cakeller

Actually I've seen that discussed before (about deleting 3DC and clean installing... the license is PROBABLY just coinciding with the clean install... and probably not the problem with the lock-up. However you did the right thing to fix it! --- there ShOULD be a dialog that pops up when the license is not valid... which is why I say it's prolly just coincidental about the license.

Since I'm a tinkerer, I ofter TRY whatever to see if I can break things... much to my excitement, 3DC is INCREDIBLY ROBUST and ... shall we say, trusting of it's users MOST numeric entries can accept values that are outside the widgit's imposed slider limits. By typing, I can put in a depth of 50000 for the Voxel "Fill" tool, and I get more than just an interesting effect, it becomes a very FAST fill tool. MANY things in shaders are fun to explore out of bounds values... e.g. the bump coefficient in Marble, when set to negative, simply inverts the bump!!! AWESOME VEINING effect... the list is nearly limitless because the software is so robust, thanks to the virtuoso behind the coding! So, with all that I think it would be GREAT if you provided a way to "unlock" this functionality in the UI, for those that know about it - I suggest that if you are holding shift, while sliding on a slider widget, you should be able to slide beyond the normal limits... this would be GREAT! It seems like something that would be relatively easy to implement and unlock a lot of really neat functionality for those that don't know you can input out of bounds numbers.

interesting... I dont' have a way to test on the mac... is it open GL version then? sorry to hear that it doesn't work on mac. I can't test it easily on mac.. but I have asked for assistance to try and resolve this issue. If anyone has this problem on mac also, or linux or pc, please let me know. It's possible, because I don't actually know what version of the GLSL spec is the lowest supported by 3DC. So it's possible since the shader was written to support GLSL Spec 1.40, if your graphics card only supports 1.2 - maybe that is the culprit. I will look into this more, but it would be of great help to know all the pertinent info of your system... OS version, graphics card, etc. version of 3DC you're using. this shader is only tested on 3.1.18 Non-CUDA, GL and DX thanks immediately after starting up 3DC, when start splash (what do you want to do dialog) comes up... press CTRL + ~ (probably command + ~ on mac? ) to bring up the 3DC Log File... copy and paste the following information about your graphics card - thanks. this is an example - (mine, when running the GL version) ------------------------------------------------------------ Graphics card and driver info ------------------------------------------------------------ OpenGL Renderer Vendor: NVIDIA Corporation Renderer: Quadro FX 3450/4000 SDI/PCI/SSE2 OpenGL Version: 2.1.2 Supported: Pixel format Multisample Multitexture (units: 4) Anisotropic filtering (level: 16) Texture Compression Stencil two side (Multivendor) Vertex Buffer Object Vertex Attributes (count: 16) OpenGL Shading Language Frame Buffer Object (max size: 4096 X 4096) MRT (targets: 4) Float texture Non power of two [EDIT] Thanks to Akira, all my speculation above - well... never mind.

Earlier I couldn't get the GL version to load (nVidia QuadroFX 3450) ...so I was stuck with the DX version - but now they both work. However, I had to change the settings for openGL to default to Generic 3D Gaming, and create a profile for 3DC. Then no more openGL problems. However, please note, that Andrew says the DX version is faster in most cases. Even on my workstation card (albeit not a new cutting edge one) I can see a difference where DX is faster... have you tried the DX version? does it also crash? have you updated your graphics card drivers? if not - don't do that first just asking. have you changed any settings on your graphics card driver for Open GL? ... anyway - good luck, hope this might help you troubleshoot... maybe it's something trivial, or... maybe it's not you at all, and instead a bug ... Cheers, Christopher

