Jump to content


  • Content Count

  • Joined

  • Days Won


Everything posted by Cerbera

  1. Another way to go would be to make a specific x-ray style shader in Cinema (usually consisting mainly of an inverted fresnel in the luminance channel), apply to the elements you need highlighted, render those separately and combine that with the masks suggested above in post. That way you get a slightly more detailed and unique look to the overlays. CBR
  2. Very nearly posted that myself Interesting that Arrimus views it as important enough to be an entire paradigm shift. CBR
  3. Yeah I'm not saying it's not possible to get close with the loft, just that the whole way that generator works is not particularly conducive to the task at hand. For example here is a fully parametric setup for that shape using a primitive with slice, an array and a correction deformer. Obviously you wouldn't use this for the whole model (indeed I'd venture you would also have to make your loft editable at some point in the process), but as far as flexible starting points with even geo go, for me this is winning over the loft approach... CBR
  4. Thanks for flagging, but this one is known, which I think is limited to nvidia video cards. Hopefully there will be a fix for it soon. CBR
  5. No, that's not how a loft works. Unlike Sweeps, the topology of lofts is entirely independent from, and unrelated to the splines within it. All topology options are in the loft object itself, including segmentation in both directions, which needs to be set one higher than the target amount (a weird quirk of that object within Cinema (now fixed in R21). If precise control over topology is the main priority here I would go as far as to suggest that a loft is not the best tool for making this form; it can be made faster, and as regular as you need with poly modelling (and optionally rotational symmetry). CBR
  6. Oh dear - that doesn't sound good. I haven't noticed that, and I am constantly saving and re-opening files in R21 / Win 10. Not a single one of those has been wrong that I have seen, but then again I don't use XParticles so haven't been able to check that. You should send support tickets about this, first to MAXON and then to Insydium.. CBR
  7. The way you are going is mostly fine, although you are working at a slightly higher polygon density than you require initially, so making a little more work for yourself. And you don't appear to have symmetry setup, which is something I'd get in place from poly 1. Lastly your loops are a little uneven in places so could do with some ironing, or better still (if you have HB modelling bundle) some Round Edge tool. CBR
  8. Their own proprietary software I would have thought. They have the sort of budgets and talent where it is feasible to just code the functionality they need. Whilst nearly all the studios use one of big 4 DCC apps (Maya and Max still winning) the film guys just write extensions to these things to make them do what they need at the cutting edge... years later that becomes a module or plugin set that might get sold off, or watered down and copied and incorporated for the rest of us in future software revisions of the DCC or as stand-alone entities... CBR
  9. In the case of a circle spline, that is why there is an nside as well. All other spline primitives contain only the points required to describe their shapes, and that is not changeable except in the case of the helix. Uniform interpolation also does let you choose amount of sub points, though these won't be visible or editable until you CStO the spline. It works like subdivision does on polygons. So if you have a 'circle' defined by just 4 points, then setting that to uniform / 1 will give you an 8 sided circle. In the case of editable splines the regular tools are enough to add additional points wherever you want them. There's also MoSpline, which can exactly match the shape of a source spline but with a more definable point count. See below, where I have made my mospline have 12 points independent of its source circle which has many more... CBR
  10. I used Disconnect from the r-click menu in edge mode. That does work, but unfortunately leaves both edges (or points) selected so they continue to appear joined when moved even though they are not. You can use the structure manager to see which points you have selected and to deselect the ones you don't. CBR
  11. I have to go out now, so don't have time to help further tonight. But I can maybe save you a bit of work if I just give you the model. Material Folds.c4d CBR
  12. Slightly refined version, with less control lops for nicer corners... CBR
  13. Thanks DV - I found the joint tool now, where you can use Shift to split existing joints into smaller ones (without breaking hierarchy), which is what I need to do for the head., as the whole top front end should bend upwards as its mouth opens, and the head joint in the fish rig does need more flexibility than its single joint can provide... CBR
