GSoC 2024 Sculpt Brush Node

:eyes:

4 Likes

Next: a brush you can use to sculpt nodes. :stuck_out_tongue_winking_eye:

7 Likes

didnt p dobarro do something like this before? :thinking:

Yes… it was just a prototype tho…

4 Likes

That’s a multi-year project, GSoC student cant do that in three months coming in fresh, nor will they be allowed to.

4 Likes

im not sure about that
p dobarro made that working node brush prototype in just a few days. and that was when geo nodes was still a branch called “function nodes” from jacques lucke, full of generic nodes only. remember that?

node brush is not something id want, but maybe its a not that big of task as you imagine. its just one brush. not a brush system replacement

1 Like

Yes. That was just a prototype, a proof of concept by the owner of the module.

A brush that is supposed to allow user to combine nodes equivalent to existing brushes.

To simplify things for most artists, more intermediate nodes would need to be added, so they don’t have to do a bunch of vector math. It would make sense to model them off of the existing Blender brushes, like plane projection for clay-like brushes. This will need design, experimentation and user feedback, to see what artists would find easy to use.

There are currently 33 brush types and almost same amount of deformation modes and relative falloffs. Their uniqueness is not just corresponding to combinations of a dozen of nodes. That is probably closer to forty, fifty nodes.
And plus to characteristic displacement, you have to create nodes for what is common to brushes : Radius, Strength, Auto-Smooth, Sculpt Plane, Stroke method, Stroke Spacing, Falloff, Texture Mapping,…

And the proposal is also talking about what does not exist, yet, in master.

It could be driven by many input nodes: camera direction, surface normal, pen pressure, pen tilt, stroke length, time since stroke has begun, random noise, texture maps, etc.

So, user would probably need a base of around 60 nodes, to feel that the feature is bringing something more than regular brushes.
And then, such brush will have to take into account masks, face sets, symmetry, attributes preservation…

When you know how picky, users can be about their sensation while using a brush, without being able to easily specify the issue …
He thinks that he can do that in an hundred of hours. He is clearly underestimating the task.

1 Like

It doesn’t matter in the end, no GSoC project has ever actually been merged into a main release of Blender (at least since 2.8)

4 Likes

I did not know that. Somehow I always thought that it’s just the ones I followed that were never merged or finished. That’s a really sad statistic. :sweat_smile: :see_no_evil:

C’mon, I mean, c’mon… that’s not true lol…

Ok, show me an example then

Start here https://code.blender.org/2019/09/google-summer-of-code-2019/#outliner-improvements

Great experience for the participants but it’s kind of criminal the amount of time and money that has been wasted on GSoC projects to never actually get committed. Even if the statistic isn’t zero, it’s so close to it that the first thing I think of when people bring up GSoC projects is how much follow up complaining there will be because it didn’t get merged, no matter how good it is.

6 Likes

4 out of 7 for 2019, combined with 0 from 2020, 2021, 2022, and 2023, doesn’t quite equal none, I stand corrected. I suppose it’s 4 out of 42, so 8ish%.

8 Likes

Hmm… I didn’t know that too, but I am not surprised. I always lost track of interesting GSoC projects after a while. I guess it because they didn’t go too far after all.

I also have a hunch that the BF isn’t really expecting GSoC projects to be actually fully finished but rather they’re vetting good developers who can code, communicate well and stay on course long term.

If that’s the case then GSoC is good for Blender anyway since it can both result in some things developed sometimes and also provides an opportunity to discover new developers.

9 Likes

Indeed, similar features from GSOCs of last year, 2022 or 2021, ended up in recent release or became part of a bigger refactor to finish in future ones.

Except select random, that is in Blender :

We had improvements of waveform loading in VSE in recent releases (4.0, 4.1), and little bit of Outliner’s improvements in 4.1.

We have that, now.
https://archive.blender.org/wiki/2024/wiki/User:YashDabhade/GSoC2022/FinalReport.html
https://archive.blender.org/wiki/2024/wiki/User:JeffreyLiu/GSoC2022/FinalReport.html
As explained in final report, Cycles X added Light Tree and Path Guiding.

We have a Curve Pen tool and a Fillet Curve node.
https://archive.blender.org/wiki/2024/wiki/User:Dilithjay/GSoC_2021/Final_Report.html

Here, it is explicitly written in Final report that changes were merged in 3.0.
https://archive.blender.org/wiki/2024/wiki/User:HobbesOS/GSOC2021/FinalReport.html
And VSE Strips previews were also a feature that ended up in Blender 3.0.

In 2020, that was a GSOC. That is a reality.
https://archive.blender.org/wiki/2024/wiki/User:Filedescriptor/GSoC_2020/Report.html

That is a feature inn 4.0.
https://archive.blender.org/wiki/2024/wiki/User:Ankitm/GSoC_2020/Final_report.html

That is also available when using Mantaflow liquid sim as written in final report.
https://archive.blender.org/wiki/2024/wiki/User:Sriharsha/GSoC2020/Final_Report.html

So, that is not true that ratio is 1/8 for last 5 years. There was no year where GSOC’s work had a null impact.
Some GSOC were related to regression testing files. Some to flamenco.
So, they were not supposed to deliver features for Blender users.
Some were about complicated physics that would require more time.
And many were outdated by refactors running at same time, or postponed because of refactor supposed to happen later.

There was no lost of important great ideas. Most of practical improvements have been done, anyways or are still planed to be made.

5 Likes

Well. To answer @Metin_Seven question, there is no display option in Blender for that.
But you can exploit the fact that FallOff Overlay doesn’t support adaption to surface normal.
By enabling FallOff overlay, Cursor stops to follow mesh normals.

4 Likes

Yes correct that option only disables it for no reason I think (it shouldn’t have to disable the other, just rather combine it)

When I’m focused on sculpting I find it distracting that the cursor keeps changing in shape, and I want the cursor to show the brush size and falloff in a clear, constant, circular shape at all times.

In ZBrush I’ve always turned the follow normals option off.

Thanks for the tip, @zeauro , I’ll try it. :+1:

1 Like

nice hack @zeauro :v:

@Metin_Seven a right click select request for that functionality would be nice

1 Like