ajz3d

Yup, if identical UV islands are stacked on top of each other I think their islands get merged, because I cannot select one of them (with Mark Seams set to islands) without selecting the other (in the UV Preview window). Lucky you.

I guess the only way is to manually select faces of one of the stacked UV islands (with Select tool) and then move it to a different UV set. It's problematic because... yeah, it's manual.

I stacked two UV islands on top of each other with CTRL+C/CTRL+V combo. Now I want to move one of them to a different UV set. How can I separate those islands now, without moving them and without changing the rest of the UVs from their set? Please tell me this is possible without having to unwrap the UV set.

Thanks Gary. I reduced retopo mesh to about 56k tris. Anything less would probably lead to surface loosing its smoothness. I had to apply thickness to some of the objects, like battery compartment doors or undercarriage metal covers to name a just a few, so it effectively doubled their polycount. There are also many elements inside the battery compartment. It is true that light bulbs have it's significant share in polycount (8k tris in total). this is because of very irregular shape of their stems. I left it in the mesh because I might want to render the bulbs with with refraction and physical lighting in Mantra sometime in the future. I can always remove the stems anytime, as they use up very little UV space. Anyway, the starship is a "hero" model that needs to keep its relatively smooth curvature on close ups, so I can't really reduce its polycount further without triangle hunting. PS. Regarding distribution of light bulbs, I had to cheat my way through in Houdini because the process in 3DC was too convoluted and buggy. I would have to spend hours on something I did in five minutes. http://3dcoat.com/forum/index.php?showtopic=19177 PPS. Sorry for any mistakes. I was typing this on a phone.

Good idea. I will drop him an e-mail.

Retopo mesh with a 2k normal map. 63677 triangles:

Thanks Carlos, but this is not neccessary. I will make it on time.

Finally, yesterday I found time to finish retopo. I'm also almost done with peparation of bake cages, so I will be posting a baked result shortly. I will be baking normal map to a 8k texture, but target resolution will probably be 4k, with 8k being used for painting only.

Any news about sculpting layers or multi-monitor support?

Kingdom for additional e-panel modes for Extrude Clay tool!

Esgaroth archers better sharpen their arrows.

Shooting in the dark here, but have you perhaps tried importing your models without voxelization? If you're importing to voxels you need to be sure that everything is watertight. The mesh needs to have some thickness, or you may run into some problems when 3D-Coat takes the process of making it thick into its own hands. I made it my habit to import everything without voxelization. This gives me 100% assurance that my mesh will import WYSIWYG. And only after my objects are imported, I convert them to voxels if it is necessary. This gives me total control over what happens with my mesh. Can you provide source geometry files which, when imported, will randomly generate this bug?

Yes, but shell tools in various applications usually utilize vertex normals as guides for displacement direction for new vertices. So intersections in tight corners will always happen if the extrusion value the user provides crosses a certain threshold. But as you can see from the last screenshot I provided, there's a distortion going on convex surfaces going on (see: the feet for example)! Everything that you said in your last post is very true, but it applies only to scenarios when Conform retopo mesh is used with tools like Pose or Move. Tools that perform an actual deformation of the surface. This is when retopo mesh snapping is required. Transform tool is a different beast IMHO. You deal with transformations of whole objects. There are no deformations going on, hence vertex snapping becomes unnecessary (at least in most use cases). To better explain what I mean, I will provide an example of my solution in Houdini. I can move my lowpoly objects anywhere I want and highpoly will conform (I chose conforming of highpoly meshes because of performance reasons in the viewport). This solution works purely on transformations of primitive groups (you can call them objects). No vertex snapping is necessary. To sum up, I think that Conform retopo mesh when used with Transform tool should be treated differently than if it is used with Pose, Move or any other tools. At least it should provide a flag, which if toggled, would force 3D-Coat to use transforms for conforming, and if unchecked - vertex snapping.

You can assign a hotkey to Rotate panorama event in Camera->Customize Navigation window.

It's a known glitch. Complex checker is horizontally (and probably also vertically) flipped in the viewport (http://3dcoat.com/mantis/view.php?id=2048). As a workaround, open the checker2.tga file from C:\Program Files\3D-Coat-V4.5\textures (or whatever your installation directory is), flip it and save a copy to %USERPROFILE%\Documents\3D-CoatV45\textures\ under the same name.

The problem I encountered can be reproduced with any pair of objects (with the default human figure and its decimated counterpart, for example). It seems that scaling the object down breaks corresponding retopo groups. It looks almost like their vertices were trying to snap to some invisible reference (and there's none present in the scene). Scaling up yields less severe, but also noticeable problems. I will report it as a bug. Meanwhile, I will position my objects in Houdini.

Thanks for the advice. I just tried Conform Retopo Mesh and right after scaling down the light bulb VoxTree hierarchy (with Transform tool), I got this: This, pretty much, is what I suspected that will happen. Several steps of Undo led me to this state: Scene reload was necessary.

Great job with the new update! Thanks for fixing "Apply Symmetry".

I found the culprit. Turns out the problem was caused by contents of /factures directory. Removal of this directory from 3D-CoatsV45 preferences folder fixes the problem. http://3dcoat.com/mantis/view.php?id=2104

To change normal direction of a selection you need to take the model to retopo room. Inside it, when you choose the "Select" tool, you will have a new set of commands available with flip normals being amongst them. However, I imagine that selecting specific faces of a mesh so dense can be difficult, so what I suggest is to import your object to sculpt room and then quickly retopologize it and create new UVs. Finally, bake the normal map and begin texturing.

I inspected your file and I think that the problem is that surface normal vectors are facing different directions. Only the normals of sideframe2, glass1 and latch2 seem to be facing the correct way. Fix your normals and you should have no more problems with texturing. --- This model could easily be reduced from ~300,000 faces to something with just just three digits. Pozdrówka.

Transform tool only allows for inputting translation vector. I need to know the rotation and scale of the instances.

Carlos, how do I find out those precise specific "positions"? Remember, it's not only XYZ position, but also rotation in local space and scale (some bulbs are larger than others)! And the order that those transformations are calculated in.

Rooattack, try this experimental software out: http://3dcoat.com/forum/index.php?showtopic=18576&hl=%2Binstant+%2Bmeshes AUTOPO is a nice tool to use, and it saved my bacon more than once, but it is also quite slow. If you have thick and watertight meshes, like in that FBX, it should work okay and you'll get the result pretty fast. Of course, you will have to tweak and reduce your retopo mesh afterwards a bit (in retopo room).

Strange. Then maybe I'm facing ye olde' options.xml problem, or something related to it. I'll have to check.