Hello! I am RossC, and I've finally decided that one of my New Year's Resolutions will be to finally finish and publish a Dota 2 item set and hopefully get it in game which would be amazing.
Brief history, two years ago I sort of took a break from the world of 3D modeling. My reasoning was that while I was comfortable with modeling and somewhat comfortable with sculpting, my texturing and digital painting abilities just weren't there. It gave whatever models I made a sort of bland look. So rather than continue learning to model I began to practice and learn how to digitally paint.
I've improved a ton and am more comfortable artistically and with the technical side of Photoshop. That's why I've decided to jump back into the chaos that is item modeling in hopes that this time I shall harness my newfound prowess and make some really awesome stuff!
I made a couple of single items in the past but they were basically foundations of learning the process and are kind of wonky in their own right.
As you can see, the texturing and sculpting wasn't exactly great but I learned the general workflow.
The last thing I worked on that I'm actually pretty proud of is this Forge Spirit:
Some further improvements but still not quite there in levels of quality. It's clear that the textures need work and there's an unbalanced level of detail or interest in the full model. This is why I took 2 years to focus on getting better at 2D digital art and get comfortable with Photoshop.
I've decided to work on a item set for Omniknight. Rather than start with a head item, which I usually attempted to do with my previous item set attempts, I'm going to start with the weapon. It's more likely to be less complicated than a head item with hair and such and allows me to brush up on sculpting hard edges. It also allows me to base the rest of the armor of the set to mirror the aesthetic of the hammer weapon. Here's a few of the rough silhouette design concepts I've cooked up:
I knew I wanted to have some sort of illuminated part of the hammer, a jewel or some sort of holy stone that powers the weapon, so working with black and white I tried to draft out some ideas.
I'm leaning towards either B or F. Most of my friends like F but I think I need to rework it's design to feel more hammer like and less axe like. I would love to have a particle effect over F, but for the time being I'm going to accept that I shouldn't take custom particles into consideration and keep things simple.
I'm liking the end of the weapon hilt of F and maybe the sort of chess piece feel of D's handle I subconsciously did. What do you all think? Feedback is heavily welcome! I'm going to do another phase of rough silhouette concepts, focusing on altering and playing with different iterations of a single concept from above before going into the actual paintover concept stage.
I'll be updating this thread with progress as I take feedback and hopefully learn and improve!
Replies
Finally did a paintover concept for B. While I and a bunch of other people liked F, it didn't really feel like Omniknight. Maybe for Axe, but my quick paint overs just didn't feel like something Omniknight would wield, so I went with B.
I really like the concept and I'm starting to get ideas of how the rest of the set will go. I'm open to any feedback on the concept. I'm not sure if I'm going to make the lower parts of the hammer red or if the center part should be red to contrast the rest of the hammer or gold.
Ex: Necro with your sickle and the Phantom Assassin's get a Sickle very accepted in game.
but i liked too of B weapon, the colours and de shape are great.
just one point caught my attention, is the Hammer handle looks to thin.
seeya!
Started working on the block out of the base model to sculpt over. I think I still sort of remember how this is done.
I need to set up some meshes to work as the indention in the hammer, but the blockout mesh is basically ready for Zbrush! Now to figure Zbrush out!
So far, not too bad? The main thing I haven't gotten down is keeping the trim smooth, especially near the inward edges, they're sort of messy and unclean.
Much more cleaner than before.
Sculpt is going pretty good!
Thanks! It sort of looks alright from a distance and I intend to have the main part of the hammer to have the most texture space, but what should I do to improve the fine scratches and their visibility? Make them bigger and more visible/exaggerated? I didn't really want to go too crazy because its meant to be more worn metal than stone, so more subtle chipping away than fully worn out edges. Maybe bigger dents or make the chipped away parts that sort of taper the edges thicker or bigger? Any advice would be greatly appreciated as I try and get better at this!
Finished the main hammer ingame model, just barely under the polygon budget of 400 with 388 polys! Now to UV Map it and see how the bake looks on top of it.
This is how it looks like when I skin it to the Hammer0_0,
And this is how the workshop tools compile it:
While the tiny hammer is adorable, what's going on here?
Now to lower the triangles!
EDIT:
Triangle Optimization Finished, 390!
I did the initial bake and it turned out alright, but noticing a few errors that seem to be brought up with smoothing groups:
So while adding smoothing groups I decided to give it another go with the UV's so I decided to try and make a better UV map. I thought I did a better job with the new UV Map but XNormal doesn't seem to think so, giving me this with the exact same settings:
Only 4 pair of vertices were edited in the model were repositioned very slightly. Are UV's that important to the baking process that two different uv maps utilizing pretty much the same model have different results? Was sewing the edges of the uv islands a bad idea? Because that's the only main difference I can figure out from the two different uv maps.
At this point I'm thinking about reverting back to the old UV map and just cleaning out some of the errors and artifacts in photoshop. But first I'm going to see if actually keeping the uv islands seperate changes anything.
Yellow is Low, Red is High:
Should the lower poly be more bloated out? I just assumed it should match as close as possible to it's higher resolution cousin? I've experimented with editing the geometry but the details continue not to get baked correctly.
I did a basic import pass with a colorized AO_map and my normal map. Looks pretty good without any proper texturing I think! There's a few artifacts I can clean up but for the most part it looks like I'm ready to start texturing the diffuse. My only issue I want to sort of address is that it seems like dota ignores smoothing groups? The bottom part of the handle has a very noticeable flat shading to it and it looks off.
Also more advice, I know a point light map is recommended for textures but is it used for a weapon? Considering the hammer is tilted to it's side is it alright to still try and great a gradient map on the hammer? Thanks!
Put out a better AO map. The smoothing group thing is still a problem though. I think I can probably fix it just by adding more polygons in the LOD0 model, but still.
Its definitely getting there. Should I use bigger and less symbols in the center? It is a bit noisy.
Working out so far. Though is it too red? I think the color scheme is two heavy on red and yellow and only now I'm seeing that it may be a bit too bright and saturated, but I'm not entirely sure what color scheme to go with instead? Platinum? Dark Stone?
Any suggestions on what to do with the color palette? I'm sort of unsure how to make it adhere to Omniknight better.
A simple base grey color scheme to the weapon itself (which is going to be hitting things and thus rubbing off any paint etc), with highlights of red and yellow seems to read the cleanest, IMO. Incidentally, like Omni himself.
EDIT: in other words, I think you just about had it in the post just before last, lol. Just brighten it a bit.