Happy New Year! To start off 2022, I’ve started adding some storage items to the longhouse model.
To begin with, I’m just using some colourful primitives as stand-in geometry. Eventually this will be replaced by models of storage items–like dried plants–that other artists have created.
First I added a couple of overhead storage poles in the vestibules of the buildings. Then I combined those with the interior storage poles that already existed, and populated them with random collections of storage items. Here are some results, with different longhouse sizes and configurations:
Once we replace those candy-coloured primitives with more realistic models, the longhouses will start to feel like real lived-in spaces!
Here are a couple of views showing how the interior might appear; one from outside, one from within the vestibule. I’ve used a wide-angle lens here to make it easier to see the surrounding structure:
Now that the longhouse model is getting incorporated into the game environment, it’s time to start adding a few more details.
I started with a supply of small logs under the bunks. I haven’t checked the reference material, but I assume this is firewood storage:
Next, I added the flat interior walls that separate the vestibules on each side from the central living area. These used a similar approach to the exterior walls: laying rectangular sheets of bark over each other in layers. The difference is the outer border needed to be trimmed to fit inside the shell of the building. The interior walls are most easily seen if we make the ground invisible and look up at the building from below:
Finally, I added some rope lashings to hold together the interior framework. I used a similar approach to the lashing on the palisade asset. This is an ‘in progress’ version, since there are some complex attachments that make it difficult to make each turn of rope behave properly. Still, it’s a useful start!
The next step is to start populating the interior of the building with props that have been created by other artists. Adding some hanging dried food, tools etc. should start making this feel like a lived-in home instead of an empty shell!
Building on Kris Howald’s work creating an accurate terrain / landscape appropriate for the time period we looked to the final Jean Baptiste Laine report as our basis for the layout of the city itself. Using the late phase floorplan, placeholder longhouse models were positioned in approximate locations relative to each other, the palisade, and the surrounding environment. Sizing was also an approximation, longhouse structures were adjusted in Unreal through changes in scale rather than in the model itself. The effect of this is a best fit “stretching and squashing” of the model and not a true structural representation of length. The single green longhouse is an example of the final longhouse model.
Working closely with subject experts we imagined the agricultural fields as wide ranging and expansive, appropriate to support the large population of Jean Baptiste Laine. Also included were paths of travel crisscrossing the fields, allowing the user clear areas that are explorable. This also had the added benefit of reducing the number of models in the level. The tree-line in the distance represents the extent of the fields as well as the extent of the player’s potential travel area.
Unreal engine’s foliage tool was used to place 13,500 models of the ‘three sisters’. Multiple levels of detail are generated in real-time to account for the large number of models – closer to the camera equals a higher resolution, farther means significantly reduced detail.
The palisade wall was created programmatically, by Youhan Guan, along a spline with single palisade posts placed on both sides to interrupt immersion breaking model repetition.
Particle based local fog was added to the interior of the explorable longhouse to create a sense of dust, smoke and low light conditions. The longhouse model and interior was taken from version 4 of the project created by Kris Howald. The current version also incorporates the colour palette and texture style he established.
With temporary longhouse models in place, it’s clear how crowded the interior city space could have potentially been. Given the number of occupants, we interpreted the ground as hard packed with patches of foliage. Grass was placed in areas away from potential travel such as along the base of longhouses and areas away from longhouse entrances.
An open question is whether the environment should be more meadow-like as opposed to agricultural. This would mean the inclusion of a variety of additional foliage types, though this might add to the accuracy of the experience, the cost to frame rate could be high.
Initially, the landscape was described to be more of an approximation. However, as the project progressed it became clear that the accuracy needed to be there. Finding topological maps of the area of Stowe Ville was somewhat of a challenge, to begin with I thought I could try to use historical aerial maps, but those were hard to come by.
There was a great little website that has somewhat accurate temperature-based height maps of the entire area encompassing all three of the sites (https://en-ca.topographic-map.com/maps/qeu/Whitchurch-Stouffville/). From these heightmaps, I was able to approximate sea level elevations for all of the high and low points. I started by looking at a larger scale map, which was a 5km x 9km area.