Please bear in mind this is a rather complex shader - and I just finished it, so there may be problems with it. However, I have tested it heavily on my own. But, you never know, different OS, different versions of 3DC... etc. Anyway, Please try exploring illegal values in the settings. It's cool because you can achieve effects that vary wildly within the same shader. It's not a perfect monolithic shader like the MIA_X material, but it is an attempt to build a highly flexible shader for use in 3DC. You can make materials from jade, translucent glass, to stone, to metals to plastics. All without even changing the 2 picmat images. When you start changing the picmats, and the texture maps, the possibilities are vast. The two picmats are diffuse and specular, although you can use whatever picmat you want, but the specular picmat has an exponent setting to choke the highlights down for lower contrast images. You can also colorize the picmats using a single base color, which is also used for ambient lighting. There is a slider that goes from -1 to 1 for colorizing each. While a value of 1... well colorizes the picmat completely using the base color.a value of -1 does the same, but inverts the base color's effect. This can be used to create a nice complimentary coloring scheme. Warm hot spots of light, and a cool diffuse ambient lighting. or similar. There are 3 texture maps that are projected along each axis. Each of these images can be scaled (again you may need to enter values above the maximum of 100, if you want to span large voxel models, I just have no way to expand the slider beyond 100) these can also be offset from 0.0 to 1.0. where 0.5 means offset by half the image... in other words, the scaling units are in voxel units, and the offset is relative to the image coordinates (u, v) Finally there is also a group of settings that affect the way selections and freezeing is displayed... ...OK I just worked out a bug that fixed it so now Specular highlights show even if the shader is transparent! (like reflections should!... no back faces, but that's outside the scope of shaders in 3DC actually... I think) I hope you all enjoy this shader - it is basically my vehicle to explore GPU shading, using HLSL and GLSL fun treat! Since I'm going to call this "beta", please let me know if you have any problems with it. It runs just fine on my machine - but I'm not with CUDA, or mac or linux. I'm on Windows XP x64 and don't have another way of checking. thanks for any constructive feedback. I attached the shader, and then a few images to show how the pose-tool's selections can look... With the selection set to Frz&Sel Object Opacity (transparent freeze and selection) 1.0: With the selection set to Frz&Sel Solid or Rainbow - at 0.0 (solid color selection, freeze ignores the color): With the selection set to Frz&Sel Solid or Rainbow - at 1.0 (rainbow color selection, freeze ignores the color): Here's sorta what the shader looks like - although it can look like a great many more things than this (NOTE THIS IS ONLY ONE SINGLE SHADER WITH DIFFERENT SETTNIGS, NOT COPIED OR DUPLICATED IN ANY WAY!!!): 2009.12.06 ----------------------------------- Name: SN_BETA_SUPERSHADER: SN_Beta_SUPERSHADER.zip removed due to bug... see below for SN_Beta2_SUPERSHADER.zip - updated Material: Advanced Dual PicMat Material Surface: it's Comlicated Color: Whatever you want it to be... Compatibility: GL and DirectX

Please post Direct X or Open GL version, CUDA or non-CUDA. Also pleas post your graphics card.

thanks Taros (Chris) I didn't think it was a bug, I thought it was a changed functionality... but if it's a bug that can be fixed!?! yay! all the better thanks for putting the post in a better place. regards, cakeller...(Chris alsp)

Actually yeah, good point! Apply sym in UV world SHOULD totally use the same UVs ... as option (some may want assym painting on teh object of course, but for some assets.... anyway - never mind, this is just a good idea. option to use mirrored UVs!

WOW sorry I missed this... HECK YEAH!!! in fact even if you don'w have a 3D mouse... you could have a gizmo to rotate the "draw on plane" or you could have a nav type hotkey, and rotate with the this is a GREAT idea... +1 (and another if I were allowed 2 votes ) the tranparent plane like the DOF location plane would be great too, because it would intersect with any and all geometry where your sculpt stroke would be... allowing you to control where the stroke goes! I hope this could be impelemented, it's so good!

