This is my thread documenting my progress in the Collab2021 project -- more to come (soon!)
Was crazy busy at work this week so I haven't had any time to really work on this until today, and I'm crazy excited to get this ball rollin'! After contemplating water wheels all week I started blocking this morning with the paddle shape that I felt seemed the most reasonable.
The concept art has chonky paddles, but due to their dimensions in comparison to the other items I broke them down into a frame of boards with a thinner veneer style face. This seemed to be the most logical way to have wood sizes that weren't arbitrarily sized to meet the art, while still preserving the proportions.
For the outer wheel rings I initially didn't separate them out into boards, but as I started to write this I realized that they probably should be. When it comes time to texture, I think it will get awkward painting wood grain for something as large and circular as these rings without them being separated into boards. Meanwhile the core axel could probably be made from a tree trunk, or metal if the wood gets weird looking in there. But the outer ring certainly would look very untrue to the concept art if made metallic.
If you look at the concept art, you cannot truly see completely through the wheel. It heavily implies that there is wood on the back face, so I filled the one side with boards. I also think this gives it a lot more character.
Basic Cobbles blocked out.
Going to start busting up the silhouettes and pulling a bit more geometry out, bevels, etc. Going to try to get this submission ready tonight, but if all else fails, by tomorrow evening.
I was sad that I didn't really get time to work on this earlier in the week, so I'll end up with less feedback to work with, but, any feedback is welcome and would be greatly appreciated!
Houston, we have bevels. Time to break up the uniformity. Currently 4800 verts when everything is converted to single users.
Got most of my tweaking done. Just some work left on the bearing/axel and cobbles. Possibly rivets, but I feel like that could probably be done with bump and painting. They're not really visible in the concept and I think giving them geometry might be too much. They'd also be really easy to add later.
Did some low poly sculpting to make the cobbles less uniform. Going to take a break until tomorrow and assess what more I should do. I'd like to pull the center bearing out into wood planks as well, but I don't think I can afford the verts/tris.
My Vert count is higher than I would like: 11.8k if Single User, but only 4.4k if we're keeping Data Linking on. I can probably remove maybe 5-10 percent once UV is done as there's a lot of hiding faces, but with data sharing on I need it all still.
Great work, I like how you're breaking everything down and trying to figure out how everything is working. Good design choices and a very efficient model. Loving it. Remember to submit tomorrow and looking forward to see how it progresses next week.
Nice workflow you are showing here. Looks like we we had the same idea about the paddles :)
@lexicolopolis Brilliant solution for the paddles. Something that irked me about the concept art was that was water tipping out of the front facing buckets, but to my mind that could only happen with a "water over" paddle wheel and not the "water under" wheel as shown in the art. Your decision to make the paddles into boxes solves both the mechanics and the artwork in one fell swoop. We can now say that rather than having water tipping out of a full bucket, the water is instead pouring off the top of a paddle box on its way through a full rotation. @theluthier
Great work!
very good work and i really appreciate your documentation of the workflow.
There were some decisions about the interpretation of the artwork. very interesting.
lucky I Noticed that when you posted last night and thought it was pretty funny :D They connect to the longer beams slightly differently but are otherwise the same.
vincav81 Yeah, I imagine we'd have some control over getting the effect back when we do some water wizardry later. There were a lot of things about the concept art that I mulled over while working this week. It sucked starting so late compared to everyone else, but it was nice having a week to contemplate things while at work.
@theluthier Here's where I'm at for this week's submission:
Vertex count is at 4,477 with shared data; 14,745 if made single users.
Once converted to single users there will be a number of hiding faces in the paddles that can be removed.
I was going to do the nails in the boards with textures, but wasn't sure if it was okay to make maps for metalness/specularity.
I've been saving along the way and can revert any of my shapes to their base blocks pretty quickly if you think normals would do the job in places. Specifically the disc of wood planks has no silhouette and could probably be normal mapped instead. I've baked normal maps before, but baking high to low poly is something I haven't done much of, so making these kinds of decisions is not my strong suit.
Nice result, I think our team will have some kick ass water wheels ;). Kent will have a hard time choosing one :).
You could just model the nails and bake them into your normal map. Another option is to create a decal map and add them later with small planes where you want the nails to be.
I agree, theyre similar in style and yet pretty different -- and all of them look great. I was initially sad that this was going to turn into highlander with so many people signing up, but I've really enjoyed all the research and communication surrounding these wheels. We'll all at the very least walk away having built something pretty cool 😎.
@lexicolopolis LoL!
There can be only one!
triple LoL! so good... the highlander reference, on multiple levels...
: D
bla...
right...
/laughs it out to make room for serious stuff...
first thing's first:
Great job!
from your research (that I happen to know of, being in the same group, same asset ; ) over design decisions to implementation...
COMPLIMENTS!
: )
cannot help but question the highlander approach to the final th0...
(no worries, not about to start a revolution, just thinking outloud... the whole cooperation - competition juxtaposition ; )
Wow! This is awesome @lexicolopolis! The most refined model I've seen so far. Way beyond the expectations of week 1 👏
I'm in love with the stone work and bevel-y imperfections everywhere. You've got a great artistic sense about your modeling.
My only question is about optimization. 4.5 vertices isn't terrible but I also see a lot of room for lowering the polycount. Namely the unseen sides of all the stones. If you continue this into a "high-res version" of the asset, including sculpted accents, an optimal retopology could set you up for amazing efficiency while capturing all the detail in a normal bake.
For week 2 don't worry about optimal geometry. Go for high res and make it gawgeous! We can optimize in week 3.
@theluthier Thank you so much for the encouraging feedback! With 130 signups I'm sure the grading is unwieldy. Looking forward to tomorrow's stream!
Set up UVs and VERY slowly figured out how to bake normals again without skewing. Also made a bunch of nails, bolts, brackets, etc to bake into places. I wanted to set up all my cages and make sure that each one bakes the extra geometry properly before moving on to sculpting anything.
Looking good, baking normal maps can be tricky. I would first do the sculpting and after that finalize the low poly model because you may have to change it after the sculpting. And then you would have to redo your UV's.