-
Posts
17,955 -
Joined
-
Days Won
733
Content Type
Profiles
Blogs
Forums
Gallery
Pipeline Tools
3D Wiki
Plugin List
Store
Downloads
Everything posted by Cerbera
-
If it's a hero object up close and you will see other angles than front-on then it might be worth modelling. And then I think the old smoothing deformer / magnet tool trick could work to add some very art-directable wrinkles, if you get the topology density and technique right. You could further use the restriction tag to confine that to the areas inside the ultra-flat edge crimp sections or just apply its all the way to the edge and re-conform or extrude the crimped bits later... CBR
-
Welcome to the Core ! 🙂 So this is a Rigid Body Bullet animation ? What you need to do is cache the dynamics (bake project from cache tab of dynamics tag), then go to the frame you want to export, do current state to object on the cloner, and delete the dynamics tag from the resulting null. Check it has worked by rewinding the timeline, at which point the clones should stay where they are. CBR
-
Yep, my methods above seem to work in the main, although of course I don't have your bottle to work with, but I am able to get a nicely wrapping label with thickness and a pleasing overlap, and have done that with all the quads and SDS. I didn't actually have to split the circular bit off in the end, but once I had applied my deformers that were making it match the curvature of my (virtual) bottle it was not trouble at all to select the polys I wanted to offset, and move them out a bit on Z to get the overlap. I also went with different deformer set to you - I used bend and FFD to match the curvature but I don't think it makes a great deal of difference to the distortion you necessarily get along the way. The last step would be rejigging your UVs (if needed) so the label looks right in place, and to account for the backside now we have added thickness. CBR
-
Yes, I think this may be the sort of topology we need to nicely deform around that bottle... CBR
-
Ah excellent, now we can see what we're working with 🙂 Slightly unhelpful that they have double-labelled your reference bottle, but what can ya do hey ?! That is actually not so much of a serious overlap; certainly nothing we can't deal with if we allow ourselves the level of control we get by manually modelling. My plan is this. Model the label flat, out of polys, then UV it, so it matches the label, but when done split off the disc on the left hand side into its own mesh. Then I would whatever deformers you are using, like you did before, but don't try and make it match the bottle too closely - you only want the general curves at this point, not exactness, which will come next. At this point I would CStO the label (to bake in the deformers), then delete half the mesh so I only have to work on one side, but would instantiate symmetry here so the other side all gets done for me as I work. I could try and skip even that stage using the shrink-wrap deformer at this point though that will squash any overlap flat, so we have to bear that in mind, and fix it manually afterwards... Now I would get poly pen in reprojection mode, and I would move the points of the label down to the bottle surface, all the way around until they almost meet at the back... at which point we can abandon the symmetry and position and stitch on our disc section to get the overlap we need (that can have been independently have been projected on to the bottle so it also matches the curves). Once these are re-combined we can add control loops or edge weighting to give us the necessary sharp corners and infinite resolution on the label if we want to do it that way, add tiny amount of parametric thickness with Cloth Surface, and further adjust the offsets of the points around the overlap area until we get a nice result. If I get time later today I may show you the optimal base topology for that label so we don't have to dick about with splines and ropey old regular grid caps topology (nasty triangles round the sides etc). CBR
-
It's a bit difficult to advise on this when we haven't seen the full shape (flat) and we don't know how much overlap it has to have ! Not only that but you haven't completed your profile so we don't even know which version you are working in, or which tools are available to you ! For example if it is a very small overlap then it is no problem to extend and / or offset the geometry manually, but if it is a significant overlap then yes that might not be the best way... CBR
-
Still very much in my golfing modeller's cave, this time I'm back with some golf ball action - always an interesting challenge to get precisely the right dimpling layout, look and UV, and in this case to design an informative story around the various layers used in its design. Models by me, rendering / music: Kane Cochran for Rooted Solutions / Octane The secret to golf ball modelling is to lay it out as 20 equilateral triangle patches that combine to form an Icosa Platonic, which is then spherized to produce the overall topology. This gives us 12 x 5-point poles and 300 slightly complex 6-pointers, something that I would not normally allow in SDS modelling, but for tiling reasons our topology requires, and the deformation in the dimples is so slight, it is unproblematic in render. UV-wise, those tri patches are then laid out so that the 3 logos that appear on the ball are lined up over the centre of the UV with nothing over the seam lines. As you can imagine, S26s new Fit circle tool was essential to this build, and saved me having to to do points to circle 362 times, instead handling them in a single move (great job Luca / Maxon)...so that was nice. And more so to be able to do it with ALL the quads 😉 This layout also met a further criteria which was that it should be able to be bisected (for cutaways) without boolean operations, and indeed there is just 1 contiguous loop that does exactly that in this pattern. With a 4 x 4 patch like this we get 362 dimples, and with a 5 x 5 we get 490. I had to make both variations for this client, as well as a set and models for a brief advertising spot showing the development history of the golf ball... all good fun, unusually mathematical challenge, and nice to get a break from golf clubs ! CBR
-
Try this one... Pebbles CBR 02.c4d Here, just 2 basic pebble shapes (one based on a cube (but smoothed), another a deformed hexasphere) in a cloner, in object mode over a landscape, with random scaling and rotation applied via Random Effector, then the whole lot deformed by a World-space noise-driven displacer. We are using multi-instances in cloner, so no scene slow-down, even with 400, 000 pebbles. CBR
-
