I’m trying to learn lscm unwrapping and rigging and all that, and I have a few questions (sorry).
One thing is the image quality. I’m getting blurred pixelization and I was wondering how you all avoid this? I’ve tried high resolutions (1600 x 1600) and there is still some jaggies visible. Is there a comfortable median I should use? I know some grahics cards have issues with resolutions over 1024, but I would like smoother gradations and sharper details.
My next question concerns subdivision levels during unwrapping and rigging. Should I work at a low poly count and subdivide later, and will the image maps and rigging translate over to the higher poly count?
The model in question has about 20k polys unsubdivided, and almost 80k after one subdivision. I hadn’t planned on rigging it, but hypothetically, how would you go about texturing and rigging it? There are multiple separate objects (body, eyes, teeth, weapons armor), so should I split the model up? Is it possible to have multiple maps at different resolutions? For rigging, I imagine you’d parent the weapon and armor to the body, but can I have different levels of subdivision for each element or is that just wrong? Image map for the body and just ramps for the other stuff?
Sorry for the deluge of questions. I’ve made a few half-attempts and false-starts at texturing and rigging, and I’ve had some fairly lousy results. I’m probably missing some points, but these are some of the more immediate questions that have arisen.
BTW I’m using Silo and Blender and sometimes Wings.
if you’re not planning on using this in a game or something the texture[s] can be pretty much as large as you want
My next question concerns subdivision levels during unwrapping and rigging. Should I work at a low poly count and subdivide later, and will the image maps and rigging translate over to the higher poly count?
… that’s a lot of polys…
if you’re doing something for a game you’d want many fewer polygons [like 5 or 8 thousand triangles].
You can use a program like nvidia’s melody to generate normal maps from the highpoly version onto the low polygon one [and they will look nearly identical]. Unfortunately the normal maps aren’t usuable in blender.
yes, you can use multiple textures of different resolutions. Unreal tournament 2004 used a 512x512 map for the body, and a 512x256 map for the head. [or perhaps that was 1024x1024 and 1024x512]. UT 2007 will use higher resolution textures, but I don’t have specifics on that… For the artwork they’re actually working at 4096x4096 then scaling down for the final maps.
the parts don’t need to be seperate objects or anything. Except maybe the weapon which you’re going to be switching out [or turning off]…
Yeah I kind of gave up on keeping him low poly. He’s got a few triangles too; enough that the only sensible way to get rid of them would be to subdivide up to 80k polys.
That’s good news about the high resolution images…I suspected it was possible but wasn’t sure. I guess keeping the specs low keeps the model open to wider use, but I like to add a lot of detail and focus on the sculpting.
Thanks for slugging through my numerous questions, zero.
Thanks, Henry. I’ll give that a shot. I know nothing of game modelling, but I suppose it’s a good rule of thumb to keep the resolutions as low as possible without sacrificing detail. I think I’ll start big and work down till the jaggies show up.
Thanks, Madcow. I’ll give that a shot if I ever get past the unwrapping stage. I think my topology is bad because I can’t seem to get workable results. I didn’t realize you could use the image texture along with the uv.
Is this a bad uv unwrap? I had to use a lot of pins to get the eyes and mouth open. It was really difficult and I think it’s because the base mesh is pretty sloppy. It seems to me that the polys are really uneven and will distort the images. This guy looks how I feel.
with it applied to your mesh look for places which are oddly stretched. I’d bet with your face there you will find a lot of them. These are the places that need to be tweaked.
Second: one of the fancy tricks people use is to "smooth’ the mesh before uv mapping. Create a vertex key [absolute key] and enter edit mode… select you face and hit the smooth button in the edit mode a bunch of times until it looks all flat. The idea is to loose the shape of everything [i wish I had a picture]. Then do lscm from that vertex key. After uv mapping you can delete the vertex key and edit your mesh or whatever…
oh, you also probably want a seam that splits the neck and forhead more, this should help with needing pinning much more.
I’ll look into that smoothing/vertex key trick…never heard of that before. I suppose I could go back redo the mesh, but dealing with these problems as they pop up is teaching me a lot.
