AbnRanger

Looks good, but yeah...if you could retopo everything, and texture it with a little wear/aging, it would be even better. That would give you a chance to try out the new boost in Painting speed (3.5.1)

I don't know, but object loading is horrendous today....10 mins and still counting just merging a voxel model (RAM is not an issue...it's not multi=threaded either...just one core working only about 25%).

I think what you want can be done without textures being brought into Voxels...you import your mesh into the Retopo room and it will snap to the one your merged in the Voxel Room. From there, you can re-snap the mesh whenever you make some changes in the Voxel Room. However, you will want to keep the changes relatively subtle, as major changes such as re-posing, would make the job to resnap nearly impossible without some manual relocation of the displaced geometry (to fit the changes in the Voxel model). From there you can export to PPP or MVP or use the texture baking to to just export displacement and/or Normal maps to reflect the changes.

I tried to find it, and it seems the link doesn't work any longer. Can anyone who did happen to get the download at some point, please post the set in the 3DC Exchange Library section of this forum?

I tested the Painting some and it seems wicked fast. Even some speed up in Voxels as well. http://www.youtube.com/watch?v=MfLnQMeUoWU

Give this a try. What you want to do is right click on the shader and click "Edit permanent shader settings" or "Construct New shader." You will have a popup to choose textures and have the ability to adjust the scale of the texture and amount of bump, etc. You will want to find an image file for the wood that you want. If you have none or want more textures, there is a host of textures available to download entire categories with one click, in the 3DC Exchange Library section: http://www.3d-coat.com/forum/index.php?showtopic=3289

The new Auto-Retopo feature does not yet work with directly with imported meshes. Andrew tried hard recently to make that happen, but had some major issues. He said he had to put it on the backburner for a bit. They are currently working on Multi-threading (with rather expensive new Intel Libraries) for the Paint Room...some of that will improve brush speed/response throughout the application.So, the answer is to merge straight into the Voxel Room, where it is converted to Voxel Clay. Then you can right click the layer in the Vox Tree Panel and select Auto-Retopo (a few different choices there). You really don't lose anything by merging into Voxels...as long as you make sure to increase the resolution enough to capture the detail during the merge. If the model comes in and only occupies a tiny portion of the grid, you will want to scale it up a bit...25-50% of the grid should be plenty, without going too far over. You will get a prompt asking if you want 3DC to preserve the original scale when you export. If you have some openings in the model, make sure to select "Make Mesh Closed" so that you won't have any artifacts upon merging. If it still appears that it didn't catch all the detail, then you can either click the Increase resolution icon at the bottom of the Vox Tree panel, or scale the model up a bit on your 2nd try.

I agree, Bezier Splines (with tangency handles) is what most people are familiar with. Although, I think the two methods now work as well if not better...especially the B-Splines. I got used to those when they were introduced into Combustion (an Autodesk Compositing app). They make Rotoscoping work so much quicker than using Bezier Splines.

I'm talking about skipping the Retopo stage and just exporting the hi-res (triangulated) mesh to 3ds Max, and there using PowerNurbs as a Retopo tool as it were. In Max, you can freeze the high-res mesh, and have CV's or Vertices Snap to frozen object. It would be a similar to having NURBS tools to Retopo with in 3DC. You have NURBS in Blender as well, so you could do the same thing. The only objects I might not use it on, would be characters or animals. Again, one of the main benefits is not only fewer loops/curves, but no concern for the snapping problems that occur sometimes in the Retopo Room. I think it would be nearly as fast to do as auto-topologizing the model.

That model is not that high in degree of difficulty retopologizing. You could manually retopo it fairly quickly using the strokes tool on all the cylindrical and rectangular shapes. Auto Retopo will work fairly well, but I would try to use the split tool in the Voxel Room to separate the major shapes into their own layer and auto-topo each one separately. Lay down guides along areas where you want to maintain hard edges or you want the edgeflow to follow a certain contour. Then the next part of the wizard is to paint the areas that you might want higher concentrations of polygonal detail. I find that auto topo works especially well for primitive shaped objects. It's the complex shapes that can throw it off a bit.I would also try a similar approach if you decide to manually retopo the model. Just separate all the major shapes into their own layer and retopo them separately instead of trying to make it all work in one retopo layer. That way...yes, you have more layers and more objects to handle, but they are more easily done individually.

I think I'm going to start direct exporting hard surface voxel objects and trace over them with PowerNurbs Pro in 3ds Max (Integrityware is the company that develops it). It seems like a good blend between the speed of Auto Retopo (drawing guides) and the control and precision of manual retopology. Instead of drawing guides, it's curves. The best thing, in my opinion, is the ability to avoid the drawbacks to AutoRetopo and the Rats-nesting you sometimes get in the Retopo room, after relaxing or subdividing the mesh.

