gbball

Has anyone tested the curves? I've played with it a little bit, but I don't know how it's intended to be used. Andrew, is it possible to make patches out of curves? Kind of like pseudo polygons? I think it would provide a powerful framework for both sculpting and retopology. The results I had were interesting. But I think I encountered some bugs in 14C. First bug -> I created a curve and I tried filling the curve with voxels and nothing happened. Second bug -> Then, I tried making a curve and aligning a curve profiles along it, which worked, but it started creating a ton of new mirrored layers everytime I moved it to adjusted it. So on one side of the symmetry plane, I had the curve with the curve profile on it, and on the other side of the symmetry plane, it stamped in a mirrored version of the curve profile on a new layer. After adjusting my curve profile for a little while, I had hundreds of layers showing a mirrored version of all the tweaks and changes I made.

Quad based because with subdivision levels ability to reuse productions assets (save time) You can increase and decrease resolution non-destructively as opposed to resampling it fits better into a pipeline better back and forth with retopo room I think it works in support of sculpt layers. Correct me if I'm wrong, but in a subdivision workflow don't higher subdivision sculpt states work off stored displacement data? And isn't image based sculpting the main component of sculpt layers? I think you kill two birds with one stone here. Last summer I had to make 5 characters for a short film and I sculpted the first character in 3D coat, then I carefully retopologized it, hoping to be able to reuse the retopo mesh for the other characters so that rigging would have an easier time transferring rig and weights from one character to another. But the task became very cumbersome, very quickly. I tried the tools in 3d coat, such as conform to retopo, but that only works with move, transform and pose - plus it's slow...having to show the sculpt with retopo data isn't performant. Those tools alone aren't enough to recreate a new character. How much easier it would have been if I could directly manipulate the mesh with the sculpt tools in 3D coat. The job would have been a lot easier in zbrush, which I don't own. I tried blender, but the sculpting tools and workflow are a bit limited. So what I ended up doing was making all my sculpts in 3D Coat, then between 3D Coat and Blender, snapping my original retopo mesh onto the sculpt. It was a big headache. Unique sculpts currently require either manual retopology or autotopo. In 3D coat, the only thing that a quad based mesh is useful for is for unwrapping or texture painting. If I want to sculpt on it, I'll lose any ability to bring the mesh back into another 3D application with the same topology/UV data/etc. The workflow currently is far too linear. Sculpt-->Retopo--->Texture Quad based sculpting could allow for sculpt<-->retopo-->texture<-->sculpt What I propose would make the program a lot more production friendly. Saving studios and artists a lot of time and money. It would really turbocharge production for big studios and indies alike.

Great addition Andrew. I will have to try it out. I hope this means that quad based sculpting is on it's way. We need 3 layer types in the sculpt room. Voxel, Surfacemode, and Quadbased with subdivision levels.

There are workarounds... You can do a retopo, then merge the retopo mesh into the sculpt room using the import tool. Prior to stamping the mesh into the sculpt room, you can subdivide a few times which will smooth things out. You might need to remove some of your voxel sculpt and bring in the retopo sculpt in on a different layer to keep your forms clean.

Here is a timelapse video of me creating this guy.

Voxel sculpting in VR would be incredible.

