I was wondering what is the workflow for creating a character...
-1 low-res modelling
-2 high res model
-3 shader
-4 joints/bind
?
Im asking because i see the tutorial make the skeleton and the skinning on low res models...
Another question, how do I keep objects like eyes and nails, attached to the character durring the animation and/or simple moving ?
tnx
QUOTE(highenduser @ 05/14/09, 02:44 AM) [snapback]307719[/snapback]
I was wondering what is the workflow for creating a character...
Theres no one workflow. Ive so far seen about 200 different workflows that are are as good as far as i can tell. Personally i do your 1 2 in oppose order nowdays.
Shading comes in paralel with all of the steps, because its often decoupled totally. Or if one works alone id say it comes LAST.
joja, Say "it depends" does not help me !
The question is, the skinning process, is to do with highres model or the low res ?
Yes, i understand that. But theres no definite answer its like asking how to make good music. Well it cant be answered in a concise and comprehensible way. Theres no right or best answer. Theres no magic if you follow this you will become king kind of solution (granted somebody WILL SELL you book with that title).
First it depends on what you are doing. Are you rigging for games, film or tv. Is the rig a human humanoid angel, horse, robot, house, train or a drop of water (polygons, nurbs metaballs or implicits, 2d or 3d, cartoony or realistic, with muscles, clother or fur, do you have render farm for preprocessing what about rendering or postprocessing)... Term rig does not mean anything other then something used to help animating the model. If you want some work flow in general you'd get this sort of meta discussion where one could mention words like assets, model, skin and yada yada.
All of these discussions would essentially be lies of some kind, because to give definitive answers the question must be definitive. Like i have this rig (download here) and i have this and that problem id like to avoid this and such how would you attack the problem. Then after a while we would gravitate to some short and meaningful solution. Problem here is keeping persons interested in your work for long enough.
So the general work flow for you frm this point on to get a worklow is.
Describe to yourself what you need to do. What do you need. What do you know allready, if so you probably dont know this but you think you do. (I don't know diddly squat, i just know more than some other people so i can know more about what i don't know). How do the tools work, what knd of animation would it need to do. Do i understand maths. etc etc.
Most people would say that you model first then you rig. But see thats a bit of a misunderstanding. Modelling is rigging they are intertwined. How much depends on the modeler and rigger in question. Most people would say you need bones, well yes and no, if you saw monsters versus aliens then the blob does not really benefit much form bones in many aspects.
QUOTE
The question is, the skinning process, is to do with highres model or the low res ?
Well that is a more definitive question. Id say Yes because i cant say depends. NO it really depends are the shapes historically dependent. If so bind the one that's more important for the effect.
So for example if you use a smooth mesh animation then yes you can rig the lowers model. Because rigging it is same as rigging the hires one. Makes sens, but then theres this rig that's not highres lowrs dependent so you skin BOTH.
QUOTE(highenduser @ 05/29/09, 06:09 AM) [snapback]308504[/snapback]
joja, Say "it depends" does not help me !
The question is, the skinning process, is to do with highres model or the low res ?
well for best performance while animating you don't want deforming geometry on the rig .. so usually there will be low res segmented geometry parented to each bone.
Then there might be a medium res "cage" that is deformed that can be toggle on or off
Lastly there will be a high res mesh for rendering, this mesh might be a smoothed verion of the afore mentioned cage or it might be indpendant.. obviously you want to try to avoid having to weight map two different skins.
There are alot of variables here and it's very easy to create a skin that will make it impossible to animate so there has to be some balance between performance and detail.
Well, it's true in a sense that there's different things that make sense for different goals, backgrounds, etc. but I think there's more to it than personal preference and some workflows are much more time consuming than others or will have gotchas. Keep in mind, too, people with the most experience often are set in their ways so what they do is not so much what's optimal today but what they have grown quick with and it might take you a long time to learn to do yourself.
From my experiences learning character model and animation, it is not a bad idea to start with very low poly like a stick man.
You can go other direction and sculpt something high first, but once you are an adequate modeler it will actually be quicker to polymodel a new character based on the sculpted one than to waste time with retopology tools. I was sure that would be the way to go but it turned out to be a big waste of time.
You get a lot of other benefits here besides that, though. It's a lot easier to bind and UV map a low poly character than a high one. It's actually better to do that beforehand because you can transfer the low poly maps onto the final target.
Keep in mind the starting place is just a stick man with a few hundred or maybe a thousand polies (not that the count matters much, it's a throwaway. You don't make a fully detailed character. It's just a huge waste of time.
So you make low poly, then sculpt in zbrush (or whatever). At the end, project the high poly mesh onto low poly so it has the right basic proportions and take it back to maya. That's when you UV map and bind the low poly. This should take little time, I can good UV maps for low poly character in maybe 20 minutes with good tools like wings3d or blender (don't use maya unless you know it very well for this). You will have to adjust these later, but this is a step that saves you a lot of work.
Now you take back the high poly (or the highest subdivision your system can handle anyway) and you model your final geometry to it. You don't need to do any snaps or anything here, or spend very long just a couple hours at most. You'll probably need to do some stuff like hands and ears separately but they are not a big deal to make - just hard to get good proportions for, so probably do this ahead of time if you don't have them lying around already.
As you make the new model (basically just extruding and welding) you don't need exactness, just topology is what you worry about. That and polycount. Just rough it in, then smooth a few times and pull out any stray verts to the more or less proper location. Then take it back in and project details of high poly onto it and bring it back to maya. Now transfer the UVs and the bind, and make any finishing touchs. And that's it, you are done.