I just thought of this and at least to me would be a very comfortable way to make an animated short. A node scene director. The nodes control all aspects of the animation, the camera, the duration of each node, NLA settings, snippets of timeline animation. Starting node–> preparation load background, activate spaceship, activate retro rockets, activate move clouds, start location of camera, Starting rendering. Kinda thing. What I like about it is you can see the whole schedule from start to finish seen in one panel as nodes. And what isn’t needed to be used is not cluttering up the planning. I realize that there some nodes that need to exist to do all of this. Maybe in v.5.0 if it gets some momentum going.
The general point of nodes - just in my personal opinion - is that they offer procedural control over “stuff”… how a shader looks, how vertex are arranged, for example. And, these various parameters can be animated over time - so, blue water can become mercury, or a solid ball can be transformed into a gel.
So within that scope of thought: what would be the advantage (use case) for using a chained serial node structure, instead of the more traditional approach of a timeline? I don’t think of the animation sequence itself as something I need to be dynamically parametric… but to be fair, I’m only somewhat in the area of “average” when describing my skills as an animator.
Understood, but animation is defined by an action(s) occurring at a specific point in time. So, how would a node structure be advantageous over a timeline with keyframes?
How would you indicate (within the nodes) that “object rotates to 10 degrees at frame 8, then to -10 degrees at frame 20, then to 0 degrees at frame 49”?
one nice thing about using nodes for animation is either reaction to other things (basically simulation, but not limited to that), or randomness (10 degrees around frame 8, +/- 4 frames, for instance), or even both. i haven’t worked with drivers yet, so mebbe drivers can manipulate keyframes on the timeline, which could accomplish the same idea.
as for the ‘how’ for your example, well it’d be an ugly node setup currently. but if we do ever end up with field manipulation nodes, a pair of those (one with ‘how many degrees’: 8, -10, 0, and another with ‘time’: 8, 20, 49) would make that fairly trivial i would think.
Silly and brilliant most likely.
The traditional digital lay it out like a video editor is done that way because its the most easy to comprehend. That is how Blenders video editor and the NLA work. Grab a rectangle and move it side to side for timing, or up and down for depth in the scene. Its easy.
This is now true - Tabs in the nodes… And hopefully will be more true in the future. But at present the same argument seems more valid for the present system. Blender is far more modal than many apps - so what is not important is not seen.
I follow Open Class on Y.T. so I have seen lots of examples of animation timing, start/stop with Geo nodes.
Personally I would like to be able to display panels from the properties area in the node tree.