yup, I GET that, but... there are times when there are teeny details I want to fill, but don't want to affect gross scale features. If the resolution of the depth is lowered by making 100 equivalent to 5000, then 1 is going to be like 50, and 2 like 100... leaving much less granularity between numbers to be able to deal with the teeny tiny features - if you get my meaning. Especially when it comes to modifying the depth on the fly - it's already hard enough to adjust settings on the fly. But, perhaps if there was a different representation than depth it would be better. what I mean is, I don't want to lose the ability to work with fine details and have controlled useage, but I admit there ought to be an easily recognizeable effect when set at 100... perhaps it would work if the depth represented a logarithmic value? 1 = 1, 10=100, 100=1000 ... or something like that. in addition, perhaps the radius also scales the effect... e.g. a radius of 5 has a given effect, and a radius of 25 has an effect (depth) 5x greater for the same value in depth? I can see flaws in that as well - perhaps it would just be best if there were an additional controller called scale of effect (like how there's growth and thaw for the increase tool, there's no reason there couldn't be depth and effect scale. both within the 1-100 range would give you an effect of 1 - 10,000, which should then make intuitive sense to you and not lose what I need from it. of course, I may be the anomale... and I definitely am a tinkerer so I never mind if there are quirks to software - but you are right - one should absolutely be able to adjust a setting with the controls and get some results without having to stumble upon it, like I did with many of the funky cool features in 3DC.

probably depends on the density of the mesh? I noticed a huge difference in the speed of fill will the SF Fill... but I was only using low res test.

I agree - it's bothersome when you have to scratch your head to figure tools out. I think it is a partial misconception of what the fill tool is supposed to do. It's not to FILL holes. It's to fill grooves. or rather, it is to ADD material only in concavities. period. Increase fills holes better than fill. (notice fill doesn't use the term "fill holes" just fill.. it fills crevices, or cuts off high parts if holding ctrl. that's the design of that brush. it has it's uses as is without changes.,,, with the exception of possibly the controls) I like the way the tool works as it is - maybe the cotrol slider's soft limitation of 100 should be relaxed so it is not unintuitive to crank it up. but it works exactly as is. The fact that you can cheat the numbers (put negatives in, or numbers larger than 100%) is one of the coolest things about the controls in 3DC, besides the fact that doing so doesn't CRASH THE PROGRAM! I PRAY THAT NEVER changes. limitations are for Wussies! hehe I Don't like to be told what I cannot do! I'm not saying there's no room for improvement, but since I use and like the tool as is, and you're JUST NOW starting to use it - perhaps give it a chance before suggesting it be changed (again, other than perhaps the slider limit) Maybe I'm alone in liking it as is?

I am suggesting a SMOOTH ALL feature for voxel surfaces. ? what does that have to do with what andrew stated today (yes, I know he said he's going to make BAKING directly from voxel SURFACEs available) or maybe I shoulda just said... HUHHHH????

I think it would be a great help in the updated sculpting tools, and make them much more useful if there was a Smooth All command (of course limit to un-frozen). It would be especially great if it was a surface preserving type smooth like Laplacian smoothing which just reduces the surface energy (works toward average edge-lengths) without losing (too much of) the form/shape. It would improve the usefulness of the updated surface tools. Since as you, yourself suggested, the surface mode is for when you need to continue to refine a model that cannot be voxelized dues to resolution limitations (4-6 M Polys ) Since you cannot re-voxelize super large models anyway, it would be really important to be able to FIX surface anomales. but really - since the surface tools are fast - not having to revoxelize them , even for lower resolution models, would be a big benefit. ------- SORRY, I meant MLS, or "RIMLS (Robust Implicit MLS) variation which is an extension of Implicit MLS preserving sharp features using non linear regression" ... Oztireli, Guennebaud and Gross, "Feature Preserving Point Set Surfaces based on Non-Linear Kernel Regression" this smoothing method works really well ESPECIALLY on high-poly models... Basically the points are allowed to move as far as they need to to get good smoothing, but are not allowed to leave the orginal surface - they can slide around...

if all you want is to create your own PicMat style shader, right click on any of the existing PicMat shaders (e.g. Red Plastic), and "Construct New Shader", then give it a name, then when the settings box comes up, Choose a new picmat image. you don't HAVE to write any code, because the shader code will be copied automatically for you. That's not to say you can't write shaders on your own... because you can. I have learned a lot just by opening up the shaders, copying them and experimenting. There are a few little quirky things to 3DC implementation of the GLSL and HLSL shaders but for the most part pretty standard vertex and pixel shaders. However, I think you really were just asking about how to use your picmat images, and the short answer - copy a shader, replace the picmat. the shader name is the name of the folder. couldn't be much more simple!

