Traffic cone Critique
polycounter lvl 3
Hey everyone, I been working on a small project to help me get a better understanding on making game ready assets. I been learning how to use the quad draw tool in Maya and I was wondering if I can get any critiques on my topology on the low poly.
Also here's a reference if needed
Also here's a reference if needed
Replies
The base is missing the rubber feet protruding out.
Can you post your high poly WITHOUT the wireframe so we can see what it exactly looks like and what you'll be baking from?
What are you trying to figure out exactly about making a game ready asset that Youtube hasn't elucidated?
So, looks like you got the right idea at least, with making a high poly model and a low poly of the high poly.
High poly looks fine. There's an opportunity to model in the highlighter strips and the rubber base if you feel challenged. Would just involve minor extrusion operations.
What are you planning to bake the normal, curvature, height, cavity, AO, etc. maps with? Maya? xNormal? Toolbag? Something else?
What are you planning to texture the asset it?
How are you planning to render the cone?
I want to texture it in Substance Painter.
I was going to render it in Arnold or download the trial in Toolbag.
High Poly
Low Poly
Those green vertices are pointless, just use the weld tool and weld it to that outer corner vert. The loops I scribbled in black aren't needed either if the sides are flat. You can just remove that loop and connect it to the corner verts like I did in red.
On the bottom it will look like this. Weld those same green verts to the corners like on the top.
Not a huge problem but its always good to optimise, especially when its a quick fix
Your UV island amounts could be greatly reduced. Why is the base plate split into quarter fragments? You're upping the vertices split count unnecessarily, it'd put more strain on render time in a game engine. Grab those faces and use a PLanar Projection. Grab the strip around the base and cylinder project it.
Also, if you have't already, i reccomend downloading Nightshade UV Editor scrupt and installing that if you're planning to use Maya towards the future. It will greatly help your UVin'g go MUCH faster.
You probably won't understand exactly what people are trying to explain about UV's until you get fully into texturing and see firsthand how bad (poorly optimized) UV's effect the texturing process. So my advice is to not spend too much time trying to get UV's perfect before going ahead and baking and texturing, but instead go forward with imperfect UV's so that you can see how it works for yourself. Then peoples suggestions and crits will become more useful as well.
You won't learn everything from a single model. It's good you are working on something simple to hone in on important skills like sensible topology, optimized UV's, etc. But, as a beginner, I believe there is more benefit to making lots of models rather than trying to make a few perfect ones. I don't mean to just focus on completing shit models above all else, but rather, focus on picking up and refining a few skills from each model, but not spending undo time. This way, each time you finish a model, you will easily be able to see from the finished result what skills you still lack, what you might improve, and you'll have a lot of forward momentum as you'll keep seeing ways to improve. Rather than poking at a single model and not being sure exactly what it needs or if you are doing things the right way or not.
Keep it up!
@aclund3 thanks for the video I will look over it before moving on.
@BIGTIMEMASTER I get what you are saying and I also think that is a good way to learn more about modeling. I will take your advice and start on some more projects.
Read this tutorial, see if it helps.
https://marmoset.co/posts/toolbag-baking-tutorial/
Also, it will help to straighten all of the UV shells in the UV space along the horizontal or vertical axis. Weird stepping artifacts can occur with slightly off-angle shells. Good step towards improving the UVs though
Any other suggestions is always welcome
It doesn't make sense for all models, but for simple things like this I like to build the low-poly game topology first. I don't bother getting the right forms or anything, just the topology I want. Then I take it to Zbrush and use the superior move brushes in that app to start working on the silhouette, and then I start subdividing. When my high poly sculpt is finished, then I drop down the lowest level and export that as my new low-poly game model. The great thing about this method is that usually I don't have to do anything with that low-poly model to prepare it for baking except unwrapping it. So long as it's not extremely low-poly, it should match your high-poly's silhouette almost perfectly.
I started out doing the high-to-low method because that's what all the tutorials out there show, but I was always having lots of frustration that way. Once I started working the opposite way, I have less issues, I generally enjoy the pace of the workflow more, and all the time I am finding ways to do more with less -- i.e. relying less on subdivision-based detail in zbrush and more from noticing subtleties in my game models silhouette, or finding ways to get more from my textures.
Well, that's my spiel. Don't take it as ultimate truth, just something to try and see if it helps you.
I like it, but this is a traffic cone. What it sounds liek you describes works for something more complicated in physical form?
But I'm just approaching one year in 3d, so that kind of workflow doesn't provide me any benefit as I am still getting to grips with the tools and also improving my art little by little. So for my situation -- a beginner -- I think I've found a pretty good workflow that helps keep the ball rolling without too much frustration.
Also, in general, it's helpful to add a VERY SUBTLE roughness noise over the whole object to break up the specularity just a tad bit.
Also, wear and tear should be in specific places. If you're not handpainting some of the scuffs, you're doing it wrong.