Hey Rich, welcome aboard...yeah the Lightwave community helped get 3D Coat rolling in the early days, as Larry Shultz (Splinegod) brought it to everyone's attention when almost nobody had ever heard of it. It has come a loooooooooooooooooooong way since then. Andrew did a lot of Multi-threading work in the Voxel Room this summer and it has taken sculpting in 3DC to the doorstep of ZBrush and Mudbox, in many respects. However, to get the most out of Voxels, you'll want to get as much RAM as you can afford, as it is accounting for volumetric data, not just points/vertices in space. Volume mode is best used to construct your model to a medium res level and then Surface mode is where you want to do your high level detailing. It ignores volume data and only displaces the surface...emulating polygonal sculpting.With more than 8GB's, you should be able to get ZB and MB levels of detail, or you can just get to a medium-high level and then do the uber-high detail with image-based sculpting (Micro-Vertex= live Displacment painting, Per Pixel & Ptex= live Normal Map painting). And to answer your question, Andrew implemented Multi-Resolution this summer as well...but it works a bit differently than in ZB or MB. Here's a better explanation: http://www.youtube.com/user/Pilgway#p/u/4/bMyLRVlHeUU http://www.youtube.com/user/javisjones#p/u/0/sasaZRqKVlo

If the roots are indeed a key part of the story and need to be highly detailed, you're going to HAVE to make some sort of compromise, I'm afraid...as displacement mapping and normal mapping only get you so far. You may try to keep the large roots and medium-sized roots low-poly, but thin roots that jut out on their own will either need to be higher poly objects or utilize a hair/fur utility to take care of those. That is how I would try to handle it. You probably have some of your own ideas how to get around high poly counts, but I'm just throwing some ideas in the air to see if any of them might help. Another option you may want to explore is using splines (you can make them be renderable), maybe see what options you have to taper the look, and a Noise modifier. Or you could also paint a few different thin roots in Photoshop and make cards from them, so you can instance those throughout the scene. for splines, you should be able to select the instances and adjust their seed in order to randomize the noise effect, independently. You can use a particle system to scatter them among the larger root models, by making them be the source emitter. What I would do from there is paint an area to distribute them where you want on the model. The following are some examples/options in 3ds Max, but you could use similar tools in other apps., such as Maya's PaintFX. http://download.autodesk.com/us/3dsmax/2011features/modeling_texturing/3dsmax2011_nfv_model_textur_objectpaint_617x407.html http://area.autodesk.com/tutorials/particle_leaves_fast_tree http://area.autodesk.com/tutorials/painting_particles_in_a_scene http://area.autodesk.com/tutorials/pflow_for_creating_a_grass_land http://area.autodesk.com/tutorials/paint_hair_growth_areas_with_vertex_paint Another feasible, low-rent option would be to paint them in at the post stage....depending on the shot. Combustion, Toxic (now Autodesk Composite), Fusion, and Softimage's built in compositer have good paint tools.

Or the Split tool?

You know, I was just thinking today that it might not be a bad idea after all to see if Andrew and the guys at MOI can collaborate on a plugin, or if Andrew can contract the developer to implement some NURBS tools within the Retopo Room. There are a few good reasons for this, I think. 1) Fewer points (CV's) to snap to the Voxel object and thus greatly reducing the rast-nesting that often occurs when relaxing or subdividing. 2) The object could then be exported in Nurb's or solid surface formats to a host of additional applications...as well as Maya, Max, Blender, etc. A lot of my current work is modeling displays and the client often would like to have the output in a Solid Surface format (that would go straight to production). Going straight from Rapid Prototype to Production. That would open up some major, major markets for Andrew...markets that neither ZB nor MB touch. 3) The object could be Retopo'ed in NURB's and exported/converted to an OBJ if needed. I'd be willing to pay extra for such a plugin or a different variant called "3D Coat XTended". I really think it could help abate some of the snapping issues in the Retopo room and make the whole job easier in many cases. It might even prove to be a good alternative to the Auto-Retopo tool...as you are doing much of the same action...instead of laying down guides, you'd be laying down a spline cage and come out with much cleaner results. In fact, I'd venture to suggest that in a NURB's mode, Andrew could use many of the very same tools, such as the strokes tool, to lay down CV Curves (closed loops) with cross-sections, in the same manner as is done with Polygons currently. He already has B-Splines in the toolset, so we could utilize those to draw CV Curves around areas like the mouth, eyes, nose, etc.

I actually like to high-level detail in Voxels, using the Surface mode...that is precisely what it is there for. To give the user the speed of polygonal Sculpting, so they can reach those ultra high details. If you only have 4BG or less, I'd recommend trying the image-based sculpting as Greg pointed out. Just take your retopo'ed model into Microvertex mode, where you can paint displacement and you can also bake out occlusion maps (give it plenty of lights for a more accurate output). You can also bake a normal map (to use along with the displacement maps) from the texture baking tool, so that you're sure to get every last detail. It might make more sense to you to do the detailing in ZB if that is what you are comfortable using, but with 8GB or higher, I don't think it's necessary to step out of 3DC to do the whole job.