Actually, you can CRANK THE HECK outta the depth 5000.0 fills holes quick like. but as Artman said, it is only for filling holes. in fact, 25,000 works smashingly well - so. but since you might want it to be more sensitive for something, it's best to just accept that you can crank it high. (as opposed to requesting that the strength be rescaled, which might be a logical conclusion) fill is nice to use on lower settings for REALLY tiny features. but when you want to blast away and fill in crevices - crank it up ! 5, 10, 25,000 etc... by the way, the increase tool is similar - you can get some really fast effects by cranking it up. of course, it's a little harder to control - but there are times when it's helpful... for filling holes that are open trough, since the "fill" tool can really only work if your brush is surrounded by surface... mostly.

That's true... physical limitations yup yup... and yet somehow he still does things that others would not given the same physical limitations - almost seems like magic already?! In truth, although I spent time re-explaining something that I thought was already explained enough to implement, it did cause me to have a couple of thoughts about special cases, and what happens if (...) ? things like... what if you want to edit just a portion of the curve segment, how do you limit the modification to a smaller region? knowledge assumptions: using NURBS curves - and assuming regular parameterization from 0 -> 1 along a single segment, or a signle change, excluding branches. 1) you could have a limiter along the curve that has a start and a stop for the region of effect, and a soft start and soft stop, to allow fo clean blending. these could be manually placed along the active curve parametrically, or using a gizmo of some kind. However, as an only method of interacting, this would be clunky and slow. this would be better suited for specific and precise reshaping of an area. 2) using the preceeding concept of a start, soft start, soft stop, and stop, as contribution parameters, with equivalent values of 0, 1, 1, and 0... and not related to the overall contribution weight, but like layer alpha in PS, it is not related to the layer's overall opacity. so with that, the start and stop are determined by one of the following - comparing the screen space old stroke to the start and stop point of the new stroke, to determine the closest projected point on the old curve, and mapping the new stroke to the curve using closest point for start mapping to start on the new stroke. if start (param 0) on the new stroke is closest to param 0.34 on the original curve, and end (param 1) of the new stroke is closest to param .37 on the original curve, then map the new curve's shape to the area between 0.34, and 0.37. 0 maps to 0.34, 0.5 maps to 0.355, and 1.0 maps to 0.37... there should probably be a percentage of beginning and ending of strokes that is consider the soft region. e.g. if 20%, then at parameter 0 on the new curve, there is no effect, at parameter 0.2 the effect is at maximum, until parameter 0.8, and then fades back to no effect at parameter 1.0 anyway... the point is, although I've spent more time than I would like to have spent on this typing and not sculpting, there was at least SOME good thought that came out.. [EDIT:] turns out I had done a screen capture video somewhat showing what I was thinking WAY BACK... I just never posted it, I think. it's my attempt to show approximately what would happen with the curves tool. all it really shows is imaginary stroke shapes, and then I go in and edit the points to show what I envision happening from each stroke. I guess I didnt' post it because I thought it might be more confusing as to what I was showing than just trying to use words.... but maybe now it could help? hopefully it doesn't fuel the confusion fire. Mock-up video

