Studio of Objects Wrap Up

Monday, September 21, 2015 at 19:48:52

Well with the Paolozzi project (Studio of Objects) finally finished, the start of the summer saw the final wrap up “event”. This was held at the suave and swanky Institute of Contemporary Arts on Pall Mall with views over Horse Guards Parade, the London Eye and, well, just about everything in central London. The beta trial version of the app with 3D navigable studio and links through “portals” of content about the art and life history of Paolozzi, along with recordings of him, was available on a number of iPads that were being demoed by some young volunteers. There was also live screencasting via a projector on to the wall (see photo). Dr Chris Horrocks (Kingston lead) gave an august introduction to the evening before drinks and nibbles led to deep conversation for all.

Some fantastic feedback from users on the beta and a hugely successful launch.

Long exposure image banding

Friday, September 18, 2015 at 14:28:26

I quite often experiment with neutral density (ND) filters for long exposure photography as part of images I produce and, over the last couple of years, have had two particular problems (neither of which I have examples of any more so I’ve pulled in some links from Google Images):

1. Cross-banding: this occurs with many variable ND filters which comprise stacked circular and linear polarisers. When you rotate the outer ring you increase the density of the glass and so reduce the amount of light. However when you overcook the density then you can get this cross-banding. Easy to fix (reduce the effect!) and easy to work out where it comes from.

2. Linear Banding: this has occurred on and off on occasion and usually when I am shooting in sunlight. My initial reaction was that this was caused by leakage of light through the filter system itself (I use Lee Filters), however no amount of fiddling with that helped it. More by accident (and then some subsequent Googling) I covered the eyepiece on my Nikon D700 (in fact most Nikon DSLR have an eyepiece shutter) which instantly solved it. Clearly a small amount of light entering through the eyepiece and, even though the mirror was up, it was leaking on to the sensor. As David du Chemin notes, its a very strange (and irregular) phenomenon!

DJI announce micro four thirds camera

Friday, September 11, 2015 at 14:16:18

DJI have announced a micro four thirds camera for their drones which seems to be specifically targeted at the film industry but has important implications for photogrammetry. In the past the GoPro3 (6.2x4.7mm) has been used which is generally poor for 3D reconstruction with a (very) small sensor and wide field of view. The Phantom 3 has significantly improved the camera quality by using a higher resolution 2/3” sensor (8.8x6.6mm), but this is still fairly small by SLR standards. The introduction of the X5 is a big improvement again at 18mmx13.5mm with, crucially, interchangeable lenses meaning you can go wide angle but have less compromise on the optics. This part of the consumer market is a very interesting space to watch.

Mapping the Syrian Conflict

Thursday, September 10, 2015 at 08:31:52

An interesting short piece over at the BBC about mapping the Syrian conflict - what is interesting is the BBCs lack of understanding about how data, (online) mapping and journalism are merging together to create a new generation of (often (remote) citizen) data journalists. The first strap line on the clip is:

“A new generation of cartographers are using social media to map the Syrian conflict”

But actually this misses a large part of the change in reporting - just look at the well regarded data journalism at The Guardian, and the related article explaining data journalism. Simon Rogers (from The Guardian) has written “Facts are Sacred”, whilst you also have The Data Journalism Handbook and Data Journalism. This is a hot ticket at the moment and has come out of the rise in open data revolution through government sites such as (and, for that matter, the leak of databases such as Edward Snowden through Wikileaks and Ashley Madison for high profile examples).

In addition to the data, we also have (increasingly free) tools to collaboratively explore, analyse and visualise these datasets (see The Guardians list for example) and the increasing profile of data visualisation and infographics (Knowledge is Beautiful for example).

Cartography (and GI Science) clearly has an important role to play in the development of this area in terms of data manipulation, visualisation (particularly online and dynamic) and, in particular, spatial data. But let’s not confuse what is the convergence and development of existing subjects and technologies that service society within the narrow constrictures of the reporting shown here by the BBC.

MCQ Best Practice

Wednesday, September 9, 2015 at 13:25:40

Following up on the last blog post, Donald also has a great post on the 10 stupidest mistakes in multiple choice question design. Everyone ends up doing MCQs at some point simply because they are easy to set, provide rapid feedback and marking can be automated (or largely automated). But, in order to boost learning you need to design them well - the first and most important is “test understanding, not facts”. Good one, but it takes effort to do this and we are all guilty of testing facts at times.

There are two are posts worth looking at in conjunction with this. The first is “10 essential points on use of open-response questions”. Open response questions can be far more effective than MCQs, yet in many situations you can still automate marking and provide detailed feedback.

And finally….. All of the Above - 20 ways to cheat Multiple Choice questions. Given how poor many tests are, use this tips to either write better tests or get better results!

Textual Double Header

Friday, September 4, 2015 at 08:35:38

Donald Clark did a good textual double header last week….. the first piece was on why text actually makes for good learning. Two 10 reasons including its fast, leaves the learner in control and is searchable. Perhaps the big downside (not mentioned) is that it requires effort, i.e. you have to WANT to read to do it, making it an ACTIVE learning process. That is perhaps one of the reasons that it is so effective - people have to engage to do it. Video content, in contrast, is PASSIVE so there will be a significant proportion of people who will watch but NOT engage.

Which leads quite neatly on to his second piece - top 10 tips for WRITING online content. These ARE good and are designed to ENGAGE learners. They include CUT CUT CUT (less is more), using images, stab points, key words and meaningful headings.

For learners and writers alike, these two articles are a great starting point.

EA DEMs Data Precision

Thursday, September 3, 2015 at 07:32:56

UPDATE: An illuminating and quite hilarious read over here. The release of the LiDAR sourced EA DEMs is garnering some nice attention around the world although Martin spotted that the elevation values are recorded to a precision of attometers. Yes that’s 15 to 17 digits after the decimal point!! As Martin points out, this is a huge waste of bandwidth and, I imagine, a default export setting. As I noted yesterday, the interface is functional, if a little clunky, and could do with some improvement.


Wednesday, September 2, 2015 at 08:29:34

As I noted earlier, the EA has now made its 50cm DEM data of the UK available at You can find both the DSM and DTM 50cm products…. From either page you’ll eventually get to this download page which is a touch slow but works well enough where you can download fairly small OS tiles (shame they cant be bigger). Ill be playing with these over the next few weeks to see what they are like.

Top of Page