Simple Lattice seems great (just tried it), but I can't seem to figure out how to use at at the vertex/component level even though the page says that it can do so. I would expect to find the option to create a lattice at the top of the mesh rightclick/context menu, with it creating a temporary vertex group + lattice based on the current selection. Or perhaps I am just imagining too much ?
it should work like that actually - not sure whats going on right now. EugeneDudavkin has been doing updates for this for quite a while - so not sure :D
I see, I must be missing something or incorrectly installing from the github files. Once installed, does the addon have any entry in the universal (legacy spacebar) search ?
maybe check the console if it errors? Hmm maybe try to rename the folder "SimpleLattice-master.zip" to "SimpleLattice.zip" or so? I rarely use the zip way of installing addons :D
I have a background render, and a 450 frame image sequence, is it possible to combine the two and save out the mp4 with the animation playing in front of the background?
Hi there @kio , I got it figured out and I can now comment on what went wrong/what may need to be adjusted.
- There is no zipped release on the GItHub page, and downloading the source as zip indeed adds the usual "-master", resulting in a zip that doesn't fit with the usual Blender addon specifications. Of course this is to be expected for anyone familiar with the way Github works, but I still think that as soon as multiple files are involved a proper zip should be provided in Releases IMHO.
- Once installed there are a few issues with the way the addon is being accessed.
1 - First, as it stands the operator name for the addon (for universal search) is "create lattice", which IMHO is too vague as it sounds too much like a default function. It could just be "SimpleLattice" or "Simple Lattice", just like the add-on name. Or, the addon itself could be called "CreateLattice" - either way would make more sense than the current naming.
2 - Secondly, it being in the RMB context menu is a bit of an issue because of the way Blender RMB menus (formally "Specials" IIRC) are weirdly duct-taped. They look a lot like the regular Object and Mesh menus, but not quite. For instance here are the Object Menu and the Object Context Menu side by side. In theory they really should have exactly the same contents (populated automatically from one to the other), but in practice and if I am not mistaken they are two separate entities, which can be populated separately. And even the common functions are ordered differently :
This is obviously not your fault as this is a Blender design issue. But this explains why I originally couldn't locate SimpleLattice for Mesh Edit mode : I was looking for it in the Edit Mode Mesh menu, which I temporarily assigned a shortcut to because I assumed it was the same as the "Specials"/RMB Mesh Context menu.
Put differently : I personally have all the RMB menus disabled, because I don't see the point of them since all their functions (outside of the ones added by add-ons) can be found elsewhere. So when it came to reactivating the Mesh Edit Mode context menu by giving it a hotkey, I mistakenly assigned a hotkey to "Mesh" (VIEW3D_MT_edit_mesh) as opposed to "Mesh Context menu" (which I don't know the command of which turns out to be wm.call_menu VIEW3D_MT_edit_mesh_context_menu).
All that to say that these two things (naming of the addon + the RMB menu oddity) can make the addon a tad difficult to find and use. I am glad I took the time to investigate all this as I now finally have a way to swiftly edit verts with a lattice and that is incredibly powerful :) But still I believe than some things could perhaps be ironed out.
- - - - -
Create Material is fantastic too, I've already been using it quite a bit and it is a big time saver already. Perhaps having the ability to have the color be done through an (automatically generated and packed) Image texture could be nice too, but then it would be a bit of a feature creep of sorts.
ill forward the feedback to Eugene Dudavkin - i agree about the release zip, that really should be there just to download. I dont have a strong opinion about the operator name and for me the rightclick context menu is a perfectly fine place to add this. I think initially I added it also to the 3d view object menu, but that got changed at some point - guess to not litter the menus too much. Would be easy to get this back tho, or make an option for this in the addon preferences.
for create_assign_material - that auto texture generation thing sounds a bit too niche for me, but maybe you can pitch the idea/workflow a bit better?
I suppose the naming aspect also came up on my end because of other lattice-related addons I already use, which are already populating search with other "lattice stuff " making SimpleLattice hard to guess/find :
That said this is trivial to rename on my end, so obviously not the most critical thing to adress :) And indeed I do understand that the way I fully disable all default RMB menus is probably not a widespread practice. Still, it wouldn't hurt to have the SimpleLattice entries also appear in the regular top menus for Object and Mesh respectively. As a matter of fact it looks like the Blender devs are strictly following the soft rule of everything being accessible from the top menus first and foremost, as I can't remember the last time I couldn't find something there.
Regarding texture generation : well, first I'll have to dive a bit deeper into my actual needs for this and make sure that there isn't already a way to do it :) The idea would be to have a tickbox next to the color selection + an input for texture size, causing a texture of the desired color be created and plugged straight into the newly created material. That way the user could start painting right away, which could be useful for all kind of things. But as said this is very much a feature creep. I can't thank you enough already for that handy little addon.
how can I make a light for cycles that can modulate how much diffuse or glossy it affects.
