The reason I wanted to start this thread is that Blender over the years has acquired a really nice toolset for areas like modeling, the only problem is that some of these tools are pretty dumb (and because of that the user might have to clean up the results).
Now in the case you’re going to bring up the point of “the app. shouldn’t try to decide for me”. Let me say that there’s numerous cases where the intended result is almost universal and can safely be decided by Blender itself, allowing you to focus more on the vision rather than fiddle with the tool to get the desired result.
In Blender, the tools can be made smarter through a combination of better functionality and a way for Blender itself to read the context of the situation (the latter perhaps looking at what faces are selected, what mode you are in, ways the selections relate to each other such as where they’re facing, ect…). To get Blender to the next level, I assume, would involve bringing in some context reading API and eliminating some glaring weaknesses in existing tools, for example.
Bevel Tool; The inability to auto-merge vertices so as to avoid overlaps can be annoying, as it can be difficult to get nice thick bevels on complex volumes. This can also be applicable for the boolean modifier which would actually allow for something close to Modo’s Mesh Fusion results.
Bevel Tool; Why can’t it just default to vertex bevel if there’s no edges or faces selecting, someone might otherwise think the tool is broken until he learns the ‘V’ key.
Inset Tool; The lack of automerge again, sometimes you have to stop at a distance and merge vertices manually
Bridge Tool; The context analysis should be able to determine if your selection islands are all facing each other, and default to ‘loop pairs’ if that is the case
Bridge Tool; For the segments option, there’s no concept of weighting for the direction of the normal for the initial selection, meaning in some cases the bridge starts off going backwards and overlapping geometry (rather than having perhaps a smooth and clean ‘S’ shape).
Bridge Tool; Perhaps some context code can look at what selections can count as a volume and what is planar (perhaps an error could be thrown if you try to bridge all six faces of a cube to a flat surface for instance).
Extrude Tool; Context analysis could look at the total disparity between the values of selected face normals, so if it’s high the tool also shrinks/flattens along the normals.
Grid Fill; Would it be possible to try to create an all-quad topology if the edges don’t match up perfectly (rather than create poles at certain points)?
Convert Text to Mesh; Unless I’m really missing something, the conversion process still uses the crappy scanline filling from the NaN days as opposed to the nice clean algorithm used when pressing ‘Alt F’. This means that certain shapes like the + sign actually gets invalid geometry (zero-area faces).
UVmapping ‘reset’ option; Ngon UV’s are circles (and as a result the only clean way to upwrap certain shapes is to use the view projection). It’s almost as if it’s never been updated for Bmesh.
UVmapping ‘reset’ option; The tool is known for sending textures in different directions (even if they are part of the same quad-based surface), the tool should try to make directions consistent when possible (which I know is possible considering how well the triangulate tool does it).
Off the top of my head, these are the immediate cases for me as to where I think Blender is lacking intelligence (cases where the vast majority of users are likely seeking the same results). Considering that 2.8 is supposed to be the big series where workflows and documentation get broken, that might be a good thing to take a look at.
Add your ideas on where tool intelligence can increase (and please don’t use this as an opportunity to bash the developers or take potshots at users). Also try to keep any suggestions in an achievable range when it comes to development.
1.Bisect tool: needs several angle restrictions. Local axis XYZ and World axis XYZ. Also a modifier to change the angle incrementally.
For example i want to cut an object on the local x axis and i want to cut it with an 10° angle.
2.Extrude tool:
There are currently 4 different modes of it and they should be streamlined into one tool.
1.pbvhtree based painting and sculpting in 1 stroke painting multiple vertex color channels or even multiple images
compartments - click object - open logic or node panel
-opens hovering window (allow multiple objects to be open at 1 time)
copy/paste nodes and or logic
particles get entry on timeline for animations per particle that is normally hidden including birth and death
The image also shows a little shortfall with which edges it chooses to dissolve (that being it does not try to keep the edges with the least acute angles with other edges). Such acute edges can get in the way of other tools such as bevel and inset.
