BBC R&D

Posted by George Wright on , last updated

At the beginning of the week, there's a big spreadsheet printed out on the wall, showing the Second Screen user test results. The team have a phone conference to work out how the triallists are going to be recruited and managed.

Theo's working on some of the presentation elements for a trip to see the production teams for Autumnwatch. Chris N makes a lot of progress on the Firehose project, learning a lot about Hadoop for map/reduce data processing, how to work with EC2, S3 and AWS, as well as beginning to plan the architecture and platform. Tony gives a talk on periphony at the Techcon at the Radio Academy's radio festival in Salford. It goes well and the Head of Technology for BBC Audio and Music references his talk in an interview for "PM" on Monday, which is a Good Thing.

Paul and I spend some time working on a new collaborative project proposal - some last minute niggles around bureacracy, but Dom gets it submitted OK and we're pleased with it. Paul and I are in a room together again for some interviewing, whilst Chris G is trying to get his head around EU project tracking methodologies. A lot of the team are off to Bristol on Thu to spend the day with the Autumnwatch production team, which goes really well. There's always an interesting challenge merging R&D with linear TX timelines, but we think we've managed to work something out.

Sam works on P2P-Next handover and work on metadata, etc. We have a breakage in the STBs that University of Lancaster are running user trials with - there's some back and forth on email but we find the (upstream) bug and then just wait for the DNS changes to propagate. Always a fun task.

Sam's also looking into the Mozilla audio API, and thinking about what web browsers will be *for* in five years time. I'm tempted by an interesting EBU event but can't go, so it'll probably be Paul. Makes up for the Service Assurance Technology Forum he has to attend in my place.. Tristan's confirmed as a speaker at what looks like an excellent November W3C event around TV and the Web

Tony's been trying to interpret the x, y & z raw acceleration data provided by the wii and convert it to a tempo based on a conductor's movement. The tricky bit is filtering and differentiating the data to account for the different ways different users handle the wiimote as a baton.

Mark manages (through a combination of the (limited) documentation, tutorials, some help from Chris N, and use of Python's dir() function) to get the BBC radio WMA streams into Python via GStreamer. This allows the AudioSync server application to listen to live broadcasts (albeit subject to stream buffering, etc.). I'm excited by this and have got a slot for a early stage demo. We're making progress on RadioDNS and have gathered a big list of things we need to address.

I act as brutal editor on a large document that has had input from 20 people from outside the BBC, and am pleased that the changes seem to have been approved. I compile a list of my top 3 telco muzak providers, which says something about the way I'm spending a lot of my time at the moment.

We've been reading a Report on the 2screen event last week that Kat attended (her report is just as good but can't link to it, hmmm, maybe it should be a blog post) along with an In-depth review of Google TV. And now we're done.