From the heat maps, I had to convert everything to grayscale so that it could be read in as a proper heightmap into the world-building software I used to generate my final textures for the Unreal Engine. I realized that the 5km x 9km dimensions were far too big to maintain the level of detail we would need for the much smaller area of the sites. I tried a 2km x 3km but ended up settling on a 2km x 2km scale. I also discovered that I needed to exaggerate the darkness of the river areas so that that detail wasn’t lost when adding things like erosion. To ensure that I was using the correct scale and that the village was placed correctly, I overlaid it in Photoshop to precisely the right size. I also brought this 2km x 2km map into Maya, where I then generated a template of the village that will be later used to place the longhouses in the Unreal Engine.
To generate the surface is I used World Creator (https://www.world-creator.com/) at first, but I was struggling to get the quality I was looking for. I decided to change gears and use Gaea (https://quadspinner.com/) which is a newer software but was able to give me what I was looking for.
Matching scale as precisely as possible was important because I wasn’t sure how much work would need to be done in Maya and then transferred to Unreal. So, I exported the geometry from Gaea to Maya to make sure it was exactly 2km x 2km. And then from Maya I sent the environment to Unreal to use a guide when generating an Unreal Environment, which has benefits such as distance-based LODs so that the geometry would become reduced the further it got from the viewer. A static mesh couldn’t do this.
Getting the Unreal environment to match was fairly easy because 1px = 1 meter. So my map resolution of 2017 x 2017 equalled 2.017 km by 2.017 km. This was slightly larger than I needed but after playing with the values I was able to nudge down the scale (99.5 in both X and Y) to exactly the proper dimensions. The height was trickier but a value of 16 seemed to do the trick.
I used Substance Alchemist to created a palette of tileable textures, that I then brought into Substance Painter to give them the painterly look. That I then used in Unreal to custom paint the ground around the Mantel Site.
I’m a bit behind on my posts due to the start of my school year. So prepare for a massive release of posts from the past month and a half. This first post highlights some experimentation I was doing with cards to populate the forests. Because they’re only 2D, they are to be used from a distance. I’ve created packs, where there are a number of trees and bushes per 1 texture file, and I’m using UV placement to split them out to the individual cards. The textures were created in Photoshop, and I did my initial testing in Maya.
I played with low res cards on basic geometry to create the 3D trees. This approach could be refined further, but seems to work with the style at the moment.
Below is a render test from Maya to test the parallax effect from layering and offsetting the planes.
Most recently, I generated a quick low poly patch of grass to replace the canned grass that was being tested out.
After a hiatus to get the fall term started, I finally had a chance to revisit the longhouse asset.
I wasn’t happy with the previous approach to the bark skin. The sheets were placed too haphazardly, due to random point placement in the digital asset. Looking at photos of reconstructions, it seemed to me that the placement should be more regular.
To recap, here’s what the original placement of bark ‘skin’ sheets looked like:
No rhyme or reason to the placement.
In aiming to make the pattern more regular, the easy part was to deal with the long parallel walls. It was relatively straightforward to lay out rows of sheets with some minor variations:
This works much better, although as you can see I’m starting to have a problem with sheets clipping through their neighbours. This problem is most apparent where the sheets start to curve more sharply approaching the roof. I’ll have to revisit and refine this.
Continuing this pattern along the rounded vestibule structures at each end proved to be much more challenging. The question of how to efficiently cover a spherical form with rectangular shapes is a non-trivial one, as cartographers mapping the globe and kindergarteners working with papier-maché both know!
I wrestled with this for a long time before realizing that my problem was that I was trying to think too procedurally. What I needed to do was to take a slightly more hands-on approach and divide each vestibule quadrant up into sections which I could treat individually. A single algorithm wouldn’t work well for the entire area.
Here’s how I ended up dividing the quadrant into smaller, more manageable sections:
As you can see, giving each section a set of parallel “ribs” makes it much easier to analyze how the bark sheets should be laid out to efficiently cover this irregular form. Hindsight is 20/20, and now I wish I’d thought to take this approach much earlier!
There is still some refinement needed, as certain dimensions of the structure can produce some unwanted results, such as “ribs” at the edge of one section sitting too close to those from the next section. But as a starting point, I’m quite happy with this.
Here’s how it looks with the bark sheets laid out according to these new patterns:
This looks to me much more like a structure built by careful and thoughtful human engineers–rather than a dumb machine throwing random points around!