We are lacking information again here. Are you trying to make large, detailed 'hero' rocks, or are you trying to make many small pebbles that will just be tiny background objects and not the focus of the scene ? 1. Rocks are very complex surfaces if the camera is right next to them, and we need 10s of 1000s (if not millions) of polys to describe those sort of surfaces if we are doing it actually on the mesh using a displacer, as opposed to later with materials and bump / normal maps etc. If you reduce poly count too much (with remesh for example), you will lose detail, and possibly complain it does not look like a rock anymore. Those sort of rocks are NOT SUITABLE for cloning into a massive pile of them - your system will just fall over with the calculation. So if you want a whole beach full of pebbles then the geometry has to be much simpler (lower density) so that you can clone thousand of them without wrecking your system speeds. And that's fine, because if they are small then the camera isn't close to them, and so you don't need the extra detail ! 2. Rocks will look like spheres to some extent if you don't do anything to the spheres before you displace them ! To make my low poly pebbles I used an FFD to change the shape of a sphere into a more pebbly kind of thing, or started with a cube, and manually modelled it into a vaguely rock like shape ! You don't always have to start with a sphere... but whatever you do start with, make sure it has very even segmentation so it all gets affected by displacers equally. CBR
-
It's a fair point and sometimes you are right, but on other occasions it can be helpful to see what steps you followed to get to a certain condition. Its the info in your profile we need most though, otherwise we have to spend the first 2 or 3 posts of every single thread you start asking what software versions and renderer you are running ! Look at my profile ! You can see instantly what I am working with, so I never have to explain that to anyone, and no1 ever ever has to ask me ! Anyway, did my answer above help at all or are you still struggling ? CBR
-
Please take care to post in correct section - this does not go in 'News' ! Moved to 'Cinema 4D'. CBR
-
For your general information that Axis mode is activated by the shortcut L, which sometimes people hit by mistake when they are trying to get the loop select or loop cut tools (U,L / K,L) for instance, and they move their mouse in between the shortcut key commands. Once you are aware of that it is easier to avoid. CBR
-
He has either got materials applied to the meshes and materials ticked in the display menu, or he is using custom display colours available in the basic tab for any object. Were these materials specifically for Octane ? Otherwise they won't work, and you will need to create new Octane materials. Additionally, your asset error warns you that you don't seem to have a required shader asset for the tongue... To be honest, you haven't made it easy to help ! If you mention a tutorial, you should always link it, your profile should tell us which version of Cinema and Octane you are using so we don't have to ask, and in most cases (though arguably not this one) you should include a scene file so we can more thoroughly see what we are dealing with ! CBR
-
Whenever you need lots of smaller pebble type rocks, I would then get the Cloner out, and clone a field of 'starter rock shapes', which will be identical initially. But if we displace the entire cloner with World Space noise, we can get a unique rock for every clone we have, like so... And if we then collapse that cloner down, we can start another one to clone all those rocks thousands of times onto a terrain object (surface mode) to make a rocky seabed. Then we can randomly affect scale and rotation and distribution to give us enough rocks that you can't notice any copies... Let me know if you need more detail on that.. CBR
-
You're welcome ! There is no image required for that file - I didn't do any texture work and all the displacement uses for its shader is standard Cinema noise. You do also have all the controls I do if you have S26, but I suspect the course you are following in Cineversity is for an older version, and there may be some differences. Click the black square to the right of the layer shader, which will open it and show you what is inside it... guess what you'll find there ?! CBR
-
Because of translation issues, there is some question about what you might mean by 'pebble'. Please find a photo reference so we can see more precisely what sort of rock you are trying to make... If you mean a round smooth pebble for example then you could start from an FFD-deformed hexa-sphere to get the basic shape. If you mean something rockier then perhaps this will help... Rock 01 CBR.c4d To get a new rock, change the random seed value in the V Large layer. To change the overall shape, move points in the FFD (Deformer). To change mid level ridges, play with the scale and seed and relative level of V Med To change the characteristic of the small scale noise change the noise type / scale of the Pits layer. Hope that helps... CBR
-
1 layer of noise is almost never enough to create displacement rocks. Your tutorial probably suggested Voronoi noise because it is very good at producing those hard angular formations that other noises lack. However that will look fairly shit on its own. You are also using the worst type of Voronoi noise - the one that gives you that dent in the centre of faces - I would go with one of the other types, which doesn't do that. So what you need is to distort that First noise layer with a second, using the distortion in the filters menu, perhaps driving it with Electric noise, and then layering up a couple of other, more detailed noise types above that on top of it (like Luka / Buya, or FBM for example), but with lessening mix amounts, and in interesting layer modes (overlay or multiply works well) to give you the mid and small-range detail. So, all that goes on within a layer shader, which you can get by choosing it instead of the noise you have now (which will automatically become the first child of it). There is no single 'right answer' or 'perfect setting' I can give you - you have to use your artistic sensibility, and play with all the parameters, and try loads of different combos of noise types and scales until you get what you want. Also, remember that displacement modelling only gets you so far. The last, smallest level of detail should come from the bump map in your material. That can also be (high octave) noise-driven, where you can adjust delta to achieve stunningly sharp final detail. CBR
-
The only reason the claws disappear is because they were ALSO set to x-ray like the Aquarium. X-ray does not support multiple levels of itself, so if you have a x-ray mesh inside another x-ray mesh it will indeed disappear and that is expected behaviour. On another note, you have some incorrect settings in the Cloth Surface - you are adding unnecessary subdivision, and not adding any thickness with it, which is why it is there in the tutorial ! CBR
-
Although the spline pen can edit tangents directly I think it is widely accepted that Move is rather better for editing existing points, and for exactly that reason. The Spline pen needs its modifier keys for other things, so I don't see how we could get that functionality as well unless there was a further tickbox in the spline pen attributes 'maintain equal tangent length', and if you're gonna click that you might as well click the move tool ! CBR