Agreed. Let me say that I love 3D Coat and it and Blender are at the core of my studios pipeline. I also teach the programs at college and university. It's the program that really made 3D accessible to me and I'm going to stick around for as long as it's being developed. I use it for everything and I learn new things all time. There is so much in the app that is amazing. Sadly us users will always be looking for more and more, but hopefully the development team is challenged by this and pushed to do their best. I personally should do a better job of reporting bugs and suggesting interesting new features. If I had a better idea of the roadmap, I would probably chime in a bit more because there are a few things that need addressing that have been mentioned before that would really boost productivity. I'm hoping they're in the works, but not sure. There are 2 really big things that have been mentioned before. -Displacement based Quad sculpting with non destructive subdivision levels This is critical, because in production, 3D Coat sculpting forces you to redo too much from scratch if you want to reuse things like hands/ears as a basis for sculpting. Also, for delicate areas like the eyelids, nose and lips, it gets hard to sculpt on one small area without breaking another nearby area freezing is not effective on a triangulated mesh because the edge border gets jaggy unless using very high resolution. Right now after doing a sculpt I feel forced to do a retopo rather than being able to reuse quad based meshes made in 3D Coat or elsewhere if I need to do more detailing or substantial changes. How do I get a uniform topology sculpt mesh without changing the quality of my mesh? I get faceting when moving low to high, or I lose crispness if going from high to low. I've found ways to work around these issues to get the results that I want, but I can see how much easier things could be and how much more production friendly it would be. -The equivalent of a topological move tool Not having this makes it difficult to work in delicate areas such as lips and it also cause big problems near the mirror when working on inner legs. Currently, I will have to separate things onto separate layers, so the top lip on one layer and the bottom lip another layer. Or for things like legs, I will delete one leg, make a mirrored instance, turn off mirror on the original layer and the mirrored layer to get the results that I want.

Not sure if it's the new version, but I noticed that the snapping is sometimes the issue. I also get pretty good results when I turn capture details pretty low...to like 10%. I also voxelize things first. I like the topology that I get for the most part, but I usually have to do a bit of cleanup afterwards. It would be great if it was really stable.

This past Thursday I made this creature for the game. Sculpt, retopo and texture in 3D Coat. Rig and animation done in Blender. I recorded most of the process except for cleaning up the rig and animating. Here is a link to the sketchfab scene - https://skfb.ly/6vBn8

How does it compare to other rigging/animation tools from your experience?

More progress https://skfb.ly/6vsHy

Hey everyone, It's time to resurrect this thread from the dead. I've begun doing weekly development session where I will work on an asset from the game. Currently the project is being made in Unreal and somethings have changed and now I am wearing several hats including developer. Which is cool, because I want to get better at coding and using Blueprints. Anyways, for anyone interested you can follow some of the progress at least in terms of asset creation on our Facebook page here-->https://www.facebook.com/pg/fullcirclecs/videos and the first two videos below. I might eventually switch to doing a twitch stream instead. Software packages I will be using during these video sessions are 3D Coat, Blender and Photoshop. I will possibly also be doing a little Substance Designer, Unreal and TVPaint.

Andrew, Can we get a repeat last stroke function? That would work great with some of the e-panel marquee draw modes.

What is 3DCPrinting? I tried downloading it but it doesn't work.

Thanks Digman, hopefully it gets fixed, but I appreciate the workaround!

I tried it again and I see what's happening. I closed my voxtree window, and set it to pop up with keypress (alt+v in my case). It doesn't work in that situation. Only if the voxtree window is docked or permanently opened does it work properly. Can someone confirm? I like keeping my screen as clean as possible recently and I've set most of my window to pop up with a keypress so I don't know if it was an issue in earlier versions.

Strange...still no luck.

I think I encountered a bug. Can someone confirm. I can't drop a voxel layer to become a child of another layer. Normally I could drag it from the layer's move icon, and drag it to another layers '+' icon and it would become a child of that layer...but haven't gotten it to work in 4.7.35

Awesome work!

This is great! Excellent design!

What I mean is for retopology. A tool where the program will be able to analyze a black and white image or a high contrast colour image and create topology projected onto the sculpt similar to the strokes tool, only using an image rather than drawing each stroke...

Just like the title of the thread, it'd be nice to have the ability to project topology on the mesh from an image/brush alpha. Like a brush stamp...with the ability to rotate, scale, change position before applying.

Glad to hear it worked out. Scaling up the model should have the same effect as reducing the objects density, or adjusting the settings in the scale/bumpiness in the shader since it seems like the shader info is tied to the object density per layer. I notice it because I change density on my sculpts a lot and it causes problems when I bake with certain shaders.

This is great stuff man!