Tips for Isolating xNormal errors?
polycounter lvl 11
I've been a user of xNormal for a very long time. I have baked out several special maps and I feel like I can avoid problems before they begin. However there are some error that I have resolved through horrible practices. The errors come in many varieties but the common factors are either (A) xNormal thinks the vertex counts are different or (B) xNormal thinks the vertices have been re indexed. I do believe there has to be a better way of resolving them since problem solving the way I have been is a MASSIVE time sink.
1: Vertex counts differ: I have seen these type of errors come up a lot. I know that this isn't true because I can select both my cage group and the low poly group in the Maya Outliner and have the same number of vertices appear in the HUD. I never figured out how to solve this. I've always just baked out different shells independently until I find the one that causes the issue, then recreated the cage for that shell. It takes to test an object that has several dozen shells since i have to bake out one at a time. sometimes I do groups of shells to see which ones work, which ones don't and target those groups but it still takes a very long time. There has to be a better way... right?
2: Issues with vertex re-indexing: I'm honestly not sure what causes these. One minute it works, the next minute i push and pull vertices without doing any edge splitting/collapsing or any topology modifying action. I even toggle the wireframe between the cage and low and they overlap perfectly. What on Earth normally causes this and is there any way to isolate which vertices in particular have been reindexed?
Thank you very much for your time.
1: Vertex counts differ: I have seen these type of errors come up a lot. I know that this isn't true because I can select both my cage group and the low poly group in the Maya Outliner and have the same number of vertices appear in the HUD. I never figured out how to solve this. I've always just baked out different shells independently until I find the one that causes the issue, then recreated the cage for that shell. It takes to test an object that has several dozen shells since i have to bake out one at a time. sometimes I do groups of shells to see which ones work, which ones don't and target those groups but it still takes a very long time. There has to be a better way... right?
2: Issues with vertex re-indexing: I'm honestly not sure what causes these. One minute it works, the next minute i push and pull vertices without doing any edge splitting/collapsing or any topology modifying action. I even toggle the wireframe between the cage and low and they overlap perfectly. What on Earth normally causes this and is there any way to isolate which vertices in particular have been reindexed?
Thank you very much for your time.
Replies
For #2, vertices will become re-indexed at any point where a component is added or removed (but not moved, scaled, or rotated). This means if you add or delete a triangle edge or even delete a face and fill it back in, the vertices become re-indexed.
I exclusively create cages in Maya and have never encountered either of these issues without it being directly caused by consciously adding or removing a component. If you have some files to provide with specific examples then it might help find a solution to whatever is causing you to ask these questions.
Also for the record, if you make a LP and your cage then you find out you have ngons or other unsavory topology, you CAN modify your mesh by adding or removing components to fix the problem. You can do this as long as you perform the exact same actions in the exact same order to both meshes. Modifying both in the same order causes the vertices to re-index in the same order as well.
https://www.dropbox.com/s/h4txrhguqvtyfeb/headphonesv01.zip?dl=0
When I tell people I never did any mesh edit they never believe me so I decided to upload a log of everything from the point where it was working to the point where it broke. I was just using keyframe and gizmo tools. No Modifiers, Mesh>* or Mesh Edit>*
https://docs.google.com/document/d/1nZ0Nxfpks3aDpFGdwa78mS5iVmxCDg33ccwSflheGZY/edit
As for the first link, I did check it out and your LP and cage both seem to be matching up both in 3D space and vertex indices.
Whatever you are doing after this stage, if you are having cage issues with this file, is likely causing your problem. I do have to ask, though, why is your low poly split into seven pieces? You don't seem to have an inside modeled to any of these parts so they really should be one contiguous mesh. It will make cage generation easier, too.