If you split UVs and keep one smoothing group you may have a seam when baking normal maps.
That is wrong. You can have split UVs and keep objects smoothed but not the other way around (you can't have a smoothing split without a UV split). Also, those shells are a good candidate for mirroring. Just make sure the triangle stripping is the same across both sides.
That is wrong. You can have split UVs and keep objects smoothed but not the other way around (you can't have a smoothing split without a UV split). Also, those shells are a good candidate for mirroring. Just make sure the triangle stripping is the same across both sides.
Right, Hasng10 has it backwards here.
Here are some basic rules/facts:
1. You need uv splits wherever you use hard edges/smoothing groups, not doing so will cause artifacts
2. you *can* use hard edges/smoothing groups anywhere you have uv splits, because it will not increase your actual vert usage(if you have a uv seam, the verts are split regardless).
3. Hard edges/smoothing groups will not cause any sort of negative artifacts if you've followed #1, however you must bake with an averaged cage. Max and Maya do this by default. Xnormal's basic ray trace setting does NOT(you have to set up a cage in xnormal or export a cage for Max).
So, if you understand 1-3, really you can just use a simple script that sets all of your uv border edges to hard, because it doesn't have any real drawbacks. The only real exception is in situations where you're doing complex partial mirroring.
The only other qualifier I would add is that if you have a contiguous smooth surface mapped onto multiple UV shells it should be smoothed together. That is kind of an obvious guideline but I have seen people make that mistake.
The only other qualifier I would add is that if you have a contiguous smooth surface mapped onto multiple UV shells it should be smoothed together. That is kind of an obvious guideline but I have seen people make that mistake.
Even then its not really necessary, I mean it might look better in the viewport without textures applied but the baker will account for it. The baker has to account for uv seams like that anyway(where uv direction changes and thus normal color) so having the edge soft or hard shouldn't make any practical difference.
I would say that Scruples's method is better because you can completely get rid of diffuse seams, and you might be able to reuse this texture piece for other tubing as well.
Replies
That is wrong. You can have split UVs and keep objects smoothed but not the other way around (you can't have a smoothing split without a UV split). Also, those shells are a good candidate for mirroring. Just make sure the triangle stripping is the same across both sides.
Right, Hasng10 has it backwards here.
Here are some basic rules/facts:
1. You need uv splits wherever you use hard edges/smoothing groups, not doing so will cause artifacts
2. you *can* use hard edges/smoothing groups anywhere you have uv splits, because it will not increase your actual vert usage(if you have a uv seam, the verts are split regardless).
3. Hard edges/smoothing groups will not cause any sort of negative artifacts if you've followed #1, however you must bake with an averaged cage. Max and Maya do this by default. Xnormal's basic ray trace setting does NOT(you have to set up a cage in xnormal or export a cage for Max).
So, if you understand 1-3, really you can just use a simple script that sets all of your uv border edges to hard, because it doesn't have any real drawbacks. The only real exception is in situations where you're doing complex partial mirroring.
Even then its not really necessary, I mean it might look better in the viewport without textures applied but the baker will account for it. The baker has to account for uv seams like that anyway(where uv direction changes and thus normal color) so having the edge soft or hard shouldn't make any practical difference.
Owned. ><
Sorry 'bout that one