cakeller

I agree about the problem. Here's a thought - maybe rather than just deleting the file name, which raises other concerns for me - having to remember what the last file-name was so I can manually increment it? yick Maybe either increment the file name automatically (if that option is checked?) or have an option to "prompt to overwrite render, and an option to change the file name" - that seems the most logical to me... overwriting a file should be caught before it happens. BUMP!

Looks nice... I've been working on a collection of new shaders too! can't wait to share them. this has a nice oily metallic look to it... an alien metal, or something. Nice start, keep it up EDIT: fyi, each folder under 3D-Coat-V3\Shaders\Custom is essentially a shader... just zip the new folder you created when you RMB create new shader, on whatever shader you start from. 3DC checks each time for any new folders and the name of the new shader is the name of that folder. so making new shaders is pretty straight forward. As a suggestion, maybe use a few letter prefix, like MH_shader-name... MH_PurpleMetal, or whatever. that way it's easy for those getting your shaders to use them. anyway... again, good to see people working on shaders! Cheers

You can import any OBJ file into the retopo room, and if it is aligned, even roughly, with the voxels, or sculpted poly mesh, the new mesh can be snapped to the reference (voxel, or big-mesh, or sculpted hi-res mesh). if these meshes are not one part, you can connect them using any retopo tools... if you DONT snap the mesh to the reference object on import (I recomend this, if it does not line up well) then when you move vertices around, they should snap to whatever reference objects are visible. --- so some of the function you are asking for, I think, is already there. However, I think the part that you mentione that would be extremely cool is if the auto-quadrangulate could be used to "fill in the gaps" not just as a whole re-topo tool, but able to use our hints. you could start with one quadrangle, or perhaps several all around in key areas (e.g. outline the mouth, eyes, and nostrils, and the ear, then let the quadrangulate tool do the rest of the blending work... that'd be such a time saver and give us more control... at least in my fantasy world perhaps the algorithms become more complex/// perhaps not?

First click "Use Background Image" then, click "Choose Background Image" there are 3 types of backgrounds - gradient, panorama, and background image. there are 2 menu picks for each (3 for gradient actually) Vertical Gradient Use Background Image Use Panorama then how to set each: Choose Top Color - choose the color for the top of the screen when using vertical gradient. Choose Bottom Color - choose the color for the bottom of the screen when using vertical gradient. Choose Background Image - sets the background image, but does not tell 3DC to USE background image. Choose Panorama Image - sets the 360° x 180° panoramic image (Also called Lat/Long images in HDRI-Shop) where X on the image, maps to latitude, and Y on the image maps to longitude, of the sky. good luck, hope this helps.

You don't have a space-mouse, or any other 3d type input devices? (Space Pilot, etc) I noticed no one talked about that possibility... as they need to have center "calibrated" or they can and will drift. And there's a button you can press in 3DC that allows you to resize by the space mouse. I am guessing this is not your problem, but since no one mentioned it, I thought it was worth at least mentioning.

Yes, actually it would be extremely useful if we could use the copy / clone brush on a mesh, instead of only for voxel layers... that way you would not have to convert the entire polymesh to voxels. In fact, it would be pretty useful indeed if you could load some entire polygon assembly (multiple parts, with their own layers) or reference any/all objects from the retopo and paint rooms while sculpting in voxels... how could this work? perhaps the idea of voxels being surfaces / or voxels could be as followes. perhaps there is simply a way to add the function to copy/clone from surface. Then add the ability to reference meshes in the retopo and/or sculpt rooms within the voxel room. to me this sounds not a simple change - although andrew is amazing, but still this is probably a lot of work, that could potentially add more bugs? this sounds like a great idea - and should be included in the future (IMO) however, is there any workflow that could help in the mean time? perhaps something andrew COULD do in the short term is allow to have a clipping cube on import - let me explain: upon import, you are worried about memory required for large reference mesh. well what if you could import a focused area... that could help. ESPECIALLY if you could repeat the import, and choose a new region but have this new object still in the same spacial relationship to the original... that could work, I think, and not have to completely rework the logic in the program. so a clipping box with min/max x, y, and z. this could be useful in a lot of other ways. for example, if you could snap the + or - of X, Y, or Z to the 0, this would allow you to import only half a mesh/model, and then mirror it later. also if the idea of a clipping cube were added, this could be another tool as well. which would allow easy clipping of FREAKED OUT voxels (which I've had, where 1 voxel somehow ends up 30,000 voxels long, which totally hoses up the voxel object. ;0. anyway... I think this could help?? anyone else like the idea of a clipping cube?

