It’s been a busy couple of days for the project and Craig, who is still down at Siggraph2015 demonstrating at the Unity booth. As you can see our project is getting a workout in the Unity 5 gaming engine! Craig’s ported the Autodesk Maya 3D models at their current stage of development into Unity 5. Even with the models in basic form Unity’s new lighting is the bomb! We’ve also been fessing about the polygon count on the models being built as we have been trying to keep them highly detailed from an archaeological research perspective, but it looks like Unity can handle it so far.
I asked Craig to also film a quick video of the Longhouse in Unity and although it’s a low quality cell phone recording, I’m really happy with the progress!
From a research perspective, it’s been a good study on the influences the process, technology and artistic craft has on the researcher when interpreting the archaeological record in virtual space. Both Sara Perry and Grahame Earl, who I spoke about briefly in Longhouse 3.1, talk about the value of this new form of archaeological illustration as being transformative and going beyond just illustrating or visualizing archaeology. The ever expanding methods and growing theories of Virtual Archaeology are helping to shape how we see archaeological material and in doing so is spawning questions researchers would have never considered. There is also a tension that 3D animation and virtual reality is “less than” traditional archaeological study, but hopefully as we continue with this project, it is demonstrating an active and valid use of traditional method and theory.
As such, this project has been following the guidelines outlined in the The London Charter for the Computer-based Visualisation of Cultural Heritage, conceived in 2006 as a means to encourage intellectual transparency in the creation and use of technology within cultural heritage studies. Essentially the Charter lays out the methods researchers should take to ensure that the digital material they create has followed academic rigour and if assumptions are made during the process (as we have clearly stated in earlier posts) that they are recorded and justified to ensure authenticity of the research itself and transparency so that the public, who are now served by experiencing these digital environments, understands that what they see is an interpretation and not absolute fact.
Lastly, even in the process of blogging, you the audience is experiencing the pains, decisions, failures and successes of the process to (re)imagine extant archaeological material. Generally this reflection comes after the project has been completed, but the Charter has indicated that it is a worthwhile effort to engage the digital process as it happens.
Carrying on from Longhouse 3.1, we were exploring the relationship of the interior supporting structure and how the exterior wall posts would be bent and attached. Craig’s quick revisions helped to better understand how Iroquoian builders might have forced the wall posts to bend at certain points to create a continuous arbour roof and wall system. It seems plausible that poles that would run lengthwise and supported by the main bunking infrastructure would be the first attachment spot for exterior wall posts. 
A second set of lengthwise poles another 4 to 5 feet higher and supported by the main interior posts would then act as part of the interior rafters and another connection point for the exterior wall posts. The interesting bit is what they did at the termination point or where two sets of poles would meet in the middle? We have no visual or written historical material to go by in determining how the very top of the roof was constructed, but it was clear through written history that longhouses had smoke holes at the very top of the roof to allow for smoke to dissipate. Assuming that Iroquoian Longhouse construction was linear in fashion (i.e. the smoke holes weren’t rounded), the very top of the roof would have been rectangular in shape to allow for square’ish smoke holes.
In a second set of revision notes, you can see that we’re now starting to play with the idea of a flattish roof and how the opposite wall poles might have terminated. I used the concept of fingers from two hands interjoining, but even that raises questions about how smoke holes would then be inserted. Would they cut the ends of the poles while on top of the longhouse, or pre-cut them before they were installed? At a starting taper of 3cm’s at the base of the pole, I would assume that after 7.m’s in length, the end of the pole would be easy to break off with the bare hands? Or did they use a saw of some sort (which I will need to check with the archaeological record)?
In addition to creating a flat roof, we also need to be cognizant of making sure our 3D objects reflect academic thinking. One of the quick methods of modelling is just to use a simple proxy (a tube for instance) and just make it the length of the longhouse to represent a support pole. However, as we know from previous discussions, White ash was likely used for support poles as they tended to be much stronger than other wood types, would grow almost entirely straight and with a uniform taper but usually no longer than 12ft (in usable footage). Hence in a 24ft longhouse, there should be at least two 12ft poles laid out lengthwise, with a slight taper. Again, there is no archaeological or historical data to support the existence of additional support beams that might have attached to bunk infrastructure to further support the lengthwise posts, but when the initial test image was rendered, it looked or more appropriately, it felt, as though more supports were needed.
A lot of the detail we are talking about will likely be lost visually when the atmospherics like lighting is added. Longhouses had two major light sources; the doors on either end, which might have been fully or partially covered and the smoke holes above. Any secondary sources of light would be coming from gaps in the exterior shingles or the fire hearths themselves, thus the detail we are agonizing over would likely be entirely in the dark once the house is imported into the Unity gaming engine. Further, too much detail in the models slows down the real-time rendering of the objects in 3D space, so it is possible that the more realistic modelling, rounded ends, nice textures, will become more angular as we reduce the polygon surface to increase speed. All considerations to examine as we continue to build.
Next in Longhouse 3.2 we will review the changes and start to see the final infrastructure that will support the exterior walls and roofing system.
Addendum: August 17, 2015.
As an addition and in response to Bill Kennedy’s thoughtful observations, I’ve updated this blog with an image which I hope visualizes Bill’s practical experience building physical longhouses. I will respond directly below in the comments section.
I mistakenly deleted Bill Kennedy’s original comment on 08/17/15. Here it is in it’s entirety:
This looks good; clearly thought out with close attention to detail. I would make one suggestion. Add ties across the short axis to connect the opposing sides to each other. You have some ties shown (colored white) to connect your arcade purlins together. You have them shown on every arcade post, which is good. I would add the same thing at the lower level to connect the opposing walls to keep the posts stable and plumb. This is an easy change. Look where you have the red lines drawn and labeled “Need horizontal supports to hold up poles along the bunk line”: instead of making these separate, just make them be one continuous tie all the wall across the short axis. In a rigid framed structure, this would be absolutely mandatory or the outward thrust of the roof would push the walls out and collapse the structure (Trust me from having seen it happen multiple times in replica structures). In a flexed structure, these ties are arguably less critical but still necessary. If I were building this structure in the real world (I currently only build rigid frames, not flexed), I would put those ties in every spot where you’ve got a post connecting to your wall plate. It makes sense because not only does it stabilize your wall plates, it would be easier for prehistoric people to execute. As far as the rest of it goes, I think it’s pretty good. When I build benches, I have my bench slats turned perpendicular to the way you have them going but I can see arguments both ways and I think your way might work about as well. If I were building in the real world, your way seems like a long span and may sag but it depends as much on the thickness of the bench slats as anything else. If you have information or accounts that they would have gone parallel to the long axis, then I say stick with what you’ve got. I can tell you from experience that installing bench slats in any direction can be quite difficult since real wood is not perfectly straight and does not like to cooperate. Nice work.