Cerbera

Welcome to the Core ! Make sure there are no axis locks in place, and then just get the scale tool, and click & drag anywhere in the viewport EXCEPT the gizmo, to scale uniformly on all axes ! Additionally you can enter size or scale values directly in the Coordinates manager (lower right hot-corner). CBR

At the moment we only have a select few producing dedicated scene nodes tutorials, but I think we can expect more and more of those as the feature set becomes more refined and 'complete' (such as these things are ever complete!). I think they are wise to work more in the background producing a solid base for the future than they are publicising / training for it too widely while it is still a work-in-progress. I look forward to the future wonders of scene nodes but I think their real time to shine hasn't quite arrived yet... CBR

Turn banking off in the sweep ? Scene scale of 1 cm should be OK, but I concur it is very jittery and unstable. I can lessen that by increasing substeps and damping in the sim scene globals, but I agree dynamics isn't playing nicely here... I think I turned on sim before generators as well, which might have helped ! CBR

Pffft - yep, the cloner way is a far better and simpler plan than faffing about with XP / constraints, and I am quite cross it didn't occur to me ! 😉 Nice one HP. CBR

You could try a Constraint tag (clamp, surface mode), but I have always found those to be a bit sketchy and I struggle to get the calculation priority right. What about using Xpresso and a point node ? CBR

First of all - are you modelling the outside only or doing the inside as well ? If it's the latter then you should start with the inside, which better governs the polys you actually need. What you have done here is fine but probably way too high poly - for example that hole on the corner should be able to be described with an 8 or 12 sided disc. By starting so high poly you condemn yourself to wrangling that increased amount of them all the way across the model, rather than only having poly density you require. CBR

If we're going to do it nicely / properly, a cassette tape is actually quite hard to model in any software, but none of it is unsurmountable if you have good reference and plan your parts, corner types (box vs inset type) and topology carefully. If you can get away with not modelling the inside (ie your cassette isn't clear material) then the outside is a lot easier than the inside, as your second reference shows... You would begin this model by outlining the holes and tape guide section, and connecting those meshes together across what is (helpfully) a mostly flat surface, which means you can do any loop expansion / collapsing on flat bits, without worrying about SDS lumpiness on curvature. In that respect it is not strictly necessary to exclude tris or ngons in these areas, but in practicality it makes little sense, and saves little time to do so. The front half of the case is notably simpler than the rear half, which contains all the extruded parts. It's not as much work as you'd think though, because we can deploy X-symmetry throughout, and all the parts that are chiral are best achieved with separate meshes... But really your first reference photo is showing you all the prime topology points you have to hit to make that inside, which mainly consists of ensuring you have enough topology to describe the hard edged inner details at the bottom, which then collapses down to less dense poly flow as they get further up where detail is less. Hence the liberal use of kite quads for loop concatenation as we see particularly in the section below... Note the box cornering that goes on everywhere there is a hard defined corner, which is required to stop this collapsing under SDS. Of course there are a number of lazier, cheat-ier ways to make the modelling easier (booles, volume builder, non SDS etc) but none of those will give you quite as nice results as if you model it poly by poly, SDS, where you have infinite resolution suitable for extreme close-ups. But if this cassette is just lying on a table somewhere at the back of a scene, then a full SDS model is probably overkill. CBR

At the risk of sounding like a broken record, it would be helpful to have the scene file here... But it looks like your character has unsuitable topology to deform properly when rigged, especially around the groin area. In any case we need to see what your weighting is doing before we can properly advise... CBR

I'd go Boole then ReMesher - perfect job, lovely topo. And then you could put that under Thicken or Cloth Surface then SDS, and you're done... CBR

As Mash said, only by turning them off. I suppose another option is the mesh deformer ? Deform your lower poly model instead, and use it as proxy for high res model... CBR

Some questions: What version of Cinema is this ? How many polys are we talking about ? Have you tried reducing the polycount via remeshing etc ? CBR

I would suspect Normals too. If there is still a Normal tag on your object, remove it. Then select all polys to check all facing camera are orange (Normals correct) rather than blue (Normals reversed). Lastly, in the phong tag you could uncheck 'Use edge breaks' to eliminate the last of the possibilities that might cause this. If still problems, then might be worth running an optimize to check there aren't any disconnected polys... CBR

Yes, in terms of least distortion. CBR

Here is a comparison of 3 of the available mapping methods. As we can see the cubic terraced islands approach produces by far the least distortion. The not inconsiderable distortion we inevitably get from any version of the UV peeler method is quite extreme if you do the outer and inner as just the single island, even in the conformal relaxed version (image C above) so by doing a cubic approach we spread that distortion out across the 4 areas we allow to remain separate, which lessens it accordingly. The price we pay for this is 4 partial seams per island* * You may notice that in A, the Cubic version I added a lot more seams than that along the base end to achieve zero distortion in that area, knowing that we probably won't see any seam lines down there... Note in the main top pic how the squares stay the same size at the bottom of the bottle all the way to the top, whereas the UV peeled ones can help but shrink those polys the higher up the bottle we go, which is why solution A is better, albeit non-intuitively so. Make sense ? CBR

