Is turbosmooth needed?
polycounter lvl 11
Lately i've been having some cases where i could only use turbosmooth on the floating geometry and still get away with the gun looking nice.
Whats the turbosmooth mainly being used for? Just to give the low poly model the high poly smoothed look? Why can't someone save us some time and create a turbo smooth looking shader?
What effects will using it/not using it have on your model?
Whats the turbosmooth mainly being used for? Just to give the low poly model the high poly smoothed look? Why can't someone save us some time and create a turbo smooth looking shader?
What effects will using it/not using it have on your model?
Replies
It's also how "current" game art generates normal maps. We don't just just run it through the Photoshop filter and call it a day anymore. Granted, some studios and some games still do it. High Poly modeling takes time...and time is money. But the better you get at it and the faster you become, you models and normal maps look 100% better and you never want to model the old school way again!
My main problem is if you have a low poly cylinder and you try to smooth that you get little black edges. This problem occurs in many other places to. Is this because im using less polys than i need to?
While we are on the subject of normal maps, why are tangent maps harder to deal with than object maps?
Turbosmooth isnt ideal for everything. Your modeling guns so there are indeed a few ways around that workflow, and the benefits may not be so obvious.
The thing about SDS modeling is the way it lets you work with a relatively low poly cage and turns it into a smooth curved surface and back without any problems. its just so much easier to work with a lower res cage of strategically placed geometry. its easier to model in the first place, easier to change, and easier to manage in uv unwrapping, and skinning.
Since something clearly went wrong in your workflow somewhere, it'd be a good idea to post images of what's going wrong, as I feel people are debating the wrong thing here.
Aside from the issue you mentioned: no, you don't need to sub-d model. You can also hard-model everything, and that's certainly a valid technique, but one with possible even more difficulties, I'm afraid. Especially in max, where bevelling edges doesn't give a very usable result where edges meet, and things get messy really quickly.
I suggest you just knuckle down and learn how sub-d modeling works, as it's really quite fast. Since you're already working with floaters, you've already sidestepped many of the more difficult issues with sub-d modeling.
Is it though?, I thought nowadays the differences were negligable, and meshsmooth is at the forefront(the ribbon) while turbosmooth is in the background.
I still use turbosmooth. I like have the modifier, so I can use a shortcut to toggle them all on/off + my symmetry modifiers I use often.
For the "little black edges," it may be over-smoothing it so that some polys are actually becoming flipped. Or problems usually start happening at the ends of the cylinder; maybe try insetting the ends.
Like Ott said, meshsmooth is exactly the same as turbosmooth, except that turbosmooth is more efficient. It's more efficient because meshsmooth allows you to edit the model while you have the meshsmooth modifier selected; you can move verts around and "crease" parts of the model. Whereas turbosmooth only allows you to view the model as subdivided, and not be able to edit it unless you're at the editable poly level of the modifier stack.
Really? I never knew that. I don't use editable mesh ever, so I guess that's why I didn't realize that was what Meshsmooth was doing.
Well, thanks for the info Perna
Im thinking alot of my problems with turbosmooth are coming from my low polycounts. My guns are usually about 1.6k quads and i always have a problem smoothing it. But could i safely double my limit and get much better, turbosmooth friendly, results?
YOU
SO
MUCH
THAT
IS
AWESOME
you still need to know conventional subd techniques to make the most of it.
I've heard people refer to this method before, but I've not tried to put it into practice before. I tend to stick to more 'traditional' subdivision techniques, and usually use Turbosmooth with two iterations (and isoline display) to achieve the result I'm after.
Mesh density is an easy enough concept to bear in mind, but could you give a simple example of how you'd build a mesh specifically with this technique in mind?
Awesomesauce. *saves*
I've tried a similar technique before with tesselate as the lower modifier, but that had rather undesirable results. This seems like it works a lot better.
Tesselate works too for boxy things, as it blindly adds edges and doesn't follow curves.
While density is important, meshflow is important too. The same rules still apply from traditional techiques, for example when there is a pole on a corner it will not smooth properly. Really it's the same standard stuff, but this is faster since you don't need to manually add in edges and bother with much control edge stuff, instead dealing with a much simpler mesh. That said, I use traditional subd methods often for simpler things so the tricount doesn't skyrocket, and so I can have manual control on particularly tricky shapes.
Yeah, problem was, not the entire thing was supposed to be boxy... Some parts needed to be rounded, which is where the turbosmoothingroups come in to play.