For exampe in eevee, its useful to turn down the the diifuse on some lights, to make more subtle lighting, but in cycles you can can only seem to turn off glossy or diffuse ie on or off
Never tried that.. but if you use nodes for the light.. then there is by default a Light Output node feed by an Emmission Shader .. so.. oh no you meant the effect on other objects.. well Cycles is meant to be an physical correct based renderer.. AFAIK only polarized light can effect the glossiness of objects.. and IDK if cycles support this.. just my two cent 😁 ... ( <== hmm latetly the emojiis doesnt seeem to work for me...)..
yuo can do it with arnold and eevee( see attached), so I a guessing it's not that hard to implement. I can't make it work properly with light path node, comes out a bit weird looking, though it does kind of work
With Cycles lights you can add nodes and do some quite creative stuff. There is a workaround so you can turn off glossiness/diffuse of a light in the properties panel, but it just on or off, not able to use a slider.
Cyles lights really need more default options like in arnold
so cycles just has on and off, where arnold has sliders for lots of different options
@gnoop We can use two connected face corner positions to get a tangent vector.
I'm not too sure how data transfer works but for this using nodes like "capture attribute", "store named attribute" or "evaluate on domain" set to the "Face" domain it will select the vectors from every 3rd corner and store it on each face.
you can use the "offset corner in face" node to shift the data.
At this time, is there any Blender version out there that has a fast enough (for lack of a better way to put it) processing speed for the boolean brush in sculpt mode ? The current implementation (either Line or Box) is just as slow as a regular boolean calculation (of course), but the 30 seconds or so of waiting are really not sustainable in practice - especially compared to how instantaneous the same thing is in Zbrush, 3DCoat, and Nomad.
Of course the instantaneous Clip line brush is good enough-ish for a lot of cases (even though I wish it had a polygonal lasso option), but still : sometimes one really needs to remove volume rather than just clipping things in.
I wouldn't even mind automatically stepping down/remeshing the model down to a fixed, pre-established low enough polygon count if that's what it takes (since such big boolean operations are mostly useful in the blocking stage anyways)
In terms of UX I would say that ideally the operation shouldn't take more than 2 seconds at max.
Or even, something along the lines of converting the current sculpt object to a gazillion small cube objects, selecting the ones to delete using a regular object lasso selection, and then converting back to a remeshed model ... essentially working in pseudo-voxels, pretty much.
Right - I mean to be fair the speed is okay when using the box variant on a rather coarse mesh - but with the lasso one being made of multiple segments, it very quickly goes back to the issue of having too many calculations to perform. So overall it's pretty much unusable at this point ...
(So I guess it needs two things really : a polygonal lasso, and faster processing)
Well, if anyone has a clever workaround I'm still all ears :)
Hi! I have a general hinderance now and then: Adding/switching between properties for multiple objects at once. For example adding a second UV map to all selected objects or make the second UV map the active one for all selected objects.
In the past I would either do it manually, or - when facing an overwhelming amount of objects - hack together some python script (which is then super specific and gets lost). How do you deal with this cases? Curious if there a functionality I don't know about yet 🤓 I know there is "copy to selected" for property values (or alt+clicking the property), but this doesn't work in the described cases.
Regarding trim brushes: Better performance on more dense geometry would be very nice indeed. Until now I only found use for the trim brushes during blockout stage when the meshes are fairly low-poly.
Yeah, this isn't currently possible. There's been talk of making everything apply to your entire selection by default, but the topic's been stuck in option paralysis hell for years now.
I'm creating a procedural brick wall shader at the moment, and I'd like to know if it is possible to somehow map a color ramp to a single brick individually using the Brick Texture node. Meaning, I'd like to map a color ramp from black to white to one of the bricks, a slightly different ramp (rotated differently) to the brick next to it, and so on. I'd like to use the result as a displacement, so the individual bricks pop out each one skewed in a slightly different way, to add variation.
Is this possible using/ building upon the 'Brick Texture' node? Can I map a color ramp to each brick individually somehow? Or do I need to set up something like this manually in a different way?
Is there a secret to making vertex snapping in Blender work 100%? I've just struggled with it earlier today when trying to adjust a model + associated blendshape/shapekey source objects. Several objects in view, one active and in edit mode and the other ones deselected. I managed to make most vertices snap to where I wanted them but some resisted and would not snap to that one vertex where I wanted them (but would willing snap to everything else).
Any tips to make this manageable (and save my keyboard from the repeated beatings) would be welcome. :)
Also, is there possibly a way to transfer/copy/paste shapekeys and their states from one mesh to another? I have them all set up and sorted on one model and it would be just great if they could magically be made to appear just like that, same order and same ranges/slider value settings on another, identical mesh that's already skinned (I'm aware I can transfer skinweights but trying to get a process down to quickly update a bunch of shapekeys done elsewhere).
Re. shape keys : I don't know it that would be of any help, but : merging two models brings in the shape keys of one into the other, similarly to how merging models preserves skin weight data. Not sure if that is precisely what you are after, but perhaps this could be a hint at a solution ?
Overall I find the very barebones/"straight to the metal" way Blender deals with weights and blend shapes to be a life saver really. It just works :)
@pior Merging doesn't seem to help as far as I can see since the shape keys only affect the merged parts of the mesh and deleting the old part mangles the vertex order. In that case it's probably better to simply transfer skinning and whatnot and swap out the whole mesh on the armature.
I have an asset containing a series of shapekeys that are set up and named in a specific order, some with custom value ranges (beyond 0 and 100) and certain default values set. I use that asset to develop shapekeys as well as edit vertex normals, paint vertex colors and adjust UVs as needed.
