Hello,

I was wondering if someone might be able to provide an insight as to what is going on in the scene of my file.

Within my “Poles” obj node, I have a blue netbox with two wrangle nodes that both ‘promote’ normals on the primitives with:

@N = @N;

There seems to be an issue with my red colored wrangle node.

Although in my scene view it shows(displays) the primitive normals as being in the ‘correct’ direction;

It doesn't render at all, with exception of the endcap.

I played around with it and get renders if I do any of the following changes:

After my ‘red colored wrangle node’ -

1) Adding a polyframe node with N as Tangent.

or

2) Use an attribute delete node and remove the points normal attribute.

Alternatively, instead of using wrangle to add normal to primitives, use attribute promote nodes
( as in my yellow netbox ).

It seems the ‘issue’ is that when there are both normals on the points and on the primitives that there is a possible ‘conflict’.

In this case maybe because the geometry that was copied to point number one has a normal in the opposite direction compared to that used for point number zero; Possibly causing the normals to ‘flip’ in their corresponding geometry.

But then again, my scene view suggests this isn't the case ( primitive normals turned on ).

So I don't know if there is a ‘display’ bug or something I am not understanding.

The copy to points is copying the bogus normal to the points on your tube. The “primitive” normal created by the wrangle is not used as the normal for rendering, because the point normal takes precedence. Either delete the point normal after copying, filter the point normal from the attributes to copy list on the copy node, disable “Copy Point Attributes” on the copy. Also, if the incoming geometry to the copy already has normal, they will not be overridden by the copy operation, and instead will be transformed appropriately. The Normal sop can be used the create point or vertex normals. This is preferred to using a wrangle to bind N, as there is more control. The reason the polyframe corrected it in your example is because the polyframe recomputes the normal on the points, replacing it with a valid normal. Another note: the primitive normal visualizer does not visualize the normal attribute present on primitives, as it only displays the intrinsic normal. It doesn't matter what the value of primitive ‘N’ was, it was always going to display correctly.

Hi jsmack,

Thanks for your response.

Reading this and playing around with the file again I now have more questions than before.

There seems to be assumptions I've had that needs some more re-evaluation.

Probably best to ‘isolate’ my new questions and post as a new threads.

I think I need to start with understanding better the relationship of normals visualized in my scene to what is getting rendered, or rather how it does or does not afect the render/s.

Thanks again for the ‘chew-on’ factor.