Advanced vertex color features for polygon2d art #35
Replies: 3 comments
-
@Xrayez, core and module are also applicable (some things I've described as core--like editor changes--might be considered module if I understand it correctly). Basically each piece whatever way it can be done, accepted obviously better. 2 sidenotes:
Tedious enough (w/o vertex painting) to edit as composite (dozens of clicks for color presets to change all points of base color) which is one reason I didn't make it more advanced (I wanted as a design on the balloon in a 3rd color, would require an advanced of a mesh to look good). To support animations would be even more tedious to edit (as composite wouldn't work). (note, example lacks AA and the color from the balloon to the string is odd because I didn't use the mesh trick to fix it)
|
Beta Was this translation helpful? Give feedback.
-
I'm not sure what could be an objective criteria to assign labels like this... In theory almost everything is possible to implement via GDScript and/or C++. 🙂
Not sure if applicable, I've recently implemented boolean operators in 2D which is mostly inspired by CSG nodes in 3D, see goostengine/goost#39. In there, |
Beta Was this translation helpful? Give feedback.
-
I'm no expert but I assume the criteria is based on suitability. Many things I've described likely can be done in some 3rd-party way, though they have varying suitability levels for how much they should be in core. Many things are suitable in idea, but just need to pass the "do enough people want this?" question (that's how I organized the sections). Which, with polygonal art being clunky/non-viable now is really a chicken-and-egg situation. I mean if there were an amazing polygonal art suite (particularly if it could be installed to Godot where it'd be accessible to multiple projects by default/simple on-switch) I'd be all for that too. Things like the editors are definitely possible there, but I feel like the more technical stuff (again: shader context and vertex color normal/specular maps) likely wouldn't be viable outside of core. There's also the problem of integration, something like the default UV editor would be mostly ignored/clutter if you're using the 3rd-party tools. Also compatibility if removed (aside from when results can be carried over and original pieces backed-up at least) but I guess that's a "it hurts when you do that? Don't do that then" issue.
For me mentioning CSG in passing, yes. Beyond that actually doesn't matter because I wasn't thinking properly, making the internal polygon editor more competent via autogen and quick tris/quads would be a significantly faster+easier workflow (esp. w/quick tri/quad brush mode, also using auto-gen after every-other-face is filled in or when only needing a few faces for details). If you ignore quick tris/quads, the train of thought I had was that you'd save time if creating internal vertices and faces at the same time. |
Beta Was this translation helpful? Give feedback.
-
Context
2D polygonal art is currently possible but not really viable due to the current workflow. There isn't even a painter currently, which I'm hoping gets added via core in 4.0 or some later 4.X release. reduz has talked about it before (in my issue 25120 mentioned below), but now @groud is working on editors.
Most of what I'm suggesting here I feel is fully capable as a core/general concept, although it is more advanced and I can understand that it might need to exist as an add-on at least to start with due to the (currently) niche status of polygonal art. That and even if it were commonly used, I could see this as a sort of "polygon2d+" add-on for those who actually need said features for the sake of cutting bloat (in the grand-scheme more than for the features on their own). Most features mentioned here I expect will need to be created by contributors anyways.
This is also related to godotengine/godot-proposals#1651 (allowing polygon2d to have normal/specular maps). I figured it might be better to create this for discussion rather than having it there.
Problem
Vertex colors need to be both faster/easier to create as well as giving more capability and modularity.
Suggestion
Less related, less core
Both of those relate to this example:
Some discussion on the vertex color trick in this reddit thread, but there is an overlap of the information here.
Even less related, but more core
Beta Was this translation helpful? Give feedback.
All reactions