVIS
WebSocket Edge: this elements translates RadioDNS/RadioVIS for the use on web widgets to be integrated in webpages of broadcasters.
All these elements consist of instances on Amazon Cloud Service and are connected using RabbitMQ open messaging communication.
This choice was motivated by the time and budget constraint of the project and the greater flexibility offered by this kind of platform.
For example the content manager instance could be run from various locations depending on the time of the day.
CONTENT INGEST AND PROCESSING
All sport events have organisations in charge of the timing and results management. Swisstiming is one of the major companies providing timing information. They accepted to make a partnership on this project by providing live raw data during the event covered.
These data consist of name of competition going on, athletes ID, intermediate and final times, global result lists with lots of other information, all this sent as text via FTP.
Standards exist for some sports such as the Olympic Data Feed (ODF) that is a XML format defined for athletics and that was used for the case of Athletics Championships in Zürich in 2014.
For sports like skiing or swimming, these data can be processed in a linear way as competitions happen but for athletics, many competitions happens at the same time and car must be taken to filter out the data.
This was the difficult part of the project to handle the data, extract information, map it to the correct athletes and completion, manage the exceptions or errors to avoid wrong or incomplete results to be displayed.