There's an identical copy in another scene, bound to an armature, using an engine-specific material-setup that is not great to directly work with.
So, I'd like to keep working on my isolated asset and simply transfer over every property including those shapekeys without it becoming a clicking-orgy every time I push an update. The list of shapekeys is growing to a point where it becomes a bit of an effort.
Oh and it's not a single surface, rather plenty of disconnected surfaces, so I think any traditional vertex order transfer is a complete no-go. Certainly the kind where you have to select the same element on source and target for it to find its bearings.
@thomasp : considering your own technical know-how I can only assume that you've already tried all the usual approaches hence I don't have much more to suggest really. Besides perhaps having a look at the Surface Deform modifier, which has been very helpful to me when dealing with complex body and facial expression transfers, bypassing the need for a thirdparty tool like Wrap3.
The (recently added, I think ?) Smooth Corrective modifier has also been of great help to me when smoothing out some kinks caused by delicate form transfers.
Outside of that it sounds like you are reaching the limit of your scene/export setup ... so perhaps it is now a good time to refactor it deeply ? Like keeping the multiple parts you mention as different models, or finding reliable ways to fully control the behavior of materials on export (an especially tricky thing to achieve with UE, as there are no explicit rules on how their importer (re)orders material IDs).
@whizzard None of these questions can be answered outside of the specific context of the models your are working on. In some cases shoving meshes inside eachother is absolutely the best approach, and sometimes it isn't. The only thing that never works is to assume that something won't be noticed thanks to the increase processing power (for instance, assuming that it is "fine" to use different number of edges on two cylinders that are related to each other, or assuming that a highres texture will hide flaws. These approaches are always bound to backfire later down the line, whereas the assets that have been carefully crafted are always going to be the most resilient. But outside of that, anything goes.
I would say that the only way to get a feel for what works and what doesn't is by developping a strong knowledge of game art history - by studying ripped models from games that have excellent performance. Nintendo games in general, PC games that run exceptionally well compared to other similar games of the same genre, and so on. This is where you will always find the most well-crafted assets to study and learn from.
@pior: Thanks for the sound advice. I will look at other's topology, which may help me understand how it was put together. I seem to end up with some parts of a model with many polygons, and others have much less or different flow. I see consistent flow and size in professional models, and I want to find a method to get closer to that.
@pior Thanks, yeah my conclusion was that the only thing that looks like it would help me is a way to transfer the morph targets like you would transfer basically every other kind of information with the Data Transfer mod. I was considering looking into making an addon but I haven't touched Python in several years ever since the big API upgrade in 2.8 and would rather just keep on making 3D things instead of digging into that stuff again.
Still wondering about the vertex snapping issue though - that was seriously aggravating. Back in 3ds max I had a list of things to look out for when the snapping didn't work as expected (surface constraints active, wrong coordinate system picked, xforms not resetted, etc) that always solved the problem in the end. I suspect I was dealing with something similar now in Blender but am lost as to what exactly caused it to keep on failing only on certain vertices.
Perhaps an issue with a shapekey affecting snapping source or target vertex even if the shapekey is set to 0%?
I'm trying to set up a dynamic blending of materials using geometry nodes. What I have is an objects that is moving across a landscape, and I would like to blend materials based on the proximity of that object in relation to the landscape.
There is a vertex weight proximity modifier, but the resulting weight map cannot (to my knowledge) be used directly in the shader tree. I would need something like the weights this modifier generates, but in vertex color format.
Can someone give me a hint on how to set this up using geometry nodes? I'm thinking of generating a dynamic b/w vertex color map based on proximity that I can then use as a mask in the shader tree to mix multiple materials.
Are there any resources / tools for building custom pie menus inside Blender? I googled around a bit but could not find much on how to write these manually. I'm looking to create fairly complex ones containing panels with various UI elements, not just the noodly default things.
I can see that its possible to do them when looking at e.g. Machinetools:
I'm asking this because I'm heavily reliant on the Pie Menu Editor addon, I'm literally useless in Blender without that. And it appears to be abandonware, increasingly incompatible with newer Blender builds, keeping me firmly on 3.3 for now.
Looks like the author is not coming back and I'll need to replace all my pie menus done in it which look similar to the ones shown above. 😱
I'm honestly holding out hope for someone (if not roaoa himself) will fix up PME, as it's a massive part of my workflow. Blender addons are open source by definition, so anyone can go in and make adjustments.
Another alternative is 'Serpens', a node-based interface for building your own addons that do export to Python (and are therefore less of a dead end).
Otherwise, you're right: python scripting them yourself seems like the best bet, and I have seen a lot of documentation on this over the years. Maybe it's best if you ask specific questions of things you're running into? Machinetools is a bit too big of a thing to easily decipher I imagine, but pulling apart a simple addon that provides pie menus could be a good first step.
@Michael Knubben At this point I've been holding out for years for a savior. Time to accept that the author has moved on (years ago) and I figure this addon is probably a bit too complex/clever for someone new to step in and decipher it and keep maintaining it. Best case to expect might a quick patch to fix the biggest compatibility issues. If only it became part of Blender itself.
