³ÉÈËÂÛ̳

Quite a few people are working on aspects of the Notifications API, which will be part of the Programmes List but will hopefully be useful for other purposes too. Duncan's been continuing work on the API itself, while Pete's delivered a nice design for basic alerts, and handed over to Theo for planning the next phase of design development.

There are also several people working on LIMO ('lightweight interactive media objects') as part of P2P Next - Theo's been working with Andrew Nicolaou on UI and use cases: Andrew's work has involved building a timeline to visualise LIMO event streams. Chris Needham and Sean are looking into GStreamer pipelines for passing external times to a streaming server for use by HTML5 video applications with associated timed data. Personally I'm interested in the potential of the LIMO file format (declarative time-based annotations for video), so I've been working on adapting one of our NoTube second screen applications to use it, and as a consequence asking Chris Needham and Sean lots of annoying questions.

Chris Lowis gave a fun talk at the team meeting about his work on the Roar-to-explore prototype - he discussed some of the problems he's been addressing in the classification of eleven animal noises made by sixty different children. This week he's been segmenting all the animal sounds into individual noises that he can use to train the classifier, and using the to extract from the audio.

A big piece of work at an earlier stage in the process is the FI Content project, an EU project about personal data and privacy. Joanne has finished off the literature review into user attitudes to personal data privacy and security that will feed into FI Content. She says: "Basically, the information people are comfortable disclosing, and the actual information they disclose, is completely contextual - including past experiences and perceived benefits". Meanwhile, Chris Needham and Barbara and others had a workshop to discuss what prototypes they might want to build around user data and personalisation for the project, prompting Tristan to ponder a different aspect of the problem - to research one aspect of a system then ideally we need a whole experience for users. But this means we either have to build the whole thing, build it on top of our current prototypes or build it into an existing ³ÉÈËÂÛ̳ product...

George had the daunting - but invigorating - task of showing RadioDNS and Autumnwatch work to the Chair and Vice-Chair of the ³ÉÈËÂÛ̳ at the R&D North open day, also having the opportunity to chat to ³ÉÈËÂÛ̳ North colleagues, journalists and local businesses about the group's work.

A few other bits and pieces:

• Marc's been exploring extracting historical location-based stories from TV news items in ³ÉÈËÂÛ̳ archives metadata and composing 'story arcs' from them
• Vicky Buser has been working out the details of the user testing for N-Screen planned for early December
• Olivier has been mostly doing W3C-related work, including poring through the
• Yves and Chris Needham have both been working on 'deliverables' - EU jargon for documents with deadlines: Yves on an architecture document for ABC-IP, and Chris on the final P2PNext deliverable on metadata and content discovery.
• Roderick has been at the Conference for Visual Media Production, discovering a great deal about computer vision and visual effects, and helping Peter Schübel show a poster for the Blue Peter Rossendale ski-jump visualisation.
• Tristan reports late-breaking news on Programme List: he got his first reminder from it on Saturday, reminding him to record The Killing.

Finally, we welcome Ralf Neudel from to the lab, working with us on an exchange for a month. Welcome Ralf!

Links of the week:

• Yves has been experimenting with , with the goal of applying it to radio streams and identify tracks in there.
• Duncan has been re-reading a post on , and also understanding .
• Tristan likes the idea of exploring some of the overlaps of .
• Olivier had a lot of fun looking at some of the (Bohemian Rhapsicord and the Drum Machine amongst his favourites), and liked the thought-provoking approach, arguing that videos which claim to envision the future shouldn't be set in a perfect, sanitised world.