How to read output only geometry nodes correctly

[proc24]

Hey there,

While I was watching the Distorting the Trunk to Add Roots lecture I was wondering if I understood reading geometry nodes correctly, so this is not a question directly about the lecture, but rather about geometry nodes in principle.

What I have been wondering about is how to properly read output only nodes, like spline parameter, Normal, Position, etc., specifically where the normals or positions come from which they output.

So imagine this example from the lecture linked above:

Do I understand correctly that both the position and the normal node operate on the geometry thats input at the very top right with the set position node?

Another example would be here:

Do I understand correctly that the first spline parameter and both named attributes read their values from the “Adapt Trunk and Branches Radius” Curve input pin? And if so the second spline parameter would then read its values from the “Capture Trunk Factors” Geometry input pin, is that correct?

I somehow found that less confusing when working with procedural materials in Stephens other course.

Thanks in advance!

[zeRgenTa]

Yes. As a general rule, data is read from the geometry node that it is plugged into.
There are ways to get it to read from other geometry, but I don’t think it is taught in this course.

If you haven’t been using the viewer node, it may be able to help you visualize things that you aren’t quite sure about. They even added a new future to it in 4.1.

Just turn on “Attribute Text” in the overlay, and it will show the values right on the geometry.

[RayMobula]

Glad you are wondering too. I had the same question at that point in the course.

I know there is an answer above, still, here is a link to my post from earlier today about this topic.

Replaced points by tree faces to make better mounds

Thank you. I like the course. Unfortunately, at times it is not clear why we do things, or explanations aren’t clear why things happen.* Thus often end up looking up the node’s documentation. And yes, there is a good chance I have to get the logic of another node-based software I used for decades out of my system. But then, this is complaining at an extremely high level, since I learn a lot from @Stephen_Wood and I’m engaged in the material - and this should be the measure how good a course …

[RayMobula]

Oh, the attribute feature is neat!

even if a bit crowded

[zeRgenTa]

It might be a little crowded, but you can always zoom in.

[proc24]

Alright Im glad I understood it correctly, because I wasnt so sure at times . Also thank you so much for both the explanation and especially that tipp with the viewer node @zeRgenTa , I didnt know about that yet, seems super useful!

[proc24]

Thanks for linking your other thread! I read it and there is nothing to be added from my side there.

I still have to write my “feedback” which I will do on monday, but its more or less what you said. I also had to look up the documentation more in this course than in the material nodes course for example. Now Im not sure if that is something that can be fixed, because when you make a beginners course and explain every little detail chances are that the course will blow up in terms of length and people might loose focus/interest. There is also the chance that geometry nodes themselves might be a topic where you will end up working more closely with the documentation than with other topics, though this is probably less of an issue than balancing the course complexity.

At the end of the day this is a very informative and well structured course I think. I learned a lot and and as I said in my other thread I think the knowledge gained in this course opens up a lot of possibilities and workflow improvements so its really valuable! If Stephen goes into a little bit more background on to why we do things there would be nothing I could say which could be improved.

I guess this answer can be my feedback thread after all