I'm not sure if this is a bug report, feature request or a demostration of my ignorance. I'm using 1.0beta2 (2b71d25, 2019-12-03)
I have a shape with a gradient fill, I've tried both mesh and linear gradients, and I'm applying the bend LPE. The shape is tranformed but the gradient does not move with the shape.
Is it suppose to, or should it or I'm I just ignorant on how to make it move? I've tried it with both move gradient with object on and off. I realize the move is different than applying a path effect but I tried it anyways.
I can see why the linear gradient may not make sense since we can't bend the gradient paths, I'm not sure why it couldn't work with a mesh but then I have no idea about the math being used.
So maybe the real question is how to apply LPE to the mesh gradient paths.
thanks for your time and effort. I have used inkscape off and on for years, just back at it again for some yule presents
I think there is already a bug report about the mesh gradient not bending. I'm not sure if it's in the current bug tracker though (just don't remember).
If you don't find one in GitLab, you can either migrate the one from LP (if you found it there) or just make a new one in GL. If you don't want to get involved on that level, let us know, and one of us can do it.
Bending a gradient is always a trick question. Was requesting a way to bend gradient meshes previously myself but can see how it can get rather complicated.
Like, the bend lpe bends only the vector data of the paths and not the fill attributes for a start.
Fine, then for a simulation -or, this case for a guide on constructing- let's first draw the mesh gradient's layout as a compound path, depicting every segment with a path with a fill and no stroke.
Then, bend that grid and see what it will end up. Tangent lines will remain tangent to the trajectory path/core path, and parallel paths of the pattern should be stretched to a parallel shape to thetrajectory/core path.
The latter one, to fit with the desired shape, will have random nodes placed here and there, for a geometrically plausible look.
Thinking now, this grid was representing the gradient mesh -you'd need extra mesh nodes added just to show the calculated geometry,
however each mesh node represents a new colour!
Also current meshes are construct on a four sided cell so it won't happen if the new "extre nodes" are not added in a way that still provides the rows and columns layout.
That in mind, you can adjust manually a mesh gradient onto a bent grid manually, "eyeballing" it as much as possible.
The extension/visualize path/plot handles can be helpful with snapping to nodes (if there are no extra nodes added to the path anyway).
Then, what else can be done?
Depends on what you are really after.
You can create a raster copy of the gradiented pattern, trace bitmap it and bend the group.
With enough scans it may look fine.
256 steps is the max so instead maybe you can try a halftoning base pattern.
If the gradient is supposed to go along the path, then a scatter effect and interpolating a solid fill colour in between can represent the same look.
If the gradient is supposed to go tangent to the core path, you can try using custom markers -can dig up a tutorial for that if interested.
Some cases if the shapes are geometric, you can make a smooth transition by combining radial gradients too.
Beyond the Basics using a LPE on an object with a gradient
Inkscape.org
I'm not sure if this is a bug report, feature request or a demostration of my ignorance. I'm using 1.0beta2 (2b71d25, 2019-12-03)
I have a shape with a gradient fill, I've tried both mesh and linear gradients, and I'm applying the bend LPE. The shape is tranformed but the gradient does not move with the shape.
Is it suppose to, or should it or I'm I just ignorant on how to make it move? I've tried it with both move gradient with object on and off. I realize the move is different than applying a path effect but I tried it anyways.
I can see why the linear gradient may not make sense since we can't bend the gradient paths, I'm not sure why it couldn't work with a mesh but then I have no idea about the math being used.
So maybe the real question is how to apply LPE to the mesh gradient paths.
thanks for your time and effort. I have used inkscape off and on for years, just back at it again for some yule presents
oh, I just tried changing the fill from a mesh to a solid and then back to a mesh after the lpe was applied and it worked, sorta.
The mesh was not bent but it was applied over the new bounding box. This is better, still not what I was visualizing would happen.
I think there is already a bug report about the mesh gradient not bending. I'm not sure if it's in the current bug tracker though (just don't remember).
If you want to investigate and make sure there's a report, you can. The old bug tracker (where I think it is) is https://launchpad.net/inkscape and the new one....there are sort of 2 places, one is here https://gitlab.com/inkscape/inbox/issues and the other https://gitlab.com/groups/inkscape/-/issues The LP search feature is almost rudimentary. The fewer search words you use, the better. I haven't tested the GL search much yet.
If you don't find one in GitLab, you can either migrate the one from LP (if you found it there) or just make a new one in GL. If you don't want to get involved on that level, let us know, and one of us can do it.
Hi.
Bending a gradient is always a trick question. Was requesting a way to bend gradient meshes previously myself but can see how it can get rather complicated.
Like, the bend lpe bends only the vector data of the paths and not the fill attributes for a start.
Fine, then for a simulation -or, this case for a guide on constructing- let's first draw the mesh gradient's layout as a compound path, depicting every segment with a path with a fill and no stroke.
Then, bend that grid and see what it will end up. Tangent lines will remain tangent to the trajectory path/core path, and parallel paths of the pattern should be stretched to a parallel shape to thetrajectory/core path.
The latter one, to fit with the desired shape, will have random nodes placed here and there, for a geometrically plausible look.
Thinking now, this grid was representing the gradient mesh -you'd need extra mesh nodes added just to show the calculated geometry,
however each mesh node represents a new colour!
Also current meshes are construct on a four sided cell so it won't happen if the new "extre nodes" are not added in a way that still provides the rows and columns layout.
That in mind, you can adjust manually a mesh gradient onto a bent grid manually, "eyeballing" it as much as possible.
The extension/visualize path/plot handles can be helpful with snapping to nodes (if there are no extra nodes added to the path anyway).
Then, what else can be done?
Depends on what you are really after.
You can create a raster copy of the gradiented pattern, trace bitmap it and bend the group.
With enough scans it may look fine.
256 steps is the max so instead maybe you can try a halftoning base pattern.
If the gradient is supposed to go along the path, then a scatter effect and interpolating a solid fill colour in between can represent the same look.
If the gradient is supposed to go tangent to the core path, you can try using custom markers -can dig up a tutorial for that if interested.
Some cases if the shapes are geometric, you can make a smooth transition by combining radial gradients too.