jerome Posted January 21, 2015 Share Posted January 21, 2015 Hi, I noticed the cylinder mesh has got some kind of "edge" on its surface when other "circular" meshes doesn't have this artefact. I rotated the cylinder in the light so this limit is visible :http://www.babylonjs-playground.com/#AM1YW (please zoom in, you'll see a separation between two colors) Has this something to do with cylinder vertexData positions or indices, or with normals computation ? I have the same artefact when I create a circular ribbon mesh even when specifying to "close" it (last positions join first ones in indices declaration) Quote Link to comment Share on other sites More sharing options...
GameMonetize Posted January 21, 2015 Share Posted January 21, 2015 What do you mean by "edge"? Sounds like you can add more tesselation to resolve this right? Quote Link to comment Share on other sites More sharing options...
jerome Posted January 21, 2015 Author Share Posted January 21, 2015 no, it's not a tesselation problem... definetly not I don't express it well. I closed up the cam near by the cylinder here : http://www.babylonjs-playground.com/#AM1YW#1As you can see, the coloration is not continuous : there are a darker blue left side and a lighter blue right side separated by this kind of surface "edge". As if the final vertices in the rotation weren't shared with the firt ones in indices array, so the normals would be computed as for non-contiguous faces (don't know, just a thought). This artefact doesn't appear on other meshes computed along a complete rotation like torus or spheres (but I've got it too when I fold over a ribbon, even when explicitly joining indices of "first" and "last" vertices). Quote Link to comment Share on other sites More sharing options...
GameMonetize Posted January 21, 2015 Share Posted January 21, 2015 Ok got it. This may be due to errors while generating indices or normals. Perhaps we can improve the algorithm to reuse more aggressively vertices Quote Link to comment Share on other sites More sharing options...
jerome Posted January 22, 2015 Author Share Posted January 22, 2015 ok I have the same problem with a cylindric closed ribbon despite I really reuse concerned (last and first ones in the rotation) vertices. So I guess it may be in the normals computation... therefore it would happen in other circular meshes then, so I don't really know. Quote Link to comment Share on other sites More sharing options...
jerome Posted January 23, 2015 Author Share Posted January 23, 2015 ok, I gues it's more a problem of vertice reuse.After the rotation (cos/sin), the last vertex indexes should be declared as the the first ones in the indices array. Not sure I'm very clear , but I think you guess what I mean Quote Link to comment Share on other sites More sharing options...
GameMonetize Posted January 23, 2015 Share Posted January 23, 2015 I guess jerome 1 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.