Yes. You can replace them with nice quad patches like so... CBR

On the contrary, that is actually rather minimal distortion, and should be functionally fine in the render... However you need to continue to terrace the outside islands together and then the inside islands together so that you eradicate 3 out of the current 4 seams on those sections, as you have started to do on the outside. Same on the inside. Modelling-wise it's mostly fine, if a little high-poly, except you could quadrify the pole sections so they become non-complex, though it is not really necessary. CBR

There should be no need for weld and relax because you can use the UV Peeler to get that result directly if you use the correct combo of up-down / left-right as you drag. And indeed this method is perfect for straight cylinders where there will be no distortion. However, as soon as your bottle contains curves - ie the radius changes up the length of the bottle, and for example, you need something like embossed text running up a curvy section of the neck, the UV Peeler can't help you, which is why I suggested the cubic method, which at least shares that distortion over 4 quarters, which is often preferable to either proportional or equalised result you can get with the peeler. Just a thought... make sure you have Distortion (from UV settings menu) turned on and up to 100% so that you can see how much distortion are getting with the various methods.. CBR

That is not a great UV map because of all the unnecessary disconnected arc sections in it, which don't need to be there, and shouldn't be. You want the whole of the outside of the bottle except maybe the bottom face and top of the lip to be all one contiguous map; same again for the inside, which is usually achievable with cylindrical mapping, or by defining a single vertical seam and using the UV Peeler (U,J). OR you can take the cubic route, for minimum distortion, where the UV for the outside should look something like this... But having said that, the whole point of Substance painter is that you can paint over seam lines, so getting the UV mapping right before export is not as vital as you might think. And indeed you can choose to have SP entirely ignore your previous UVs and generate its own, which will be massively fragmented and island-based. And you typically don't need to UV bottles at all if they are just transparent glass and the labels are separate geometry, which is often the better approach. If you do need to UV, either because the bottle is not clear, or because you want specific mapping for scratches and damage etc you can usually handle that with Triplanar mapping, which again, requires no UVs. But having said that there only needs to be the one major seam in this, and therefore that's what you should do, telling substance to use your UVs instead of its own if they are satisfactory. CBR

and yet... https://www.turbosquid.com/3d-model/millenium-falcon CBR

Ok, well that'll be why it didn't work then - you don't have new unified rope dynamics... But I have modified the scene, and adjusted force values to use the original rope solver, which might make this work with versions before R25. Give this one a go and let me know if it works. I can't test it in R21 myself as I don't have it installed anymore... CBR spline swing legacy solver CBR.c4d

200 posts and we still don't even know which version of the software you have - why isn't this information in your profile ?! CBR

Like this ? spline swing CBR.c4d CBR

Lols. Yep, a bit in 'night mode' - only a small overlap between me and daylight at the moment... If you are happy with the VF approach (personally I am struggling to find any input sources for that that give me the sort of detailed, curvy shapes we see in photo refs) then Variation shader (Standard / Physical Render) does do a reasonable job of quantizing and mapping a gradient and some variably random tones over the pieces... ...but I rather think that this is only helpful in a cartoon / map style kind of way. I can't help thinking that if you want this looking 'real' in any sort of meaningful way then it has to be somehow based on those top-down satellite images, whether you pay for a solution like DEM Earth, or find a free way of implementing it. CBR

It's an interesting problem, because none of the fracture or noise based tools we have tend to create those sorts of very specific patchwork divisions. You don't actually say exactly what you are trying to achieve via modelling versus what you intend to do with textures. However 3 things may help. 1. Google maps does let you hide the labels and roads in satellite view and does provide that perfect top-down reference like the following I randomly grabbed just now... You could use that or something like it as photographic base I suppose, applied via flat mapping to a shallow landscape object or similar, and then add certain 3D elements on top of it using area scattering or cloning ? 2. I think you have a version of Cinema that has the variation shader ? That might be ideal for colourising sets of fields (providing you find a decent way to get those sort of shapes), and getting an organically varying colour range if you are not relying on photo texturing to do that. 3. If you fed something like the image above into illustrator (or equivalent) that should be able to trace the photo with spline paths which you could then refine and output into Cinema to help you get the field shapes, though uncertain so far what the full workflow for that would be - just an idea... I'll pop back if I have any more ! CBR

Like nearly all questions on this site, we could do with the scene file so it is clear what your setup and circumstances are, so that nothing is left ambiguous, and we are not left guessing what you want... CBR