DMG

I'm enjoying getting familiar with the new Multires tool. I did initially think it was bugged, because I was getting weird collapsing polys... But I figured out it was because I was smoothing with the SHIFT option set to 'Powerful Smoothing', and this option has decimation included. The 3DCoat new feature video: ...says that (around 1m 20s) that when multiresolution is in operation, dynamic subdivision is disabled. Perhaps a the 'Powerful Smoothing' should also be disabled, or maybe automatically switched to normal Smoothing while multiresolution is in use?

Hi MJonathan, I don't have the test builds, so cannot test this, but it looks like the first cube in your video is made up of quads, while the second cube in 3DCoat is composed of triangles. This will certainly account for some difference between the two results. Would be worth testing with straight quad cubes in both for a better comparison. Derek

I know Andrew and the team will be busy working away, but I'm really excited to try the new multires modifier when it reaches a 'stable' release.

I've just discovered Jama's works; what a very talented artist. I thought I'd investigate who created the gorgeous sculpt of the bearded man used in promoting 3DCoat, and came across Jama's very varied portfolio. Like others, I am very curious as to the brushes, tools and textures he used to create that sculpt, and others like it. It is incredibly realistic. I have found the 'Jama clay' shaders, but that's about it! Derek

That's a nice simple video. I learned the same years ago from reading articles by Bay Raitt, who was the first I'd seen to teach use of edge loops. He modelled Gollum in Peter Jackson's LotR movies, and if I'm not mistaken, he used to frequent this forum some years back. Fragments of his tutorials are here: http://www.theminters.com/misc/articles/derived-surfaces/index.htm

Updated Chart... The rows are: 3DCoat's GGX, Handpaint, Render; then Blender's Eevee, Cycles. * I think the Handpaint render (compared to the GGX one) has slightly lower intensity in the bright highlights, as well as ever so slightly more diffuse reflections when the roughness reached mid-levels. And compared to the 3DCoat render, the bright highlights in the Handpaint render are a touch more intense, and as with the GGX ones, much more defined. Looking at the Handpaint renders and the Blender Cycles renders, they're actually very similar. This is excellent. I feel that if I use the Handpaint preview, I will be seeing a very close approximation of what I'll get in Blender. So my chart is now redundant, based on the reason I created it in the first place, but it has nevertheless been very useful. Derek

I had a quick check with the 'Handpaint' Paint room render, and it at a glance, the light hotspots aren't quite as intense as the GGX ones are, so I suspect that will be closer visually to the Bender renders. Definitely worth updating my chart to include that.

Hi Carlosan, I have that shader selected, the GGX Burley one. If I have time, I might expand and include the 'Handpaint' one too, but it won't be 'till the weekend at least.

A texturing request... I understand the workflow surrounding clipmasks for layers. I get that they work on transparency rather than greyscale values. However, if we have a layer in greyscale that we wish to use as a mask, as I understand it, we have to export that layer to an external paint program and convert it into a image with transparency based on either the black values or the white values, then re-import it. It would be a great time-saver if there was an option on the right-click menu for layers, which could convert a greyscale layer to one with transparency. The standard appears to be black=fully transparent; white= fully opaque; and the grey values have varying levels of transparency in between. I'm not a programmer, I don't imagine it would be too hard to script a function which looks at a layer's luminance, and apply that to the layer's transparency. If this feature already exists and I just haven't found it yet, then I apologise... Derek

I was going between 3DC and Blender a bit trying to hit the sweet spot I wanted for something, and I noticed differences between the two renders, especially caused by varying roughness values. I decided to get a bit scientific about it, so here's the results of my tinkering. I used the studio.exr panorama from Blender in both programs for a fair comparison. The 3DCoat environment settings were left at default, but the light was deactivated. The HDRI was normalised when I imported it. The HDRI strength was left at default value in Blender too. A simple sphere, identical colour value in all renders. I'm making no judgement as to which renderer is 'better'. This is just purely for comparison, to see how roughness values appear in each in order to make texturing a bit more predictable. I did notice that in the HDRI there is a tiny but bright spotlight high up adjacent to the other light, but for some reason this isn't picked up in the 3DCoat renders. Perhaps it's just a tiny pinprick and too small to be picked up. I don't know. There also appears to be a light down low, right of centre, but only on the renders for 25% to 40% roughness in the 3DCoat renders. I don't know where that's coming from. It isn't apparent in any of the others. The most noticeable differences is that the reflections in the 3DCoat Paint Room preview are a touch more intense than both of Blender's renders, but the same reflections are a bit feinter in 3DCoat's Renderer. Also, at around 35-40% roughness, the 3DCoat Renderer has greater blur on the main reflections near the centre, which has almost become a single bright patch, whereas all the others still have two distinct light sources. Like I said, this isn't about picking a 'winner' or 'favourite'. It's just to help me predict how my textured objects will appear and render in Blender to save time to-ing and fro-ing making adjustments. Just thought I'd share. Regards, Derek

