Comments on: Confused about Normals http://blog.blackkettle.org/2006/08/28/confused-about-normals/ Things of Occasional Interest Thu, 30 Apr 2009 14:30:22 +0000 http://wordpress.org/?v=2.9.1 hourly 1 By: alex http://blog.blackkettle.org/2006/08/28/confused-about-normals/comment-page-1/#comment-5 alex Thu, 31 Aug 2006 03:26:14 +0000 http://blog.blackkettle.org/2006/08/28/confused-about-normals/#comment-5 Oh, I understand the principle - it just strikes me as conceptually odd that polygon normals are secondary to vertex normals. For starters, you don't always want to smooth polygons, and if you do, surely it's better to let the hardware calculate the averages than to specify them? I guess I can't really see a use case for vertex-only normals given that front/back information is already given by winding order. In Shaxam, there's only one combination of surface properties that causes the generation of normal data at all, and that's where you've got an atlas UV map on a smoothed surface - and it's only necessary in that case because it's the only way in WPF to get rid of unsmoothed seams between discontinuous UV coordinate sets. Oh, I understand the principle – it just strikes me as conceptually odd that polygon normals are secondary to vertex normals. For starters, you don’t always want to smooth polygons, and if you do, surely it’s better to let the hardware calculate the averages than to specify them? I guess I can’t really see a use case for vertex-only normals given that front/back information is already given by winding order.

In Shaxam, there’s only one combination of surface properties that causes the generation of normal data at all, and that’s where you’ve got an atlas UV map on a smoothed surface – and it’s only necessary in that case because it’s the only way in WPF to get rid of unsmoothed seams between discontinuous UV coordinate sets.

]]> By: Alan Gasperini http://blog.blackkettle.org/2006/08/28/confused-about-normals/comment-page-1/#comment-4 Alan Gasperini Wed, 30 Aug 2006 20:41:57 +0000 http://blog.blackkettle.org/2006/08/28/confused-about-normals/#comment-4 This is actually not that uncommon. The idea is that at each vertex, you can average the normals of all triangles to which that vertex belongs. This will give you smoother shading, since the normals are interpolated across the triangle. This is actually not that uncommon. The idea is that at each vertex, you can average the normals of all triangles to which that vertex belongs. This will give you smoother shading, since the normals are interpolated across the triangle.

]]>