Again, there are some problem areas that will need to be addressed (like the centerline, and the top of the roof). There is also still the problem of some sheets clipping through their neighbours. But I feel that these are solvable problems and the base approach here is a solid one.
I ended up creating a variety of pipes based off of the four main shapes that I initially modeled. The different decorations were all inspired by the images from Guided by Smoke: A Comparative Analysis of Early Late Woodland Guided by Smoke: A Comparative Analysis of Early Late Woodland Smoking Pipes from the Arkona Cluster Smoking Pipes from the Arkona https://ir.lib.uwo.ca/cgi/viewcontent.cgi?article=7843&context=etd
I created the base club and three variations in Maya and Zbrush using references provided by the research team.
Retopologizing makes a simpler version of the club geometry.
The last step I completed was texturing the clubs in Substance Painter using a provided smart material and painting in some details by hand. I added some different wood textures to make each club unique and to give them more visual appeal.
The three sisters! I loved learning about the three sisters – squash, corn and beans, the core of First nation farming. I was aware of the practice of rotating certain crops to replenish soil nutrition (nitrogen fixing bacteria in legume plants) but this was the first time I learnt of companion planting and how they support each other in more ways than nutrition. Even more intriguing was the legend or traditional story of these 3 sisters.
Apparently the story of the three Sisters varies from tribe to tribe and this version is from an oral account by Lois Thomas of Cornwall Island found in the ‘Indian Legends of Eastern Canada’.
Once upon a time there were three sisters who lived together in a field. These sisters were quite different from one another in their size, shape and way of dressing. One of the three was a little sister, so young that she could only crawl at first, and if she wanted to stand up she had to twine herself around her eldest sister. This sister wore velvet green with delicate tendril ribbons. The second of the three sisters, wore a frock of bright yellow and had a way of running off across the field when the sun shone and the soft wind blew in her face. The third sister was the eldest. She was always standing very straight and tall above the other sisters trying to guard them …
It is clear that the eldest sister is a corn plant, the second sister squash, and the youngest a bean plant. The story goes on to share how they feed a young Iroquois boy who came to visit them. They all lived together and supported one another. Just like the story the three plants are grown together in what is called a ‘mound’.
The height of the corn supports the bean vines, which tie the corn stalks together for added stability. Beans fix nitrogen, which means they take it from the air and transfer it to the soil, making it an available nutrient. The beans feed the corn and squash, which are heavy nitrogen users. The large leaves of the squash plants act as a groundcover to provide shade, conserve moisture, and suppress weeds. (https://www.naturespath.com/en-ca/blog/planting-a-three-sisters-organic-garden/). Researching these weren’t related to my task in modelling but I was so intrigued I delved into it and had to share here!
So I started with the task of first modelling the three sisters and its variations. I did the squash first. I borrowed some gourds from Sonia and textured them in Zbrush to add the bumps found on squashes. I used the alpha again but changed the size to give it irregularity. The alphas I used were 23, 32, and 07 for the bumps and alpha 22 for the vein texture.
For the corn I took a simple dynamesh sphere, stretched it and drew the kernels on with a dam standard brush. I used the ‘snakehook’ brush to fashion leaves out of additional spheres and textured it using the ‘rake’ brush.
The last were the beans. I googled some images of heritage beans to get a good reference for modelling. Then made simple variations of my base model. Additionally I made single dried bean seeds and that is how they would be eaten also.
Unfortunately I could not get to the mounds yet. It will definitely be a challenge to create something that will be low in poly but be repeated enough to make a field.
The bow and quiver was also fun to do but maybe a bit trickier than the others. This is because the reference image provided below has the bow cover and the quiver attached.
They were blocked out in Maya and transferred to Zbrush for detailing. Since all extrusion of surfaces and hollowing out were done in Maya it wasn’t a problem to detail in Zbrush. Similar to the spears, I used alphas to give the rough texture of leather, and used Alpha 60 with the dam standard (Zsub) to add the scratches found in natural leather.
I reimagined some string all along the side of what I’m assuming is the bow cover (without the arrows) kind of like shoe laces, for easy opening and closing of the cover. The strings also allow different sized bows to fit in it. Although the reference here seems to be made of suede leather, I made the quiver a bit thicker to mimic thicker hide/leather for durability. For one of the variations I made the bow cover and quiver separate with separate handles.