The same concept can also be applied to make the Boolean Intersect tool smarter.
Lets talk about Modifier. I would really like to have the option to apply Modifier to an object AND all of its child objects.
Like create a null, in the outliner throw a selection of objects into it and apply something like a solidify, subdiv, array or mirror.
Symmetrize option in Mirror modifier. Cut model on given axis, throws away one side (controlled by menu) and then do actual mirroring. Some other programs works that way and it’s super useful, especially with rotational symmetry controlled by object.
Monio: Agreed! I use Automirror for this now, but it’s destructive.
Speaking of modifiers, wherever an object can be used in a modifier, have a button besides the object-picking one that just creates a new empty in a logical location (so in the case of the mirror modifier, at the object’s pivot point)
Extrude; When extending a profile (ie. just a line of connected edges), have the tool extend the UV coordinates as well (because even if the user ‘may’ not want this, it is still better for many cases than to simply leave the X or Y scale at zero).
Extrude; Prevent the new geometry from inheriting the ‘pinned’ value (from the UV window) if the vertices being extruded from is pinned (otherwise you end up doing an unwrap and wondering why it didn’t seem to work or why it produced a messed up map).
Shrinkwrap Modifier; For the nearest surface mode, add a bias that causes it to favor nearby geometry opposite of the normal direction (not only does it make it more sense, but it means re-topology in very tight spots would become far easier).
Mirror; Actually recalculate the normals when this operator is invoked (currently, it feels like a glorified version of just pressing ‘S’, then ‘X’ ‘Y’ ‘Z’ and setting the scale to -1).
Extrude; Piggy-backing on the last one, recalculate the normals if there’s no non-manifold geometry in the result after the invoke. This goes back to that context system idea in the first post.
Symmetrize; Use the local axis as the direction of mirroring and the object origin as the center of mirroring (the tool is a good idea, but it’s currently a tricky deal to get good results).
MOUSE WHEEL reliant tools cannot be remapped to work with Tablets(Wacom):
bevel tool - cannot change divisions
proportional edit - cannot control size
others
// I personally use AutoHotkey with numpad ± emulating mouse scroll
BEVEL op:
if you choose a large area and are zoomed in on your mesh, beveling center is at very awkward place on/off screen making it difficult to find and control bevel. Should be at screen center or mouse position?
BEVEL MOD:
remove spiking. I have applied a little hack inspired by 3dsmax Clamping that allows user to specify distance instead of automatic calculation( https://youtu.be/BDjsJin78sQ?t=40 )
MIRROR:
- custom split normals (e.g weighted normals for games) when used with MIRROR get messed up.
REMOVE DOUBLES:
currently welds to random nearby vertex. Would be good to have option to weld to common CENTER which would cause smaller mesh distortion. Also vital for below mentioned modifier.
* REQUEST: AUTOMERGE/WELD/REMOVE DOUBLES MODIFIER:
It can be part of decimate or anything else, but it’s extremely useful modifier. Ex:
Ability to close SCREW revolved shapes
Enables more stable Bevel modifier behavior after using Booleans as nearby points are merged.
I will shame myself by positing minimalistic BMO_ solution that works, but I have been told off not to do that. Asking developers for proper solution for benefit of all:
UI: The amazing PREVIEW panels (template_ID_preview, template_icon_view etc) currently require pressing on ICON to open the selection. Would be great to have option to IMMEDIATLY open the selection window, bypassing needless click. This can be useful for example in Sculpting when choosing a brush (similar to Zbrush):
Hi!!
Though perhaps not very usefull in most advanced 3d working workflows, I’d find extremelly usefull having a rulers and guides system, similar to those in vector and bitmap drawing softwares, in order to help fast concept 3d modeling.
Aren´t there plenty of material libraries out there for free download? Like the Chocofur Shaders that AFAIK come packed as shader node groups that sort of simulate a so called uber shader?
