Well friends, this afternoon I sat down at my desk fully intending to do some work on the next chapter of my thesis, I went to brush my fringe out of my face and instead poked myself RIGHT in the eye. As result I think this painful incident excuses me from doing any real work today…and also has me questioning whether I actually have the IQ to write a thesis in the first place… #naturalselection
Anyway, I’ve decided to take a bit of time to write a blog entry instead as it’s been a while since I’ve written anything just for the fun of it! And currently the thought of making a start on writing up my next thesis chapter makes me feel like this….
I get a lot of enquiries about the software I use in my work so I thought I’d take a little time to write a post explaining what I work with and why.
First of all it might be helpful if I explain why I work with digital media for visualisation rather than more traditional methods. As an archaeologist it’s essential that your methods allow for flexibility in the interpretive process and digital mediums mean that edits and experimental changes can be done quickly and intuitively. Fundamentally though it’s a medium I enjoy working in and one which I find both challenging and satisfying. That may seem like a very simple justification but I think being passionate about your work is hugely important. I’ve worked a lot with more traditional mediums, hand drawing, painting, comics, though I always return to digital artistry as my tool of choice.
Over the past few years I’ve done a lot of work with laser scan and photogrammetry datasets. Before these can be incorporated into any reconstructed scene or model they need to be processed. The first stage in this process is usually to clean up some of the raw data point clouds, for example chopping out the blurred point trail of a low flying seagull or a passer by….or far more likely me, see previous post. Depending on the type of survey I then need to register the data. In order to get decent coverage of a site when scanning a structure the scanner is moved to various positions and numerous 360◦ scans are taken. If not done via a traverse (a method of scanning over large areas which uses targets and tripod heights to allow the scanner to place each scan correctly) multiple scans then need to be merged to create a complete model, and this process is called registration.
I usually work in a bit of software called Cyclone for the point cloud part of the process, before creating a solid model of the site in Polyworks through a process called meshing which works by joining three adjacent points together to create a surface. At the meshing stage a little cleaning up is usually required to fix any mangled edges produced through the meshing algorithms and for this I usually use a great piece of freeware called Meshlab.
Then comes the fun part! Once I’ve got myself a decent cleaned up model of the site I can convert it to a compatible format and take it into 3ds Max which is where the magic happens! Here I model the reconstructed elements (architectural elements and artefacts etc), add materials, texture, lighting and atmospheric/environmental elements like fire, smoke or haze.
Laser scans are great for capturing the aesthetic essence of a site. One of the drawbacks of software like 3ds Max is that it’s very architecturally focussed and is very good at straight lines and right angles, but not so great at intuitively modelling more organic shapes and textures, which is why for a lot of the ‘final touches’ I turn to a brilliant piece of modelling and texturing software called Mudbox. Mudbox operates along the concept that the model you import into the software is like a big lump of clay that can be squished and patterned into shape. Unlike 3ds Max the texturing tools in Mudbox can be ‘spray painted’ directly onto the surface rather than draped over a surface using UVW mapping.
Though I can’t say I use giant toads in my reconstruction work very often, I do use human and animal characters in a lot of the scenes I visualise and for that I use Poser Pro 2012. In the CGI industry people spend their whole careers specialising in character modelling and rigging – unfortunately I have to do everything myself as I don’t have an Avatar-sized crew of digital artists working for me (yet, ha!!) so to save time Poser Pro has pre-rigged character models which means they’re ready to be animated without hours of painstaking work.
One of the issues I’ve had with Poser Pro is that the hair tends to be very well groomed, enter a nifty piece of software called Hair Farm which is a plugin for 3ds Max which allows for much more detailed (and more importantly, messy) styling of hair and fur.
Working with a lot of tools and plugins developed for the digital art/film industry and not really with archaeologists in mind means that often I’m having to work out how to achieve what I want to with the tools available and not a lot of examples or help available. It’s difficult to make CGI look dirty, rough and natural!
I’ve been working on a project on the side of my PhD research for the past few months which requires the generation of a fairly vast vegetated landscape and although I’ve used Vue in the past, this time I’ve been working with a plugin called Multiscatter to distribute iPlants and Xfrog vegetation. I usually leave vegetation and atmosphere to the very last stage as both kill render time and slow my pc way down!
Once I’m relatively happy with a scene I’ll then take the rendered result into either Photoshop (if it’s just one still image) or After Effects (for animations) to apply the final touches. This usually involves tweaking the lighting and contrast a little, patching up any unwanted artefacts in the image (artefacts in the sense of unwanted light spots of splotches where a rendering algorithm may have messed up a little, not in the archaeological sense of the word artefact!) or adding in effects such as fire.
Throughout the entire modelling process I keep a reconstruction diary to keep track of creative decisions made along the way. For example if I make a decision to model a roof or rebuild a section of walling in a particular way based on a paper or report I’ve read I’ll keep track. Similarly if I’ve had an interesting debate about how to represent a scene I’ll bullet point the key arguments and reasoning for or against the interpretation I adopt.
For example for the visualisation of House 7 at Skara Brae I’d chosen to represent artefacts associated with ritual in the foreground, with more domestic items to the background. These decisions were made based on the fact that archaeologically House 7 stands out from the rest of the village and seems to have been treated very differently to the other dwellings. Physically House 7 is peripheral to the other houses as it can only be reached by an uncomfortable crawl down a narrow passageway, descending into darkness. House 7 also had controlled entry in that it is possible to secure the door from a chamber outside the structure, ultimately locking people inside. Furthermore, there were two female burials under the right hand side bed, made when the foundations of the house were laid – a determining factor in choosing to represent the seated character as a female. I think it’s important to note that with creative work like this it would be very easy to create a fiction about the site, but I don’t because I have to stay true to the archaeological record and respond to the evidence.
Depending on the project, and whether it’s a still or animation this process can take anywhere from a few days to a couple of months – but it’s a lot of fun and always a challenge I love it.
Well, I should probably get back to writing my thesis…