Basically I'm looking for how to build these panels, map functionality to them and have them show up in a pie menu in context. I'm not even using any of the fancier functions of Pie Menu Editor like macros but it's a lot of pies that will need to be replaced. I agree Machinetools is probably a complex case to use as reference but I haven't come across too many of these kinds of pie menus containing panels and entire elements taken from the standard UI.
Serpens looks very interesting to build the interfaces at minimum - thanks for the tip!
The typical advice for clearing up the greyish boundary-line artifacts that show up in the Metal-Rough workflow is to use a higher resolution texture. The results aren't great & it comes with a serious performance hit. By using a power node, those artifacts can be hidden by pushing them towards metallic. There isn't an ideal number that works for every situation. Depending on whether both metal & paint are both light or dark, or one is dark & one is light, or reverse that combination, you may have to either raise or lower the power by which the metallic mask is raised. I haven't tested this using Unreal or Unity but it works well for both Cycles & Eevee.
On a recent project using a baked alpha mask -with fine writing & scratches- texture filtering was causing me issues. I achieved a 'good enough' result using pow(64.0) to tighten up the edges so the bake results more closely matched the original procedural material. When you're having issues with masks, try a power node & sweep across a wide range of values to hone-in on something that helps.
-Setting the alpha texture filter to 'closest' works well too, just didn't think to do that before I took this screengrab. After setting it to closest, I was able to drop the power factor substantially so the scratches don't become as narrow(not shown here).
ps. I keep switching between writing 'closest' & 'nearest' filtering. Sorry about the confusion. For that type of filtering, Blender uses the word 'closest', OpenGL uses the word 'nearest'. It seems I'm easily confused.
Great news, I emailed Roaoa (the PME dev) and finally got a reply just now. It turns out they’d been locked out of their accounts for a long time, and just now solved it. Updates to their addons are coming!
When adding new textures to a material (by creating a texture node and then browsing for the source on disk), is there a way to somehow duplicate a given texture object and point it to a different source texture by simply editing the path/name - hence bypassing the need to browse for source file all over again ?
For instance here I have duplicated the metalness texture node to be a starting point for the roughness node.From there I'd like to simply link this node to the source texture for the roughness by (somehow) simply renaming "metalness" to "roughness". This would save a ton of redundant file browsing and would speed up the creation of materials tremendously, but the input filed is grayed out, hence not editable.
So far the only workaround I know to avoid having to browse for a texture source consists of visually drag-an-dropping the file from Explorer onto the shader editor. This is useful, but not quite as fast as potentially being able to simply rename the suffix inside of a texture object and it has its own set of limitations.
@Udjani : Well, that's what the "scene world" button does !
@kio : Ha, that's interesting (and it works indeed, that makes sense). However I now realize that it doesn't quite help after all, as the new texture path affects the original image object (in my case here it turns the metalness into roughness, as opposed to operating on a new texture and node). So I guess what I am *actually* looking for is a way to dictate the source path of an image object/node without having to browse for the source manually, by typing in the path directly (or pasting it). But that doesn't seem to be possible, because an image object only shows a source path field once a source has already been browsed for … I guess it's a case of an UX trying to be too clever for its own good.
Ah, well if you duplicate an image node it still references the same image datablock, you need to click the little number first before editing the path - this will create a new image datablock. Its pretty much the same concept as with instanced objects.
Ha - I am not too used to multi-user stuff in general (mostly familiar with the duplication of materials) but not so much of textures). Alright, that does work, thanks ! I'll have to see if that is efficient/fluid in the long run, but I already like how it allows to skip browsing altogether, even if that means doing a bit of unpacking/packing.
That system is really not well communicated, and also a bit an outlier comming from other DCC's. The whole thing is based on reference counting, the numbers show you how many times a datablock is referenced (so every reference is a 'user') . If reference count is zero - the thing wont be save into the blend file → things are gone after you reload the scene. If you switch the outliner to blender file it shows you all the currently stored datablocks from the file.
Im sure it wouldnt be hard to create a little custom operator which basically just duplicates the selected image node and makes it unique etc.
Yeah, I got familiar with the basics of it a little while ago when creating/sorting out sculpting brushes, but I definitely didn't have the intuition to think that it could also apply to unpacked images textures aswell. Pretty useful and powerful stuff !
Ha indeed yeah that's the other way around - but, one can use the rotate HDRI addon to at least control the rotation through a mouse gesture (personally never looked back once I started using this)
Replies
Simple Lattice seems great (just tried it), but I can't seem to figure out how to use at at the vertex/component level even though the page says that it can do so. I would expect to find the option to create a lattice at the top of the mesh rightclick/context menu, with it creating a temporary vertex group + lattice based on the current selection. Or perhaps I am just imagining too much ?
it should work like that actually - not sure whats going on right now. EugeneDudavkin has been doing updates for this for quite a while - so not sure :D
Hi there @kio, understood, I'll keep digging. Is there a specific version you'd recommend ? If possible compatible with 3.1.
just tried latest version with blender 3.1 - for me its working just fine.
I go to edit mode -> right click context menu -> create lattice (top most entry for me)
I see, I must be missing something or incorrectly installing from the github files. Once installed, does the addon have any entry in the universal (legacy spacebar) search ?
maybe check the console if it errors? Hmm maybe try to rename the folder "SimpleLattice-master.zip" to "SimpleLattice.zip" or so? I rarely use the zip way of installing addons :D
Gave the material thing a bit of polish:
https://github.com/BenjaminSauder/create_assign_material
Hello,
I have a background render, and a 450 frame image sequence, is it possible to combine the two and save out the mp4 with the animation playing in front of the background?
