Hi
I have been making a very low poly version of Ezio from the Assassin’s Creed 2 game and have just about finished the modelling. This is the model:
But now I find myself needing to texture it, which I have tried and failed miserably at. 
Anyone got any tips?
Thanks in advance
Nice character here!
Separate logical parts of mesh, add seams. Unwarp parts, bring images to picture editor of choice and colour them. Assign material to parts, add painted textures…
What was the workflow you failed, actually?
Thanks for the response.
I managed to unwrap it all but as soon as I started to paint the textures it just looked really bad.
Any ideas on how to texture it to get the same sort of effect as this:
Thanks
@Silencer
The picture shown above is made by some pro. so, don’t think you’ll even come close on first try.
As @eppo said, assign materials to different parts and unwrap them. for starter, boots, pants, belt, and the rest of the gear, assign material with desired color. unwrap your model, bake te textures and retouch your baked texture file in gimp or photoshop.
All the above good looking cuts, coloring is done in Photoshop. these kind of materials are not possible in 3D applications.
compare what you have done with the high-quality-model as reference-
there are a lot things you will fail.
Only one: the arms
you made the faces of the arm the way with an edge where the other model has no edge and its the place of this metal-armature-arm-shield-type thing.
The idea is, you need faces at the model that can be painted in the image like a flat picture. You can try to set seams and unwrap like hell, but you never will have all faces like this. But if the texturing “breaks” at places like the seams under the arms, then its hard to see it –
If you look at the painted diffuse-map, then you easy can find larger parts, that are nearly flat painted - you even could project a photo of something to those parts without big re-sizing and re-“warping”.
I am none of those pros! But i would suggest you create a first unwrapping and try to get large parts of your model in such a flat way out.
Then - dont use such a small picture, use a large one - minimum 1024*1024 - and use the blender-paint-texture to paint a first coloring. Save this and the uv-image (the one with the edges, to be uses as a background mask in your paint-program)
and use some photos as stamp/brush/clone to check what it looks like in blender with this texture on your model.
The very last step - if you ever want to do it (you dont need if the model is not for such a low-resource game-thing with small texture-sizes) is to scale the image down and then again edit single pixels that were distorted too much. There are different methods to do this down-scaling (its called like dithering, because multiple pixels are combined to one but without loosing the structure, so you dont end with a total blurry image).
Sorry for taking so long, but I finally got around to finishing it up and texturing it. The no. of faces has increased to 572 because I realised that I missed some small details in the first render.
So, here it is:
Tell me what you think of it, because I think that it turned out alright. Here is the texture, it is 1024x1024 at the moment but I might try and make it smaller, maybe 512x512:
(Don’t know what happened with the black next to the hood in the texture)
Anyone know why the texture goes all blurry in the render, it makes it look bad in my opinion.
Thanks for the help everyone.
One comment: The example you posted above (post #3) states the art assets are @ 256x256. You can bet they were not painted at that resolution, but scaled down to fit a usage that requires small art assets. As mentioned, for decent detail use 1024 or better, and you can scale that in Photoshop or Gimp and use Unsharp Mask + touchups to retrieve some of the detail clarity.
anytime you re-size an image you lose image quality. Photoshop has an option called smart object. Right Click the layer that contains your image and choose “Convert to Smart Object” now you can resize your image anyways you like and the image will never lose quality. If you need to make edits to the image you must rasterize the image first. NEVER re-size the Image with out converting it to “Smart object”
@Silencer: use a simple cube and unwrap it to your created texture and
create a render of this cube with the texture to check if any material-settings
are wrong.
There are different settings that will change the result – interpolation, lighting…
and
last, if the unwrapped texture-space of a face of the model is too small, then
you will always have the result, that one pixel of the texture is spread over a
visible part of the model-face and you cannot create sharp things like lines for
this part. This is not bad for those parts of the model, where you dont need sharp
looking things. You can try to “cheat” a bit and for example the big symmetrical ornament on the stomach could be cut to the half and the uv-unwrap of the 2 stomach-faces should use the same texture space (but mirrored). Then the half part of the ornament can use the double size and a single texture-pixel will use a smaller part of the face (it is projected to).
But first check the different options for texturing in the material and texture-image-settings (and use a simple cube …so your view is not ?distorted by the uv-unwrapping (which may have too options to be optimized)).
