You're such a geo nodes master, although it's hard to explain to my partner what it is I was watching, if it was fun or why it was interesting: but it really was! I'm looking forward to possible future parts. I think the space around the intersections may be larye because of the dashed center line; perhaps it starts with a dash there? I'm unsure how your dashed curve node works but perhaps that's the cause. Can I ask a really of topic question? I'm working on creating an animation of a Rembrandt etching and I want to convert a pixel image of one of his works to curves in Blender, so that I can make it animate and appear in a reasonably realistic way. I've tried vectorizing the image with Illustrator and such but it always ignores the line thickness and can't detect lines as individual lines that cross each other accurately. Do you think there is a way I could try to achieve something like this within Blender's geo/material nodes?
I don't think there would be an easy way to do that. The problem is that the input is pixels...or to think of it in generic terms...the grid where a cell can be either a line or not a line. There is no data about the direction of the lines. We would need to construct that. The only idea I have, that might produce an interesting result is to randomly drop points onto the image, delete any that are not on lines, then give them all a velocity and steer them so that they stay on the lines, keeping track of the path they follow as a curve along the way. Kinda similar to this experiment I did when simulation nodes first came out: ruclips.net/video/6k7Mi2ufoa0/видео.htmlsi=CJYQyT36R7CobyDj
@@looseEdges Thanks for the thoughts, I will try if this would bring me any closer, thanks again! I've seen the slime experiment and I have used it in the past to try to 'simulate' rain on a windshield, but didn't come far :)
Thanks Daniel!
You're such a geo nodes master, although it's hard to explain to my partner what it is I was watching, if it was fun or why it was interesting: but it really was! I'm looking forward to possible future parts.
I think the space around the intersections may be larye because of the dashed center line; perhaps it starts with a dash there? I'm unsure how your dashed curve node works but perhaps that's the cause.
Can I ask a really of topic question? I'm working on creating an animation of a Rembrandt etching and I want to convert a pixel image of one of his works to curves in Blender, so that I can make it animate and appear in a reasonably realistic way. I've tried vectorizing the image with Illustrator and such but it always ignores the line thickness and can't detect lines as individual lines that cross each other accurately. Do you think there is a way I could try to achieve something like this within Blender's geo/material nodes?
I don't think there would be an easy way to do that. The problem is that the input is pixels...or to think of it in generic terms...the grid where a cell can be either a line or not a line.
There is no data about the direction of the lines. We would need to construct that. The only idea I have, that might produce an interesting result is to randomly drop points onto the image, delete any that are not on lines, then give them all a velocity and steer them so that they stay on the lines, keeping track of the path they follow as a curve along the way.
Kinda similar to this experiment I did when simulation nodes first came out: ruclips.net/video/6k7Mi2ufoa0/видео.htmlsi=CJYQyT36R7CobyDj
@@looseEdges Thanks for the thoughts, I will try if this would bring me any closer, thanks again! I've seen the slime experiment and I have used it in the past to try to 'simulate' rain on a windshield, but didn't come far :)
Thank god am the first person
I can confirm, you are indeed first :)