Hi there @kio , I got it figured out and I can now comment on what went wrong/what may need to be adjusted.
- There is no zipped release on the GItHub page, and downloading the source as zip indeed adds the usual "-master", resulting in a zip that doesn't fit with the usual Blender addon specifications. Of course this is to be expected for anyone familiar with the way Github works, but I still think that as soon as multiple files are involved a proper zip should be provided in Releases IMHO.
- Once installed there are a few issues with the way the addon is being accessed.
1 - First, as it stands the operator name for the addon (for universal search) is "create lattice", which IMHO is too vague as it sounds too much like a default function. It could just be "SimpleLattice" or "Simple Lattice", just like the add-on name. Or, the addon itself could be called "CreateLattice" - either way would make more sense than the current naming.
2 - Secondly, it being in the RMB context menu is a bit of an issue because of the way Blender RMB menus (formally "Specials" IIRC) are weirdly duct-taped. They look a lot like the regular Object and Mesh menus, but not quite. For instance here are the Object Menu and the Object Context Menu side by side. In theory they really should have exactly the same contents (populated automatically from one to the other), but in practice and if I am not mistaken they are two separate entities, which can be populated separately. And even the common functions are ordered differently :
This is obviously not your fault as this is a Blender design issue. But this explains why I originally couldn't locate SimpleLattice for Mesh Edit mode : I was looking for it in the Edit Mode Mesh menu, which I temporarily assigned a shortcut to because I assumed it was the same as the "Specials"/RMB Mesh Context menu.
Put differently : I personally have all the RMB menus disabled, because I don't see the point of them since all their functions (outside of the ones added by add-ons) can be found elsewhere. So when it came to reactivating the Mesh Edit Mode context menu by giving it a hotkey, I mistakenly assigned a hotkey to "Mesh" (VIEW3D_MT_edit_mesh) as opposed to "Mesh Context menu" (
which I don't know the command ofwhich turns out to be wm.call_menu VIEW3D_MT_edit_mesh_context_menu).All that to say that these two things (naming of the addon + the RMB menu oddity) can make the addon a tad difficult to find and use. I am glad I took the time to investigate all this as I now finally have a way to swiftly edit verts with a lattice and that is incredibly powerful :) But still I believe than some things could perhaps be ironed out.
- - - - -
Create Material is fantastic too, I've already been using it quite a bit and it is a big time saver already. Perhaps having the ability to have the color be done through an (automatically generated and packed) Image texture could be nice too, but then it would be a bit of a feature creep of sorts.
hey @pior,
ill forward the feedback to Eugene Dudavkin - i agree about the release zip, that really should be there just to download. I dont have a strong opinion about the operator name and for me the rightclick context menu is a perfectly fine place to add this. I think initially I added it also to the 3d view object menu, but that got changed at some point - guess to not litter the menus too much. Would be easy to get this back tho, or make an option for this in the addon preferences.
for create_assign_material - that auto texture generation thing sounds a bit too niche for me, but maybe you can pitch the idea/workflow a bit better?
Hello !
I suppose the naming aspect also came up on my end because of other lattice-related addons I already use, which are already populating search with other "lattice stuff " making SimpleLattice hard to guess/find :
That said this is trivial to rename on my end, so obviously not the most critical thing to adress :) And indeed I do understand that the way I fully disable all default RMB menus is probably not a widespread practice. Still, it wouldn't hurt to have the SimpleLattice entries also appear in the regular top menus for Object and Mesh respectively. As a matter of fact it looks like the Blender devs are strictly following the soft rule of everything being accessible from the top menus first and foremost, as I can't remember the last time I couldn't find something there.
Regarding texture generation : well, first I'll have to dive a bit deeper into my actual needs for this and make sure that there isn't already a way to do it :) The idea would be to have a tickbox next to the color selection + an input for texture size, causing a texture of the desired color be created and plugged straight into the newly created material. That way the user could start painting right away, which could be useful for all kind of things. But as said this is very much a feature creep. I can't thank you enough already for that handy little addon.
cool feature in blender, hopefully it's self explanatory :)
how can I make a light for cycles that can modulate how much diffuse or glossy it affects.
For exampe in eevee, its useful to turn down the the diifuse on some lights, to make more subtle lighting, but in cycles you can can only seem to turn off glossy or diffuse ie on or off
Never tried that.. but if you use nodes for the light.. then there is by default a Light Output node feed by an Emmission Shader .. so.. oh no you meant the effect on other objects.. well Cycles is meant to be an physical correct based renderer.. AFAIK only polarized light can effect the glossiness of objects.. and IDK if cycles support this.. just my two cent 😁 ... ( <== hmm latetly the emojiis doesnt seeem to work for me...)..
yuo can do it with arnold and eevee( see attached), so I a guessing it's not that hard to implement. I can't make it work properly with light path node, comes out a bit weird looking, though it does kind of work
With Cycles lights you can add nodes and do some quite creative stuff. There is a workaround so you can turn off glossiness/diffuse of a light in the properties panel, but it just on or off, not able to use a slider.