I think I'm trying to do something similar... I have a wood texture, but I want to desaturate the lighter parts of it. At first I desaturated it in 3DCoat, and adjusted the contrast and brightness to get a good 'mask' layer. However nothing I tried would apply it as a mask. I eventually remembered that clipping masks in 3DCoat work off the transparency layer, NOT greyscale images. I exported the colour layer, opened it in Krita, used the 'Colour to Alpha' tool to convert the light areas to transparent, and leave the darker areas opaque; then re-imported this into 3DCoat to use as my clipping mask. I'm not expecting 3DCoat to switch to using greyscale images as clipping masks, that appears to be a large change which has been requested before. However, I think a simple script on the layers right-click menu which basically turns greyscale images into layers with varying levels of transparency (The standard seems to be white to fully opaque, black to fully transparent, and varying levels/gradient inbetween) would be very helpful, and would mean users no longer need to jump outside of 3DCoat on such occasions. Cheers Derek

Just updated, great to see the 3D primitives scaling issue has been corrected, Thank you!

Was just having a quick look inside PixPlant, and it has the ability to clone simultaneously over multiple maps. I think this solves my last issue with simultaneous cloning and keeping displacement maps 16bit. I'm not saying there isn't a way to do this inside 3DCoat (and I may very well continue to investigate layers, alpha channels, etc) but PixPlant is an app designed for this very purpose, so it is the logical tool for that part of the job. Workflow complete. Just need to fix up these test maps. I'm still open to suggestions regarding the UV issue I highlighted, and cloning 16bit displacement maps at the same time as the diffuse map. Thanks especially to Carlosan for all his assistance and for pointing me in the direction of PixPlant.

Yes, I have changed those settings. (Showed that in my second post) A later post shows that the UV points are close, but not exact... (The inaccuracy varies along the edges. I think it the whole UV island might be rotated by a fractional amount) ..but even though that amount probably falls under the definition of 'close enough', it doesn't help me if the mesh I'm unwrapping isn't square, because I'm then back to manual guesswork to scale and position it. It's not a workflow killer though. Depending on the situation, I can either live with it being 'close enough', or export the retopo mesh to Blender and snap the UV points there. Would be nice if there was a 'rotate-and-stretch-to-fit-and-snap-to-the-edges-of-the-UV-area button inside 3DCoat though. Cheers.

Hi Carlosan, yes, I've been using the Texture: Import functions. Here's a bit of a breakdown of some workflow points. It might better explain what I'm doing, and/or where I'm going wrong, and an issue I've come across with my displacement maps. Bit of an epic post, sorry. UVs: The UV issue is more to do with creating the UVs for my baking mesh. Here's an example (it's not from a photogrammetry object, but that doesn't matter for the purpose of this illustration). The surface I want to capture and turn into a tiling texture could be an irregular or curved surface such as a tree trunk, etc. I'd import my photogrammetry object, and create a retopo object to bake the displacement and colour to. Thing is, I want my UV to completely fill the UV area. I can use various tools to straighten the grid up, but I can't see any way to make the UVs expand to fill the area so that: the vertices on the left snap to a U value of 0; the vertices on the right snap to a U value of 1; the vertices on the bottom snap to a V value of 0; and the vertices on the top snap to a U value of 1. I can eyeball it, which is okay, but being able to expand it to fit would be superior and more accurate. I've also not found a method of using the gizmo in the UV window to restrict the movement of the UV island to left/right or up/down. There's handles to restrict scaling to those directions, so I'm a bit surprised I can't see similar handles for translation. None of this is not an issue when setting up to paint the textures to tile them, as the Paint UV Mapped Mesh: Tile options import a plane with UVs right to the edge. (Thanks for highlighting that option) Normal Maps No issues here at all. I just wanted to highlight the maps I create and use. The normal map generated by 3DCoat is perfect. However, there's a lot of overlap between it and the displacement map. The normal map can easily be generated from the displacement map later on. If I'm painting maps to tile them, I'm not going to bother painting three maps when I can do two and derive the third in a single click later on. The high-detail normal map I generate later creates tiny surface details from the diffuse map which is not captured in the scan/photogrammetry model. This is best created after the diffuse has been tiled, and it's not functionality that 3DCoat has or needs. I use Shadermap and now PixPlant to create these maps. Depending on the situation, I could use either of these normal maps, or a mix of both of them. Here's tiny swatches from the 8K maps: Diffuse (baked in 3DC): Displacement (baked in 3DC): Normal map (Baked in 3DC): Normal Map (derived externally from the diffuse): Displacement Maps: I have realised that my displacement maps need to remain 16bit. If they are reduced to 8bit textures, then 'stair stepping' artifacts become visible when rendering. I had tested copying the displacement into the roughness channel so I could paint the diffuse and displacement together, but the roughness channel seems to only support 8bit. A bit of research late last night suggests that alpha channels in 3DCoat do support 16bit (most of the posts on this forum about them seem to focus on using them to create brushes), but right now I have little experience of working with those, and I am not sure if they could even be painted simultaneously with the diffuse when I try to tile them. I fear this may ruin my plans... More experimentation needed, and suggestions welcome... Cheers, Derek