Hey guys, so I've finished modelling this vending cart based on this concept (second from bottom right):
Here is a render of my model with a 1024x512 tiled road (texture is up for debate):
I've been suggested to that I should consider hand painting the textures to go along with the style of the concept (once I've unwrapped it). I was hoping to get some suggestions or opinions on which approach to take in the texturing phase. I can also post a wireframe if need be but the model itself is fairly tight and under 4,800 polys. Also, I figure I can make this into a full scene, I'm thinking of expanding beyond the curbs of the road and adding grass and foliage. Any pointers there would be helpful!
Thanks a ton guys!
Replies
And i think u should put it in the middle of a town or something. Might be a lil much, but if you can just bust out some 3-4 modular buildings u can put something together quite nice
Questions.
Why make a full scene ? how would scene help your portfolio ?
Why make a single prop ? how would scene help your portfolio ?
How would making a hand painted texture help your portfolio ?
How would making a photo sourced texture help your portfolio ?
Why did you start this project ? and what was the goal ?
lets clarity things at the most basic level first, before moving forward with anything else.
This will help clear ur mind
@Hiten: Why make a full scene ? how would scene help your portfolio ?
I thougt that making a whole scene would make a nice addition to the portfolio because its something I haven't really done. That and I wanted to play with some daylighting and texturing grass. Do you think its a good idea or should I leave this as more of an asset?
Why make a single prop ? how would scene help your portfolio ?
The goal of this assignment is to have a nice asset with good modelling and solid texturing. I'm just thinking of taking it a step further.
How would making a hand painted texture help your portfolio ?
The idea for taking a hand painted approach is that it seems like it would lend to the style of the concept well, that and I feel like it would be something cool to learn.
How would making a photo sourced texture help your portfolio ?
Photo sourced could work too, but again, it feels like it may take away from the style of the concept.
Why did you start this project ? and what was the goal ?
I started this project to have a nice, stylized asset with the idea of making a mini scene to show off multiple areas of study. The goal is to have this portfolio worthy.
Thanks for making me think about all that!
@Tigerfeet: Thanks for the headsup on the source! I was wondering about that. As for wireframe, here you go:
Other then that I think it look great, nice job
Sorry for the messy english
In case the text is hard to read (I don't write well with a tablet >.<) here's what things say, with some elaboration
The red lines indicate loops that could be removed entirely with little or no silhouette loss
Blue is where things could be condensed into tris
Orange where things definitely need to go unless you plan on mirroring the UVs along that line (which I wouldn't recommend on some of the spots I found them)
Things I think you should keep in mind going forward
1. Don't be afraid to float geometry, there's no reason to go through the hassle and extra geo of building certain pieces into the same mesh when the same effect can be achieved simply by butting up two separate pieces.
2. Tiny details, like the raised bit on the big wheel can be completely in the texture, if they don't drastically alter the silhouette it's much more poly efficient to simply paint or normal them in (depending on whether you're doing a normal map or handpainting)
3. Smaller cylinders don't need as many sides as the larger ones. Even on objects that are the same mesh you can progressively collapse edges as things get smaller to be more efficient.
4. If and when you do model more complex objects as a single piece, you should collapse unneeded loops as soon as they stop defining silhouette.
With that in mind here's one of the best ways to determine which edges are vital. I challenge you to do it as a sort of homework.
1. Put a pure black material on your model so all you can see is the silhouette
2. start collapsing and deleting edge loops
3. If you see no drastic change in the silhouette, they are unneeded
Remember skill at modeling isn't just about making something pretty but also conserving when ever possible. It's less about the actual tri count and more about making sure that every tri has a clear purpose on the model. Speaking of tris it'd be great to know the tri count rather than polycount as that's more accurate (polys can be 2 tris, or more if there's N-gons, tris are the base level of flat surface)
My paintover is not the be all and end all, you may find a few I suggested to remove are necessary, I can't see the thing in 3D to know for sure, you make also find many more places to collapse things that I missed.
Also try not to render things out with a black background
It's looking nice so far and I'm excited to see how it turns out.
+1
Going back to #4 of what I said earlier you can still bring lots of edges to tris where you're still continuing them through the whole model
Like I said before as soon as they stop being useful terminate them.
There's no need to keep things in clean even quads unless you're planning on bringing it into a sculpting package like Zbrush or Mudbox. Otherwise tris are completely valid and can really help reduce things
Case in point, I modeled out two of your shapes real quickly and made two versions, they retain the exact same silhouette but have very different polycounts
I can tell you're carrying over edges in lots of places: the two loops down the middle of the main cart they might define something under the piece with the meat on it, but by the time we can see them there's not reason for them to be there, they can be collapsed together into one and then even target welded to one side to save more tris. You've also done this a lot on the posts and edges coming off the main cart.
Personally I would reduce the cloth at least a little bit, if you can draw a nearly straight line between two points along the geo and there's an edge loop in between it can probably go. At the very least like I said above reduce into tris wherever an edge loop isn't actually doing anything. But obviously it's ultimately your call. The only reason it would need to be that dense is if it were flying free and going to be animated. Something that attached to the cart won't really move.
I would also still reduce the number of loops on the tori/rings as well as the smaller cylinders of the pots and the back wheels
Always good to mention if it's a high or low poly just so people don't get confused, plus there's no point in giving a tri/poly count for the high as it's never actually going into a game engine. Which is probably why people assumed it was the low and started asking for wire shots.
Anyways my crits still stand for when build your low. And at that point I think you'll be really surprised at just how low you can get while retaining all the detail of the high.
IMO:
Go hand painted texture style. and have only one of these in ur portfolio. unless you REALLY LOVE that style. and wanna work on Darksiders 3 :P
Stick w/ the one asset focus in mind. But continue "planning" on a "small" , very small tight scene. Don't go Micheal bay on that scene...
Now for your solo prop, define what done looks like and what doing is.
I'll be waiting for your response.....
Now I have the normal map in a normal bump slot under 'screen' which I've found gives the cart itself a kind of animated style to it (which I was kinda aiming for to go along with hand painting). What do you guys think about it?
Here's my diffuse so far for the cart:
And a normal map I baked from the zbrush sculpt onto my low poly through xnormal (which is still being tweaked as it came out messy in spots).
Lemme know what you guys think and what should be improved! Thanks as always.
Thank you kindly!
@Selaznog:
Thanks! Still have a bunch to texture still, but for sure you are right. Grunge on the tires makes sense.
I'll give it a go. Thanks.