What meshopt_SimplifyErrorAbsolute does ?

[GiuseppeCassisi96]

Hi,

I wonder what meshopt_SimplifyErrorAbsolute does. Does it allow meshopt_simplify to return the error as an absolute value instead of relative to the mesh extent?

[zeux]

Yes, as described in https://github.com/zeux/meshoptimizer/blob/master/src/meshoptimizer.h#L335. If the mesh is 50 units across (eg positions XYZ are from -10 to 40), an error 0.01 without using this flag is equivalent to error 0.5 when using this flag.

[zeux]

Yes, if you want to specify the target error as relative.

meshoptimizer/demo/main.cpp

Line 646 in b092eee

void simplifyClusters(const Mesh& mesh, float threshold = 0.2f)

example does this.

[GiuseppeCassisi96]

Dosen't it should transform the error to absolute instead of relative ?

[zeux]

If you know the target error in absolute units you can just use the flag. If you want to simplify the mesh with an error of 1% but use absolute flag, then you need to convert yourself.

[GiuseppeCassisi96]

Ok many thanks. I would like ask you another question. Why you resize the meshlet_triangles array in this way after the mesh splitting ?

const meshopt_Meshlet& last = meshlets[meshlet_count - 1];

meshlet_vertices.resize(last.vertex_offset + last.vertex_count);
meshlet_triangles.resize(last.triangle_offset + ((last.triangle_count * 3 + 3) & ~3));
meshlets.resize(meshlet_count);

Thanks

[zeux]

This just keeps the triangles array aligned by 4 bytes (or by 4 indices since each index is 1 byte); technically meshlet_triangles.resize(last_triangle_offset + last.triangle_count * 3); would work as well but it's usually better to keep data 4-byte aligned if it's going to be read by GPU.