Sign in to follow this  
dpadam450

Mesh Mender ???

Recommended Posts

My triangles are non-indexed, so for each triangle there are 3 unique vertices. 1.) Why and how can mesh mender be spitting out even more triangles than the original mesh? 2.) Also, there is an assert in there claiming that a computed normal is coming out to be of length 0 causing it to crash. Even if I pass in my own normal data, and tell it not to compute the normals, it will still crash.

Share this post


Link to post
Share on other sites
I haven't used mesh mender (is this the nvidia library?) but if it is already non-indexed and then it gives you more vertices and triangles, then I would guess that either one of the input parameters are not correct or the library isn't working. Have you compared the input to the output? What is the additional material? You will probably have to post some of the results to allow someone to help diagnose what is going on...

Share this post


Link to post
Share on other sites
If you just type "mesh mender" to Google the topmost hit display the following result:

Quote:
theVerts - Should be initialized with your mesh data, NOTE that when mesh mender is done with it, the number of vertices may grow ...


I suggest reading this particular document to find out why the number of vertices may grow as stated in the reference. But some reasons you can easily think of are:

- the triangle layout is not optimal hence the mesh gets remeshed
- you get better indexing schemes by inserting help points
- a totally different polygonization of the incoming points is used
- ...

Edit: Google is your best friend

As for the crash:
It should obviously not do that. Trusting nvidia to some degree I would assume that your input data might not be valid causing this crash.

Share this post


Link to post
Share on other sites
I know I read this and it talks about adding more because of smooth groups and splitting.

Each of my verts, being non-indexed, don't share triangles: therefore there is no normal averaging / smoothing to be done.

I will check the output data. But the only way to not crash is to take out the assert, because a normal is computed to be zero length.

Share this post


Link to post
Share on other sites
If you have vertices at the same location but belonging to different triangles mesh mender might melt those vertices. But in this case the vertex count should decrease. Without knowing anything about mesh mender I cannot tell what may cause it to increase the vertex count.

But you say no vertices shared among triangles? Your mesh might not be a closed continuous mesh which mesh bender might require? This could screw up the internal algorithms so I suggest finding out more about the input constraints mesh mender sets upon you. I would still say the input data is invalid as in terms of what mesh mender can process.

If this does not help either try looking for people using mesh mender and let them try your input data. You should find such people at nvidia forums I would think.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this