Cyles lights really need more default options like in arnold
so cycles just has on and off, where arnold has sliders for lots of different options
This seems to work
Thanks a lot Defunct. Could you explain a bit of how you get face tangent. I see it's working but still can 't comprehend how . What is general idea?
@gnoop We can use two connected face corner positions to get a tangent vector.
I'm not too sure how data transfer works but for this using nodes like "capture attribute", "store named attribute" or "evaluate on domain" set to the "Face" domain it will select the vectors from every 3rd corner and store it on each face.
you can use the "offset corner in face" node to shift the data.
Thank you very much Defunct. The pictures are really explaining
Hello all,
At this time, is there any Blender version out there that has a fast enough (for lack of a better way to put it) processing speed for the boolean brush in sculpt mode ? The current implementation (either Line or Box) is just as slow as a regular boolean calculation (of course), but the 30 seconds or so of waiting are really not sustainable in practice - especially compared to how instantaneous the same thing is in Zbrush, 3DCoat, and Nomad.
Of course the instantaneous Clip line brush is good enough-ish for a lot of cases (even though I wish it had a polygonal lasso option), but still : sometimes one really needs to remove volume rather than just clipping things in.
I wouldn't even mind automatically stepping down/remeshing the model down to a fixed, pre-established low enough polygon count if that's what it takes (since such big boolean operations are mostly useful in the blocking stage anyways)
In terms of UX I would say that ideally the operation shouldn't take more than 2 seconds at max.
Or even, something along the lines of converting the current sculpt object to a gazillion small cube objects, selecting the ones to delete using a regular object lasso selection, and then converting back to a remeshed model ... essentially working in pseudo-voxels, pretty much.
FWIW I am currently on 3.1.0.
@pior , I'm on 3.4.1 and the trim brush calc time is crazy long. Pretty sure they will whack these brushes into shape soon.
Right - I mean to be fair the speed is okay when using the box variant on a rather coarse mesh - but with the lasso one being made of multiple segments, it very quickly goes back to the issue of having too many calculations to perform. So overall it's pretty much unusable at this point ...
(So I guess it needs two things really : a polygonal lasso, and faster processing)
Well, if anyone has a clever workaround I'm still all ears :)
Hi! I have a general hinderance now and then: Adding/switching between properties for multiple objects at once. For example adding a second UV map to all selected objects or make the second UV map the active one for all selected objects.
In the past I would either do it manually, or - when facing an overwhelming amount of objects - hack together some python script (which is then super specific and gets lost). How do you deal with this cases? Curious if there a functionality I don't know about yet 🤓 I know there is "copy to selected" for property values (or alt+clicking the property), but this doesn't work in the described cases.
Regarding trim brushes: Better performance on more dense geometry would be very nice indeed. Until now I only found use for the trim brushes during blockout stage when the meshes are fairly low-poly.
Yeah, this isn't currently possible. There's been talk of making everything apply to your entire selection by default, but the topic's been stuck in option paralysis hell for years now.
Here's a free addon that does what you want: https://blenderartists.org/t/multiple-object-batch-uv-tools-plugin-free/1450400
I'm creating a procedural brick wall shader at the moment, and I'd like to know if it is possible to somehow map a color ramp to a single brick individually using the Brick Texture node. Meaning, I'd like to map a color ramp from black to white to one of the bricks, a slightly different ramp (rotated differently) to the brick next to it, and so on. I'd like to use the result as a displacement, so the individual bricks pop out each one skewed in a slightly different way, to add variation.
Is this possible using/ building upon the 'Brick Texture' node? Can I map a color ramp to each brick individually somehow? Or do I need to set up something like this manually in a different way?
There is the Brick Trick Addon..
..which do work also in 3.2... aaaannnd.. (wait a second).. 3.3.2.. there are several nodes in the shader editor available.. for example... :
with th CellId you can also pick any individual brick.. theer are also other pattern.. swirls and also hexagon..
Is there a secret to making vertex snapping in Blender work 100%? I've just struggled with it earlier today when trying to adjust a model + associated blendshape/shapekey source objects. Several objects in view, one active and in edit mode and the other ones deselected. I managed to make most vertices snap to where I wanted them but some resisted and would not snap to that one vertex where I wanted them (but would willing snap to everything else).
Any tips to make this manageable (and save my keyboard from the repeated beatings) would be welcome. :)
Also, is there possibly a way to transfer/copy/paste shapekeys and their states from one mesh to another? I have them all set up and sorted on one model and it would be just great if they could magically be made to appear just like that, same order and same ranges/slider value settings on another, identical mesh that's already skinned (I'm aware I can transfer skinweights but trying to get a process down to quickly update a bunch of shapekeys done elsewhere).
Re. shape keys : I don't know it that would be of any help, but : merging two models brings in the shape keys of one into the other, similarly to how merging models preserves skin weight data. Not sure if that is precisely what you are after, but perhaps this could be a hint at a solution ?
Overall I find the very barebones/"straight to the metal" way Blender deals with weights and blend shapes to be a life saver really. It just works :)
@pior Merging doesn't seem to help as far as I can see since the shape keys only affect the merged parts of the mesh and deleting the old part mangles the vertex order. In that case it's probably better to simply transfer skinning and whatnot and swap out the whole mesh on the armature.
