Hi all. I’m following this tutorial for hard surface. He suggest to use alphas to easily add details to the surface, and I find it very convenient, but mine has pixelated borders no matter what.
The tutorial poly count are around 740k scene vertices but he still got smoother results than mine. (check down for more images, I can’t upload more than one).
I can have decent details just around 1.5 mil polycount. I think there are too much to sustain in multiple objects in a complete project.
If I add details, then voxel remesh down, I got orrible results. I can decimate but I’ll lost wire, then add details in the future will be pain.
I turned on “smooth shading”, that fixed a bit. But I wonder, is that just a visual smoothing or it really modify the mesh for a cleaner result ?
I found out that jaggies lines appear more when the alpha is not aliened with the grid.
Blurring the alpha a little helps, but not fix.
Any tips to improve the situation?
(I already asked to the author and he suggested some of the mentioned fix, but I’m not satisfied with the result, and I thought that asking to the community would be fine. I don’t want to depreciate the tutorial than, beside of that I found very helpful)
As knacki mentioned, @Razuminc should be able to help. It’s his tutorial so you should contact him. Maybe he has Dynamis Topology enabled or something?
We already spoke privately, but unfortunately following his tips I can’t get a good result. Therefore I decided to ask to the community, maybe somebody else has a similar problem.
My intention is not criticize Razum tutorial, just ask for a third opinion about the problem. If this result not appropriate I can remove the thread.
You should turn on the wire frame and zoom in a bit closer to your object, so you can better understand why the edges are jaggy. It’s like a photo, the further you zoom into the image, the more pixelated it gets.
You need the high voxel density where you want to use the alpha image and not on the whole sphere … 500,000 is not necessarily a high resolution if the voxels are distributed over a large area.
Use as the others already said dynamic topology or 2-3 or 4 million … then it will be sharper and sharper, but also more and more memory consuming.
And YES, the alphas look like this at 500k if they don’t have blurred edges.
Your alpha image can also have an impact, because this image also consists of pixels e.g. 2000 x 2000 or only 500 x 500…
