³ÉÈËÂÛ̳

Pete made experiential improvements to the World Service Radio Archive project, ensuring visual and behavioural UI consistency across the site and designing a brief overview and feature-tour for new users. I implemented those changes in the prototype which makes the whole site look and feel a lot more unified. Meanwhile, Chris L made a couple of small bug fixes and started looking into segmentation using the results of the diarization process in earnest. He's written a library to help him work with segmentation data, including visualisation, and he's also using some of the C99 code he wrote a while back to start looking at inter-segment similarity based on the co-occurrence of speakers within each segment (see screenshot).

An episode of the Today programme annotated with the segments derived from the Chris' segmentation work

Chris L also went went to MediaCity:UK for two days to attend the Audio Research Partnership anniversary. He said: "It was great to catch up with old and new faces as well as hear about all the exciting work that is going on as part of that partnership. It seems like a great way for the ³ÉÈËÂÛ̳ to structure and communicate our collaborative work with universities."

In other Chris-related news, Chris N continued to work on the egBox TV authentication prototype and demoed to the management team on Wednesday. He got all the software installed on a dedicated laptop PC, so we can test the full end to end system. We've also just received a couple of very slim nettop PC's that we'll install egBox on. Dan and I were testing the end-to-end UX for the authentication work. Bugs were fixed and the UI was getting tweaked. All in all, everyone's very satisfied with the results. Barbara says: "We have achieved all the objectives we initially set thanks to the fantastic job everybody in the FI-team has done."

New Chris on the block, Chris F, was researching and prototyping the display of radio program waveforms for Snippets' upcoming snipping tool for radio. After trying various methods of display (canvas/SVG) and various frameworks (, , ) he's settled on a combination of SVG and d3.js to display the data and has now begun the process of prototyping a snipping tool.

Meanwhile, Anthony continued work on the Snippets Keyframe tool where there are a few memory issues to sort out on the second-screen implementation. He demoed it at the Thursday show-and-tell and got feedback from the team.

Chris G was orchestrating our impending move to the ³ÉÈËÂÛ̳'s Media Village campus at the end of the week. Olivier's going to miss the walk through Fitzrovia to our current office and getting his daily antipodean greeting at the coffee shop round the corner. However, he's cheered himself up by completing a first pass of analysis of all R&D's past publications, so we can start to integrate them in the new R&D website CMS and look into making them more browsable and searchable. Meanwhile, Theo is magically designing with real content inside the CMS, and it's starting to look quite nice. Sean and Dan have both been learning more about the Locomotive CMS which is going to power the new R&D site. Sean's also been reviewing candidates for the job of Senior Software Engineer on the Snippets team. Again, it's a very good selection. He says it's going to be tough deciding based on their applications - which is a nice problem to have.

Michael had a successful kick-off meeting with an indie about a music matching prototype and Libby's been preparing for the VistaTV workshop on iPlayer data which is in a couple of weeks. There are 25 people coming so far so that's all good - Olivier and Tristan have been helping her to plan it. Libby's also launched her twitter-bot including a .

Finally, George, Yves and Dom, along with Mark from the World Service, went to the to demonstrate our work on the World Service ABC project and .

It went down very well - Yves presented two papers, and the World Service project featured in at IBC (daily roundup of interesting things.)

Radio DNS was featured in the and George was about this, the World Service work, and our overall approach to R&D.

George hurried from the IBC to a meeting about the next phase of the FI-CONTENT project and then came back to umpteen other things, all too dull to mention.