Hi there - of course yup. Still worth mentionning that reordering verts is not an issue thanks to this tool :
https://barre.gumroad.com/l/transfer-verts-order?layout=profile
That said I am still having a bit of a hard time visualizing what you are after. Mind showing a practical case ?
I have an asset containing a series of shapekeys that are set up and named in a specific order, some with custom value ranges (beyond 0 and 100) and certain default values set. I use that asset to develop shapekeys as well as edit vertex normals, paint vertex colors and adjust UVs as needed.
There's an identical copy in another scene, bound to an armature, using an engine-specific material-setup that is not great to directly work with.
So, I'd like to keep working on my isolated asset and simply transfer over every property including those shapekeys without it becoming a clicking-orgy every time I push an update. The list of shapekeys is growing to a point where it becomes a bit of an effort.
Oh and it's not a single surface, rather plenty of disconnected surfaces, so I think any traditional vertex order transfer is a complete no-go. Certainly the kind where you have to select the same element on source and target for it to find its bearings.
@thomasp : considering your own technical know-how I can only assume that you've already tried all the usual approaches hence I don't have much more to suggest really. Besides perhaps having a look at the Surface Deform modifier, which has been very helpful to me when dealing with complex body and facial expression transfers, bypassing the need for a thirdparty tool like Wrap3.
The (recently added, I think ?) Smooth Corrective modifier has also been of great help to me when smoothing out some kinks caused by delicate form transfers.
Outside of that it sounds like you are reaching the limit of your scene/export setup ... so perhaps it is now a good time to refactor it deeply ? Like keeping the multiple parts you mention as different models, or finding reliable ways to fully control the behavior of materials on export (an especially tricky thing to achieve with UE, as there are no explicit rules on how their importer (re)orders material IDs).
@whizzard None of these questions can be answered outside of the specific context of the models your are working on. In some cases shoving meshes inside eachother is absolutely the best approach, and sometimes it isn't. The only thing that never works is to assume that something won't be noticed thanks to the increase processing power (for instance, assuming that it is "fine" to use different number of edges on two cylinders that are related to each other, or assuming that a highres texture will hide flaws. These approaches are always bound to backfire later down the line, whereas the assets that have been carefully crafted are always going to be the most resilient. But outside of that, anything goes.
I would say that the only way to get a feel for what works and what doesn't is by developping a strong knowledge of game art history - by studying ripped models from games that have excellent performance. Nintendo games in general, PC games that run exceptionally well compared to other similar games of the same genre, and so on. This is where you will always find the most well-crafted assets to study and learn from.
@pior: Thanks for the sound advice. I will look at other's topology, which may help me understand how it was put together. I seem to end up with some parts of a model with many polygons, and others have much less or different flow. I see consistent flow and size in professional models, and I want to find a method to get closer to that.
@pior Thanks, yeah my conclusion was that the only thing that looks like it would help me is a way to transfer the morph targets like you would transfer basically every other kind of information with the Data Transfer mod. I was considering looking into making an addon but I haven't touched Python in several years ever since the big API upgrade in 2.8 and would rather just keep on making 3D things instead of digging into that stuff again.
Still wondering about the vertex snapping issue though - that was seriously aggravating. Back in 3ds max I had a list of things to look out for when the snapping didn't work as expected (surface constraints active, wrong coordinate system picked, xforms not resetted, etc) that always solved the problem in the end. I suspect I was dealing with something similar now in Blender but am lost as to what exactly caused it to keep on failing only on certain vertices.
Perhaps an issue with a shapekey affecting snapping source or target vertex even if the shapekey is set to 0%?
I'm trying to set up a dynamic blending of materials using geometry nodes. What I have is an objects that is moving across a landscape, and I would like to blend materials based on the proximity of that object in relation to the landscape.
There is a vertex weight proximity modifier, but the resulting weight map cannot (to my knowledge) be used directly in the shader tree. I would need something like the weights this modifier generates, but in vertex color format.
Can someone give me a hint on how to set this up using geometry nodes? I'm thinking of generating a dynamic b/w vertex color map based on proximity that I can then use as a mask in the shader tree to mix multiple materials.
Pretty simple actually, for anyone who is interested:
Another topic for me: Pie menus!
Are there any resources / tools for building custom pie menus inside Blender? I googled around a bit but could not find much on how to write these manually. I'm looking to create fairly complex ones containing panels with various UI elements, not just the noodly default things.
I can see that its possible to do them when looking at e.g. Machinetools:
I'm asking this because I'm heavily reliant on the Pie Menu Editor addon, I'm literally useless in Blender without that. And it appears to be abandonware, increasingly incompatible with newer Blender builds, keeping me firmly on 3.3 for now.
Looks like the author is not coming back and I'll need to replace all my pie menus done in it which look similar to the ones shown above. 😱
I'm honestly holding out hope for someone (if not roaoa himself) will fix up PME, as it's a massive part of my workflow. Blender addons are open source by definition, so anyone can go in and make adjustments.
Another alternative is 'Serpens', a node-based interface for building your own addons that do export to Python (and are therefore less of a dead end).
