Manipulate texture coodinates


I´ve watched the great video from Bartek Skorupa about texture coordinate ( Thanks, Bartek!

I´ve created the Wood Grain Base, based on the video, but, I´d like to ask how to rotate the the texture. I know that Bartek has shown the formula, but I´m not programmer. And he didin´t explain how to create the node setup for that.

Anybody could help me with this?


In the attached .blend you’ll find a node group to rotate a vector around an axis by a specified angle.


cycles_vector_rotate_node.blend (100 KB)

1 Like

Thank you Stan! Can you help me to use it? The image shows where I plugged it. Is that right? I´ve made some tests, but with no success…

You need to also plug something into Axis input, or edit it directly in the node. X axis is (1,0,0), Y is (0,1,0), Z is (0,0,1).

EDIT: I’ve attached an example: rotate around the normal.


rotate_tex_coord.blend (111 KB)

I tested in a cube, and it worked! I think the problem is my object… But, thank you very much, anyway. :wink:

Oh, you’ve replied before I’ve finished editing. I’ve updated my post above. And you’re welcome :slight_smile:

Thanks, again, Stan!

Excellent tutorial/speech, but unfortunately shows a strong Cycles weakness - a lot of nodes needed for basic things making the procedural texture too slow in the end to be useful as a rendertime effect, and have to be mapped out to a normal texture. In which, isn’t there better options out there (does require UV unwrapping or simple shapes like that floor).

As Bartek said in the talk, he bakes the result before the final render, so all that extra calculations are only being made once. :slight_smile:

Calculations are made a lot of times during construction of the material, probably 100 times more if I were to do it since I don’t really know what I’m doing :slight_smile: But yeah, reduced to regular sets of textures before the final render. Probably a lot more efficient with dedicated procedural texturing tools (for the pattern variations) than using Blenders lacking (or rather, too basic) nodes. Like why do we have to construct trigonometry nodes (and learn the maths behind what we want to achieve :)) instead of exposing the inputs so we can connect them to a group input?

Don’t get me wrong. I realize the nodes are quite powerful (a few things still cannot be done due to the nature of the rendering pipeline), but as his setup shows, you still need a whole lot of stuff going to get good control over geometry nodes. His video intro also shows some stuff that is not explained in the speech, like transformation between different coordinate systems, say from cartesian (parallel built in) mapping to spherical coordinates or disk mapping. With my limited maths skills, this is an impossible task. And even when done through nodes, it gets so slow it becomes unusable due complexity.

But I have to say, Bartek is amazing. His knowledge level of Blender is beyond achievable :slight_smile: