Retopology Design Task

Retopology design task

https://developer.blender.org/T67997

17 Likes

Not a fan of a new mode just for that. Blender has far too many modes already.

9 Likes

I’m looking forward to seeing how this works. It sounds kind of like they’re planning to add features very similar to Retopoflow, which I won’t object to.

I agree. I think this makes more sense as a new object type instead of a mode. Each object type can have different modes already (only armatures have a “pose” mode, etc…). We could avoid adding an extra mode that you won’t use in 99% of your meshes by having a “retopo” object type and converting that to a normal mesh when you are done (maybe with a button on the object data tab).

I think this approach would work better for the skin modifier as well. When people are new to blender and don’t know what the skin modifier does, they may end up adding it to dense meshes causing blender to freeze.

I’ll just wait and see how it turns out. I’m not against it being a mode, nor would I be against just putting some tools in edit mode or whatever.

The fact that Blender has any retopology capabilities at all is by sheer accident rather than deliberate design. So I’m just happy to see Retopology getting attention, regardless of specifics.

3 Likes

Exactly! They should unify things more, instead of spread things apart even further.
I’d like to have access to all the modeling tools when retopoing, so yep, edit mode would be the right place.

2 Likes

A proper shading to help retopology using the Topogun method is strongly needed, “hacky crashy” workarounds no more if Team Devs can implement it officially.

2 Likes

There’s no reason to think a retopo mode wouldn’t have all the same editmode tools +tools that are just for retopology. I mainly just don’t want to have yet another mode for normal mesh objects in the mode switching menus. Six is enough, or at least there would need to be a better reason for adding another mode.

Or at the very least, the workflow wouldn’t be smoother than it would be if you were to use a special retopo object. If retopology is a seventh mode on every mesh object, then the workflow would have you switch to retopo mode, build the retopologized mesh, then finalize it somehow as a new mesh object that will also have a retopo mode menu in its options.

If we have to convert the work done in retopo mode into a normal mesh, then we might as well just have a new object type that you have to convert to a normal mesh when you are done. This way you have a similar workflow, but we won’t have to memorize a new direction for the mesh object pie menus.

Might you be able to offer a screenshot of that, since as someone who never has used topogun, I’m having a hard time visualising what that might be like in my head.

By principle, I am against another mode for retopology.
Tools for retopology are mesh editing tools. if they are useful for retopology, they can be useful during a mesh editing session.
If we create a better snapping for retopology, why will we have to suffer a worse one in edit mode ?

2.8 is supposed to adapt display to workflow.
We are able to set a different display in a different workspace for same mode.
Creating a new mode just to use another display does not make sense anymore.

If Loopcut tool uses preselect highligting, it is because it makes sense for tool.
But a recurrent demand about this tool is to make it faster by creating a loop where mouse pointer is.
So, preselect higlighting often corresponds to slow tools.

  • A way to draw ‘patches’ which are auto-filled with quads
  • A brush that lets you ‘push’ verts around while snapping to the underlying target
  • A tool that lets you draw lines across limbs, which will automatically be filled with quads

These tools would be good edit mode tools.
Currently, fill tools (Grid Fill, Bsurfaces, Fill Holes, F2, etc…) are in edit mode. So, a way to draw patches would just correspond to a preselect highlighting drawing for same result.

A brush to push verts in edit mode that will not be more exotic than proportional edit mode.
We already do that for UVs. We called that UVsculpt but that occurs in Mesh Edit mode.
We could have that in 3DView with a different name and rename UVsculpt to be more consistent.

A tool that let you draw lines around limbs : we have a knife, a bisect in edit mode. Do you really think that would not correspond to same type of tools ?

These kinds or large behavioral changes for how snapping, display and tools all work don’t fit well inside any of the current modes, and other apps essentially do something similar. Just like Sculpt mode is a separate mode for mesh editing, separate from Edit Mode, we think Retopology mode should do the same.

So, that conclusion is just not true.
In Sculpt mode, you are using almost only brushes. Mask/Hiding/Remeshing operators are exceptions.
Brushes are completely different tools than Edit mode tools, dedicated to deform geometry.
Edit mode tools are deforming, generating, cutting, splitting, merging etc… selection. Between each step you do a selection.
Would there be a need for a selection in Retopo mode ? Of course. So, what step is different than in edit mode ?

I would accept as a point that Edit mode has too much tools to make them easily discoverable for retopology purpose.
But for that case, that would be a good occasion to show how UI can handle customization to simplify such complex mode for another specific task.
A different workspace with a different toolbar that would be great.
But a new mode just for a different toolbar : no. Original design does not imply that we need to create a new mode, each time, we need a new toolbar.

4 Likes

100% agree with you. Could you add this in the comment section of the task…? You just explained a LOT better than I could why this shouldn’t be a separate mode.

If the task is just to explore ideas and define the design of the new retopology tools/mode for Blender I think is better to have this conversation going over there than here.

Agreed. It’s overkill.

I gotta say I entirely agree with you. Blender is flexible, let’s not lock ourselves into yet another context within which we won’t be able to use edit mode tools (and edit mode couldn’t access retopo tools? makes no sense).
Retopo should be a workspace preset rather than a mode, using existing drawing/shading models (x-ray on the new model, solid on the old model for example) and there should be a new scene property targeting an object (or a collection) as the mesh to be retopo’ed. I mean, that’s one way it could be done.

I mean modes are there to manipulate different data structures and access different toolsets, but this task (retopo) is fundamentally about creating and deforming geometry, and that’s the job of edti mode - it does not justify having yet another mode, especially when it comes with so many drawbacks.

5 Likes

auto-retopology and semi-auto should be in the sculpting mode no question about that, but for a final retopo mesh then it’s acceptable to have it in a seperate mode because that’s what most softwares do, because if it’s in edit mode then you’ll have to hide it behind a basic tool/brush with little options as possible.

1 Like

I don’t get what you mean. Why would retopo tools in editmode need to be basic? There is already a free retopology addon called bsurfaces (comes with blender) that works entirely in editmode. I always felt that most of the trouble we have with retopo is the setup process. We could possibly use more options on how to visualize the retopologized mesh as well.

To me these things are best served with a new object type that has an edit mode that is geared toward retopology. All the editmode tools that are useful could go there (we probably don’t stuff like this spin tool), along with anything that would only make sense for retopology (I don’t even know what that would be).

new object type that has an edit mode that is geared toward retopology.

so basically you’re flipping the idea here instead of a new mode u want a new object that has a new “retopo edit mode” specifcally for that type, that makes it even more complicated i think,i would imagine this instead you select your high mesh enter retopo mode and immediately start retopoing with points,quads…etc it does everything for you.
have you seen other dedicated retopo tools like retopoflow for example? they come with a handful of tools with lots of options so that’s why putting them in edit mode will have the effect of becoming just basic like polybuild.

This doesn’t make sense. A tool is a tool. It’s power/features has nothing to do with the mode it is in.

There’s really no need for a new mode.

In my opinion, a dedicated object type is just a different way of adding a new mode (but more complicated on the user end because even though it’s not near as unique as text or metaballs, it wouldn’t be able to access many general mesh and object tools and properties).

As for whether a new mode is needed, it depends on whether the toolset would be unique enough to make it rather difficult to pass them off as general editing tools. I’m thinking of powerful tools like being able to draw feature lines and patches for automatic retopo (on both a local and model-wide scale).

2 Likes

unless you have 10s of features that you have to split them into each tool, edit mode will be overflowed with tools by then.