I have a Galaxy GTX 275, and even though it's overclocked, I have noticed no problems whatsoever with CUDA. I've encountered plenty of bugs in the past several months and reported them to Andrew (they got fixed pretty quickly too, I should mention...thanks Andrew) and none of them were related to CUDA. I think you'll be perfectly happy with the card. CUDA doesn't make as much difference on the older NVidia cards, as the Manual states that CUDA performance scales in relation to the number of cores on the card. If you only have 64, it says you won't see any improvement over a quad-core CPU. But the 470 has about 448 cores and almost 1.3GB of VRAM, so you should notice a pretty big difference. You'll probably giggle the moment you have it installed and test it out.

I have 8GB and wish I had more...but then we are always prone to push every program to it's limits. Back to the original question, you want to use lower resolution when using large brush sizes and making broad sweeping changes, and increase the resolution only as needed when you detail...even then, that is where you would usually want to switch to Surface mode as it is designed to be much faster than Volume mode (it ignores volume data and only accounts for surface level data). Sometimes though you need to go back and make some large changes even after you have reached the detailing stage. That is where users are going to HAVE to start taking advantage of the Multi-Res capability. It's there for this very purpose...to make fast work of situations that slow you down in Volume mode, and it's wicked fast. If smoothing gets slow, for example, cache that layer (I like to stay to 2x level Proxy) and see what kind of difference it makes.

That's just too complex of a piece to expect near perfect results from the Auto-Retopo tools. Split the model up into logical parts, and try a combination of auto-retopo with guides, and manual retopo of areas that are just too problematic for the auto-retopo process. It can shave a lot of time out of the process, but not THAT much.

There are others here who have complained about the painting speed, particularly with large brush radius'. You can try to scale that down a bit when you notice any lag. Also, 1GB VRAM+ is recommended to help deal with such cases. We have been trying to get Andrew to Multi-thread the Paint Room (and mirror the Surface mode tools from the Voxel room over to the Sculpt Room...and multi-thread that room in the process). That will likely culminate in V4. Andrew has been finishing up the work with the Auto-Retopology feature set, so maybe something will start to develop on the Painting front soon. If it's anything like the multi-threading work he did with Voxels this summer, Painting will end up being crazy-fast.As for adding details to a Voxel Sculpt, you can get fairly high levels of detail in voxels with the Surface mode. That's what it's there for...to emulate the speed of polygonal sculpting, where it ignores Volume data and just accounts for surface level data. You can always add more detail after you have retopologized...just re-snap the mesh after making your changes and then you can re-export or re-bake your maps in the Texture baking tool. You also have Image-Based sculpting at your disposal in the the Paint Room. In Per Pixel mode, you're painting live Normal Map detail and seeing the result in real-time. Same thing goes for the Micro-Vertex Mode, except your are painting live vector displacement maps. If you need to make any large-scale changes to the mesh, you can use the Sculpt room for that purpose. If you need to make any boolean type operations on your model, Voxels are the best way to go. You can detail in ZB later if you need, or you can stay in Voxels and use the Surface mode for the ultra-high levels of detail.

Perhaps the hidden gem within Voxels itself is the ability to cut/boolean, trim, etc with Photshop-type selection and spline tools (as well as B-Splines)...and do so with abandon. Can't do that anywhere else.

Try this one...no bridge in it but you can glean how to do that from the many other tutorials available:http://www.youtube.com/user/Fugazi3D#p/u/14/cw1Xs5vqs0w

Ability to multi-select layers like you can in Photoshop. If you have many layers and sub-layers, and want to delete a bunch for any reason...right now you have to do so one layer at a time. That is way too tedious.As for the color indicators beneath each layer, for Color, Depth, and Spec...that would clutter the panel up something awful, I think.

As long as the black and white image has decent resolution (no jaggies) and is clean, it's a 2-click process. One to select the Logo Tool, and 2 to select the image. Bam...it's created for you in the viewport. You will likely want to increase the resolution on the layer at least once if it comes in at a small size on the grid. If it's large on the grid, don't. The larger the scale of a model the more voxels it contains.You can scale it immediately after clicking the image. You can use the Pose tool to skew or scale a portion as well. There are lots of things you could do with it from that point http://www.3d-coat.com/tutorial/voxel-video-manual/

Check out the "Split" tool in the Voxel Video manual. That will split the letter/number into it's own layer, after which you could use the transform tool to place them where you want them individually.http://www.3d-coat.com/tutorial/voxel-video-manual/