Last week, Nick Kraus and I had the pleasure of travelling to Chicago on a Field Museum visiting scholarship! Way back in October, Nick and I submitted an application in collaboration with Anthropology curator William Parkinson, to take photos and generate 3D models of the Field Museum’s cuneiform tablet collection using the Ortery 3D PhotoBench 260 system.
The ultimate goal of our work at the Field Museum was to produce 3D models of the cuneiform tablets that would eventually be hosted right here on the From Stone to Screen website. The reason we chose cuneiform tablets for this pilot study was threefold:
1) cuneiform tablets are often plentiful in museum and research collections, and are both durable and portable, allowing for easy future replication of our workflow;
2) Nick is currently pursuing his PhD in Assyriology at Yale, and so has expertise in working with these objects (this came in handy when deciding which tablets to choose!); and
3) while many objects can be viewed and studied through photographs alone, objects like cuneiform tablets require a three dimensional view for comprehensive research – they often have writing on all sides, and the ability to rotate a tablet is paramount to its interpretation. Photographs can only go so far towards understanding and reading these objects, so they are a perfect candidate for 3D imaging!
The Field Museum currently houses about 80 cuneiform tablets – which seems like a lot compared to UBC’s 6 tablets…but pales in comparison to collections like Yale’s, of about 40,000! Of these 80 tablets, around 60 are currently uploaded to the CDLI. On our first day at the Field Museum, we spent the morning learning the ins and outs of the Ortery equipment from Sean Bober (that equipment gave us some real start-up issues, as technology tends to do!) and once we had a few test objects completed, Nick selected around 30 tablets for us to scan that spanned a good chronological range, from the Old Akkadian to the Neo Babylonian period.
With the help of Collections Manager Jamie Lewis, we spent all day Tuesday, and part of Wednesday morning imaging these tablets and marvelling at the ease of the Ortery set up….until we realized that we’d been doing it all wrong! We should have known that everything was going too smoothly. Essentially, because we wanted images of all sides of the objects, we had selected a ‘spherical’ setting on the machine, which allows us to take a pre-programmed amount of photos on one side of the object and then prompts us to flip the object over so that the machine can take the same number of photos on the other side (you wouldn’t need to do this for an object that stands up on its own, like a statue, for example). The two halves are merged together later in the post-processing stage (with a program like Agisoft PhotoScan). The problem is, because we were flipping the objects horizontally (top-to-bottom) and not vertically (end-to-end), the data encoded in the photos would not allow the two halves to match up. If you’re worried about this happening to you when using the Ortery setup up, have no fear! There is a handy diagram that shows you exactly how to flip your object…just make sure you pay attention to it!
We spent the rest of Wednesday taking the images (correctly!) of the remaining tablets, and decided not to re-image any of the ones we’d done incorrectly until we could see whether there was an easy fix at the post-processing stage. That night, I ran one of the incorrectly-imaged tablets through Agisoft PhotoScan to see if somehow technology might magically come to our rescue…but it did not. As we thought, the two halves did not match up at all (and I ran the images both as one chunk, and then again as two separate chunks and tried to merge them). Unfortunately, because I only had my laptop with me, this processing took several hours, so I let the program run all night and woke up to disappointing yet expected results on Thursday morning. At that point, I thought it would be worthwhile to run a correctly-executed set of images through PhotoScan to make sure that the images we had redone wouldn’t present us with any problems! I started the processing, but had to leave the computer running while I left for the day – I spent Thursday at the University of Illinois at Chicago, while Nick and Jamie valiantly scanned all of the rest of the tablets, and re-did all of our incorrectly-imaged tablets too! They put in an impressive amount of work while I wasn’t there to bug them about posing for pictures that I could insert into tweets!
We were all feeling pretty proud of ourselves – we’d surpassed our goal for the number of tablets that we wanted to image, and that was including 17 tablets that we had to redo! If we hadn’t run into our flipping error, we probably could have imaged close to the entirety of the Field Museum collection. That night, when we returned to our hotel, Nick and I were already planning on taking the next morning off to see the Terracotta Warriors featured exhibit at the Field Museum – surely we’d earned it? And then I saw my computer…or, to be more specific, the attempt at a 3D model of a ‘correctly’ done tablet on my computer. Not only did the two halves not match, but the quality was extremely poor and grainy, and certainly not at all comparable to a regular old photograph! Now, I am no photogrammetry expert, but I have spent my fair share of time with PhotoScan, and I knew that this was NOT the result that we were hoping for. Unfortunately, I was at a loss for how we could improve our post-processing results. We needed help!
Luckily, the next morning we ran into Kate Webbink in line for coffee (the great unifier)! Kate works in the Technology Department at the Field Museum and had been with us during the Ortery equipment training. She asked how our week had been and I explained the difficulties I’d been having in trying to generate a 3D model, to which she suggested that she put us in touch with JP Brown, Associate Conservator at the Field Museum and resident photogrammetry expert…Resident. Photogrammetry. Expert!! Luckily JP was available that morning, and together we went through my workflow on PhotoScan…and he produced largely the same quality of results that I did. This was good for my ego, but very, very bad in terms of the quality of all of the images that we had taken thus far. Since this was our last day and we were working to beat the clock (we had tickets to the Cubs game at Wrigley Field that afternoon!), we raced back upstairs to the Ortery equipment and Jamie pulled the tablets back out of storage. We had just enough time to scan one tablet, and we drastically altered several components of our imaging process in the hopes that these changes would produce a workable 3D model: first, we adjusted the zoom significantly so that the frame consisted largely of the object and had minimal white space; second, we placed a black cloth beneath the tablet because the Ortery’s turntable is transparent, and the reflection of the object can cause processing problems; and third, we adjusted the number of frames per hemisphere from 12 to 24 (generated from cameras, for a total of 120 total images per side instead of 60). We had just enough time to transfer the files to our portable hard drive before we had to leave for the day – but I still haven’t been able to bring myself to process these images in PhotoScan because I’m so worried that they won’t produce a useable 3D model.
So, cross your fingers that at the very least when I finally muster up the courage to run these images, we’ll have a replicable workflow for the future that we can share with anyone else embarking upon a similar project! All was not lost, though: Nick and I got to spend an amazing week in the Windy City, and met some of the wonderful, generous people working at the Field Museum. Nick had the opportunity to study some of the tablets he is using in his dissertation firsthand, and we both learned the basics of the Ortery 3D Photobench 260 system, which is some very impressive equipment to work with. The Ortery also generates .html viewable ‘models’ of the images, so we have (at the very least) rough working versions of the tablets that can be somewhat manipulated. We learned a lot during our week in Chicago, and overall had a great time representing From Stone to Screen!
Oh…and the Cubs won!