I agree. I actually suggested this a while ago. you should be able to freeze (as you said lock) the retopo layers. I think it's just that the retopo room isn't intended as a modelling room (although it works as such in some cases, like what you suggested) So the only work-around I know is to either move the object to per-pixel painting with no subdivision, or export the layer as a separate obj. then you can import at any time.

Thanks. Absolutely. That's actually exactly what I started doing. I use many voxel objects now. I crank up the resolution on, say, an eye-lid so I can sculpt down to the tiniest precision on a small portion of a model. Then I can reduce the resolution a little to keep the model relatively manageable. Then use the copy/clone brush to blend (somewhat) the edges in... smooth, copy, smooth, copy... but really,even if I merge the higher res object back down into the head, it's SO much better than if I just tried to do that detail at lower res in the head. Thanks all for the kind comments.

Thanks, You're absolutely right, this was early - and I was having trouble understanding how to make folds and stuff (which you can't really do in voxels.... it'd be like sculpting something out of gum, or mercury heheh... it flows together. however, here's my follow up head model... (not the same character though, and I should add the voxel model to the previous post (a few more images to reveal how bad it really is ... just a beginning. OK, here's the next one:

So, share I will... here's a Head model I started, using some reference images, just to see if I could. (I'm not a trained sculptor, so be kind) I am open to any kind of constructive feedback, thanks. This is just a first attempt (done a while ago)... Mostly it was about learning 3DC...

I COMPLETELY AGREE!!! this should actually be the way it is like in maya... ALL customizations should go in the user documents folder... and anything of the same name in the "customizing folder" overides that of the same name in the app folder. e.g. maya has scripts, called mel scripts, and any scripts you alter and place in your "scripts" folder in your <user documents>/maya/scripts/ folder, overrides even the maya factory shipped scripts in the app folder. This would be great for things like shaders, and etc. that way users don't MONKEY with the install folder. And in the event that a customization messes something in 3DC up, you just delete, the customization folders in your docs. (or move it somewhere else temporarily to make sure it's your customizations)

It would be helpful if you post your system specs, OS, and which version (DX, OpenGL, CUDA / NO-CUDA), how much RAM, and Graphics Card, etc you're running. If the graphics is under-powered, that could affect the creation of the DirectX or OpenGL "Device Contexts" I tested both GL and DX versions (no cuda here) on Windows XP x64, 4GB Ram... never mind, look at my sig (has my system stats, as does phil's, makes it easy to not forget to include on a post ) anyway... defragmenting your hard drive, believe it or not, can actually SLOW things down for a short while until the dust settles. Like that line from that 80's song, "She's tidied up, and I can't find anything" - Thomas Dolby --- but give some more details, and maybe we can track something down for you.