What's confusing about blending 2 curves? Case 1 - pure blend (think of a surface ruled between the two curves... and the new curve is some % from one curve to the other... and all the attributes could also be blended at the same weighting, or even individual weightings. Case 2 - modifying a sub-segment of the original curve... the new stroke is shorter than the original curve and close to a portion of the original, there would need to be a blend beyond the ends (simple) or within segment... to create a smooth transition (just leave it up to the artist to blend - like well it'd be your responsiblity to align the beginning and ending or feather it out. Case 3 - limited movemement allowing the blend of the original and new stroke to occur in only some of the axes etc ... meaning settings to make the averaging work only along the X axis or only X and Y, or such that the only movement is in screen space, so the projection of the new curve from the viewpoint, and then the original is blended such that the entire curve is still parametrcially at the same depth, but from the camera, it is blended with the new stroke. ---- but all of this SOUNDS complicated... just like trying to explain to someone how a car works, how to shift, and clutch, and gas etc... but in practice it's so easy ---- all of thise could be done simply by adding a few small changes to the curves tool - I would appreciate specific thoughtful questions about how this would work - but would appreciate it if folks would refrain from saying "I think this isn't necessary" since that's literally pointing out that they don't get what I am talking about. Call it arrogance, ok, but I KNOW this would rock - an if you have a specific question, or ask for clarification - I'm happy to oblige, but I'm not going to try and re-explain, until I actually have some time to illustrate it... besides, this has been around for a long time - and I haven't heard Andrew mention being interested in it, so - - - maybe it's just a lost cause - - - and that makes me sad.

wall of text hah funny... but this is nothing like soft stroke - soft stroke is another good feature, but doesn't in any way fulfill this feature request. HISTORICAL STROKES that are EDITABLE!... do I need to draw a picture here guys?

how about PERSISTENCE!!! keep it the way I did it last I'd like that (persistence) on most settings, and most are that way. but if not, would always be nice! (sure have a reset button to restore the default, but save my settings between uses) Retain all previous settings, not just a few.

Good idea! +1

Fair enough, I guess. Regarding your tablet complaint, THAT IS EXACTLY WHY the multi-stroke concept would work AWESOME and intuitively, and I thought I had explained why - but if not, here goes again. A signle stroke is difficult to get just right. there is a visual and physical disconnect between the tablet and the screen, which makes getting a precise stroke, challenging. However, with practice a stroke can become better, but rarely perfect. Enter the multi-stroke concept. Forget for a moment, about 3DC and think only in terms of a pencil stroke on paper. Have you ever watched an artist sketch, where they lay down light sketchy lines until they find the line they really want, and then they begin to push harder where they know the line is good... they might even rotate the paper to get a better physical angle to move the pencil. In the end, the line will have character and motion and emphasis where the artist determined it. Imagine suggesting to that artist, ok - make your line on the paper, then we'll make it rubbery, and you can have little points to move it around - ugh, AWEFUL! not intuitive, not fluid, and very inorganic, ESPECIALLY for thick and thin lines etc.... on the other hand a combination of the two methods would be great... do the direct sketching/resketching and then have the ability to go in and edit the curve based on the spin, and such. so the idea here, is taking that into 3 D... but same basic idea... the direct manipulation of the curve by the artists hand, not through widgets and gizmos, but the extension of their mind - meaning their hand. so --- back to the multi-stroke If you can make multiple strokes averaging them, it will be MUCH smoother, but allow you to hone in on your desired shape. your complaint about tablet is exactly the benefit you would see with a multi-stroke shaping of a curve... (see the math above for how the strokes could be combine) in addition, it's completely reasonable to think that you could limit further strokes to affect only X, Y, Z, or Screen Space, or PRESSURE The point is, the implementation I envision would be incredibly user/artist friendly to stroke a curve at any time - pick the curve-entity, not a whole curve tree, but one branch (or the trunk)... this is already the way curves works anyway. stroke along the curve (or not along it, so you can reshape it), with varying pressure, and from multiple views each time the shape is being averaged with the current stroke and previous strokes.... the curves are then averaged based on a weighting... you could set the weight so that each additional stroke only contributes 20% compared to the original which contributes 80%... therefore you'd have to stroke the same stroke 4-5 times to get the original curve to move half way between the two curves. or if you set the weighting to 50%, it'd take about 5-6 strokes to get almost to the new stroke. keep in mind, the cool part of the idea is that if the curve is only allowed to move in screen space (depth from screen is not changed) then you can create complex curves by sketching. there really isn't a 3D interface that can do that with a single stroke - since the stroke itself is 2D. so the only real way to get 3D is either manipulate manually (not intuitive, or fluid) OR intersect 2 strokes... but better, is to be able to intersect multipl strokes and be able to continue to reshape it.