Otherwise, you're right: python scripting them yourself seems like the best bet, and I have seen a lot of documentation on this over the years. Maybe it's best if you ask specific questions of things you're running into? Machinetools is a bit too big of a thing to easily decipher I imagine, but pulling apart a simple addon that provides pie menus could be a good first step.
@Michael Knubben At this point I've been holding out for years for a savior. Time to accept that the author has moved on (years ago) and I figure this addon is probably a bit too complex/clever for someone new to step in and decipher it and keep maintaining it. Best case to expect might a quick patch to fix the biggest compatibility issues. If only it became part of Blender itself.
Basically I'm looking for how to build these panels, map functionality to them and have them show up in a pie menu in context. I'm not even using any of the fancier functions of Pie Menu Editor like macros but it's a lot of pies that will need to be replaced. I agree Machinetools is probably a complex case to use as reference but I haven't come across too many of these kinds of pie menus containing panels and entire elements taken from the standard UI.
Serpens looks very interesting to build the interfaces at minimum - thanks for the tip!
The typical advice for clearing up the greyish boundary-line artifacts that show up in the Metal-Rough workflow is to use a higher resolution texture. The results aren't great & it comes with a serious performance hit. By using a power node, those artifacts can be hidden by pushing them towards metallic. There isn't an ideal number that works for every situation. Depending on whether both metal & paint are both light or dark, or one is dark & one is light, or reverse that combination, you may have to either raise or lower the power by which the metallic mask is raised. I haven't tested this using Unreal or Unity but it works well for both Cycles & Eevee.
On a recent project using a baked alpha mask -with fine writing & scratches- texture filtering was causing me issues. I achieved a 'good enough' result using pow(64.0) to tighten up the edges so the bake results more closely matched the original procedural material. When you're having issues with masks, try a power node & sweep across a wide range of values to hone-in on something that helps.
-Setting the alpha texture filter to 'closest' works well too, just didn't think to do that before I took this screengrab. After setting it to closest, I was able to drop the power factor substantially so the scratches don't become as narrow(not shown here).
ps. I keep switching between writing 'closest' & 'nearest' filtering. Sorry about the confusion. For that type of filtering, Blender uses the word 'closest', OpenGL uses the word 'nearest'. It seems I'm easily confused.
Great news, I emailed Roaoa (the PME dev) and finally got a reply just now. It turns out they’d been locked out of their accounts for a long time, and just now solved it. Updates to their addons are coming!
Hello all,
When adding new textures to a material (by creating a texture node and then browsing for the source on disk), is there a way to somehow duplicate a given texture object and point it to a different source texture by simply editing the path/name - hence bypassing the need to browse for source file all over again ?
For instance here I have duplicated the metalness texture node to be a starting point for the roughness node.From there I'd like to simply link this node to the source texture for the roughness by (somehow) simply renaming "metalness" to "roughness". This would save a ton of redundant file browsing and would speed up the creation of materials tremendously, but the input filed is grayed out, hence not editable.
So far the only workaround I know to avoid having to browse for a texture source consists of visually drag-an-dropping the file from Explorer onto the shader editor. This is useful, but not quite as fast as potentially being able to simply rename the suffix inside of a texture object and it has its own set of limitations.
Thougths ?
Can you link the viewport shading with HDRI preview with one that can be used with rendering so it's easier to control it?
Hey I think the issue is the packed texture, for regular image nodes the path editing works just fine.
@Udjani : Well, that's what the "scene world" button does !
@kio : Ha, that's interesting (and it works indeed, that makes sense). However I now realize that it doesn't quite help after all, as the new texture path affects the original image object (in my case here it turns the metalness into roughness, as opposed to operating on a new texture and node). So I guess what I am *actually* looking for is a way to dictate the source path of an image object/node without having to browse for the source manually, by typing in the path directly (or pasting it). But that doesn't seem to be possible, because an image object only shows a source path field once a source has already been browsed for … I guess it's a case of an UX trying to be too clever for its own good.
Ah, well if you duplicate an image node it still references the same image datablock, you need to click the little number first before editing the path - this will create a new image datablock. Its pretty much the same concept as with instanced objects.
Ha - I am not too used to multi-user stuff in general (mostly familiar with the duplication of materials) but not so much of textures). Alright, that does work, thanks ! I'll have to see if that is efficient/fluid in the long run, but I already like how it allows to skip browsing altogether, even if that means doing a bit of unpacking/packing.
That system is really not well communicated, and also a bit an outlier comming from other DCC's. The whole thing is based on reference counting, the numbers show you how many times a datablock is referenced (so every reference is a 'user') . If reference count is zero - the thing wont be save into the blend file → things are gone after you reload the scene. If you switch the outliner to blender file it shows you all the currently stored datablocks from the file.
Im sure it wouldnt be hard to create a little custom operator which basically just duplicates the selected image node and makes it unique etc.
Yeah, I got familiar with the basics of it a little while ago when creating/sorting out sculpting brushes, but I definitely didn't have the intuition to think that it could also apply to unpacked images textures aswell. Pretty useful and powerful stuff !
@pior But then you can't control the hdri rotation and strenght with the little panel
Ha indeed yeah that's the other way around - but, one can use the rotate HDRI addon to at least control the rotation through a mouse gesture (personally never looked back once I started using this)
https://blenderartists.org/t/rotate-an-hdri-for-blender-2-8-2-9x/1178283