Vector displacement is absolute and needs no coefficient if you use Zero is Black, and Not Normalized. Then the values in the displacement are absolute, based on the chosen space (Tangent, Object, or Abolute)

Voxels are, at least at this point, [on a regular spaced grid] and highly optimized. I am almost certain that the octree is extremely efficiently culling out unnecessary regions, and freeing up resources that so nothing is wasted on large empty voxel space. However, there have been discussions about multi-resolution voxels, and Andrew has said he is working on a way to make this work... I'm not holding my breath because I doubt it will happen overnight, but am interested to see what he comes up with. And to be honest, the guy is so fast, it wouldn't surprise me if it happens sooner than we all expect. But I also think, right now, he's working really hard to un-bug all the little bugs before re-working the core of voxels. Anyone feel free to correct me if I have misspoke

When you export vector displacement maps, if you use Zero Level is Black, NOT NORMALIZED, then the displacements will be absolute and not need any coefficient. It is the most reliable way to use vector displacements, as ALL the data is retained. Of course you have to choose the correct "space" If this does not answer your question, perhaps someone else could answer - or you might elaborate why it does not. By the way, I think a lot of people might not know this - Normalization means to bring it within the range of 0-1 (or possible -1 to 1) but of course, that will mean you'd need the coefficient... but if you want the actual value, don't normalize it.

Actually... I just had a thought... it'd be cool if Andrew could add the RMB, MMB, and LMB to the list of "EVENTS" then you could customize any key really to function as that button...

And I wish the "Maya" setting ACTUALLY worked like maya. When the Y Axis is upside down, the rotation should be inverted (so the visual rotation matches the mouse direction) as is, left / right movement ALWAYS rotates around the Y axis in the same numerical direction. so when upside down you slide your mouse right and the model rotates left.

I HIGHLY recommend getting a 3 Button mouse if you're going to be doing 3D. Or better yet, check out the new multi-touch mouse! that thing looks AWESOME! And if you use a wacom with multiple side buttons, you could always assign the side buttons to MMB and RMB... It is a huge slow-down, IMHO to have to use extra keys to simulate MMB and RMB... especially since CMD+MMB and CMD+RMB are already navigations... I'm not saying you are wrong to want this, just offering up an alternative that you MAY be able to use

Click the [Camera...] in the upper right corner of the screen, choose the sub-menu "Background" then click the menu-item "Choose Background Image" unfortunately, the background image stretches to fill the screen (at least it does on XP x64, GL and DX versions... so you have to somehow get the aspect ratio right, which I have no idea if that can even be done. Also, you're supposed to be able to save out alpha with the image, so you could always composite your render in your fav image editing software later. good luck, hope this helps.

I had this problem in earlier version (the one uploaded per your request on your thread, and had problems with it. And still in this I have the same problem... however I fixed it for me by adding an additional input parameter in the XML... Attached is my edit of the XML, and although it adds a useless parameter (I think it's useless) it prevents the non-replaceable CustomSampler4.dds (because what happens for some reason is that even though it's trying to change 4, it saves over 3... and 3 is not used in the shader... it's a workaround. See attached XML I added this between the CustomSampler2.dds and CustomSampler4.dds <ExShaderParam> <ID>CustomSampler3</ID> <Type>texture</Type> <!ShortName>RainbowMapUnused?</!ShortName> </ExShaderParam> seems to fix it alright Variables.xml

Yeah, this works better for me too... I HAD A TERRIBLE time picking verts on surfaces that fold a lot... and as artman said, it's just a working setting, you can set it back to 0 or whatever.

in customize navigation you can add a key+mouse combination to "roll" the camera. OR you can, as I do, use rotate around screen space instead of around vertical axis which allows easy (to me anyway) positioning of the model at whatever angle. Which, is important for exactly why you said, for seeing areas. either way... in the camera menu, click on customize navigation. add an action, change it from whatever it comes up as to "roll" or "roll in screen space" and then "Assign Keys" to make it work. (As you can see I used Ctrl+Shift+RMB) good luck...

ARG, I have these rings! I assumed the rings were just an artifact and part of the way it was supposed to be (although they didn't look nice, I just don't use them) Anyway, I have the nVidia QuadroFX 3450 with 256MB and the rings show up in GL and DX... I still get all the same rings as mentioned before... any thoughts?

PLEASE! With Sugar on Top

Andrew has since fixed it so that fall-off and brush alpha and focal shift are taken into account for the move tool!