I think one of the big issues he's going to have is - which direction the offset happens... sure it makes sense for any of the brush tools, like pince or voxel-shift, pinch etc... but mathematcially what direction is "UP" when you're not stroking the surface? so for any tool that ONLY works on surfaces (pinch, draw, etc) then it's no problem, because it doesn't generate geometry. for any of the tools that can generate geometry out of thin air, the normal vector to offset along is potentially hard to determine.... look at your initial stroke - which way should it embed or offset??? show a mathematical proof that explains a non-ambiguous determination based on the data of the stroke that says - ok, I know which way is up at all times. As I said, it's mostly useful on the non-CREATING tools anyway - so... the offset shouldn't be hard for that. for the free-space ones, it's probably more difficult.

Brush Depth is the control of how much the brush is going to apply (don't use that term, because it gets confused with the already existing setting "DEPTH" Embedding (how much is something embedded) Offset (how much is something offset - or in negative, inset - from the surfacE) but Stick to Ground (do it, or don't, there's no implication of a quantity... ) It's also a setting used in Snakes, and Toothpaste, so it doesn't makes sense to confuse the two - they aren't the same. Stick to Ground is a behaviour - the stroke sticks to geometry, Embedding is how much the actual geometry is created offset from the stroke by the surface normal. they are two separate characteristics. Please call it STROKE - EMBEDDING or STROKE - SURFACE OFFSET and leave Stick to Ground alone because it will confuse the subject and get the request ignored. (A GOOD IDEA REQUEST IT IS, and I WOULD LIKE to see this happen)

why do you need to stroke exactly over the top of a previous stroke? the idea is to BLEND multiple strokes to be able to re-work strokes and reshape them. How is it better than adjusting individual points? - uh, because it's an artists intuitive stroke! that's why... the idea here is to have a "sensitivity" similar to the way an artist thinks / draws / sculpts. NEVER EVER, except in rare savant type cases does an artist lay down their strokes perfectly on the first go. think blue-line drawing as comic book artist do before inking. or sketching of any kind.... the tablet is integral to using a program like this, if you don't have a tablet ! GET ONE ! it isn't required, but if you're serious about doing any kind of meaningful work - I would highly recomend it. it would be AWEFUL if it were not intuitive, but it's quite simple.... mode = shape, then additional strokes on a curve "re-shape" the curve, blending the previous strokes with the current stroke. hold control to select a different curve segment to re-shape. mode = create, lay down curve strokes like snakes tool / sphere brush if you're so in love with the sphere brush, you should really check out the snakes tool - it creates almost the same type of stroke, except it's a smoooooooth curve, where the pressure is thick to thin etc in a smooth way, not like the sphere's chunkiness. anyway... I've already suggested the pseudo-math for this, and it seems really really really easy to me, I don't know why it would be confusing, but then that depends on how much other 3DC stuff is new to you... This is a concept we've been discussing here for several months - A lot of thought and effort has gone into this, and it's pretty well thought out. Feel free to add new thoughts, but hopefully you try and thoroughly understand what has already been discussed before dismissing it. of all the requests out there, this is my #1 request - I WANT CURVES II... a 3D sketching tool with the ability to re-shape the strokes. there's no other way to do it unless you can stroke, rotated the view, stroke, etc... unless you manually edit (which is NOT artistically fluid)