  14. Yep, that can all be handled by regular SDS poly modelling. Start with a cube of the right dimensions, 8 x 6 x 8 then make editable. Then select these edges and disconnect them. Now move these polys out and down a bit... Then select these edges and ctrl-drag them under bottom of the model, then scale in slightly... ...and do the same with the other side of the corner... ctrl dragging inside the flap first a tiny amount (control edge) then a lot further, then a small amount again to finish. Now we just need to add SDS control loops like so... Select all polys and add negative thickness with Extrude, invert the Normals then fix the inevitable intersection to get something like this, which you can tweak into the shape you need.... if you find you are struggling to do that you could try adding the thickness before you do control loops, which will simplify things a bit. Then you can delete all but one quadrant, and use double symmetry to do all the remaining corners, followed by SDS to get the smoothness. CBR
  15. I found this quite helpful, though granted it's a fish rig, not a quadruped... The biggest rule is to keep all the dots inside the mesh. The second one is that most component points should be moved to the rotation point of where the physical joints on the real thing would be, hopefully as indicated by your mesh in the case of quadruped legs, which are the most confusing bit. If you google some reference vids of horses running in slow-mo you should be able to see which joints are where and where they rotate from. hovering over adjustment points should tell you their names, which are also helpful in determining which dot is which joint. CBR
  16. That's excellent Might well end up doing that, as I am getting a lot of 'spine drift' as it moves along the spline, and although I can get pretty close, no controls in the IK dynamics seem to be able to fix it 100%, whereas this would. CBR
  17. That is behaving as I would expect - stuff gets fractionally dimmer the further away it gets from the light source, mitigated by the near 100% amount of luminance you have mixed in, but entirely overridden and blown out by the glow settings which are inverted and way too powerful. Try 150% inner, and 0% outer glow and a slightly bigger radius for nicer results. I would also dial down luminance and light strength. Not sure if you intend for the light to be inverse volumetric - meaning each butterfly casts a 'shadow' of light which is opposite to what a butterfly would do, glowing, luminant or not. But that might be exactly the effect you want. CBR
  18. I'd normally do that with 'ghost colliders' ; objects or shapes that don't render but still collide cloth... No idea how you'd make it fall into place then be belted though. Hopefully someone else will... CBR
  19. Oh OK, you've convinced me; that's a nice tight curve . Kinda wanna use an IK spine though so I can get the main swim motion automatically, but as long as that can bend that much where it has to (main controller will be following a spline path I imagine) that should be totally OK. Cheers Dan. CBR
  20. Hey animation whizzkids. I'm very new at all this character animation lark, and so am cheating initially and using the Character Rig to make my learning curves less vertical. I have a question about using the fish rig in the aforementioned Character Object, if anyone would be so kind as to show me the way... When I add my IK Spine component it comes in with 8 joints, which is probably not enough for what I am doing (white shark with very tight turning circle). What's the best of way of say, doubling that number, but keeping everything else about the setup the same ? Cheers CBR
  21. Excellent - always nice to hear when a plan works and days get saved, and even nicer when OP thinks to mark their thread solved ! Gold standard forum-ing there my friend CBR
  22. Thank you dude. Still got to add the raggedness round the fins, a lot of which will be handled by geo rather than sculpting, but pretty happy with the main form so far. I am starting to look at movement now. For anyone also building a great white, here is possibly the most helpful set of shots you could find in that regard, certainly for the movements where the shark is not hunting, which are surprisingly slow, graceful and languorous - very hypnotic and trance-inducing to look at when you're having a break from the modellings... CBR
  23. I don't feel qualified to answer this, but do have a sneaking suspicion it will be a priorities thing, so if it is I am posting here in the hope that whoever can tell us what the correct priorities might be and why, I for one would be most grateful, as this is something I have been trying to understand for years, especially in how these priorities work with or are different from the hierarchy / ordering of the Object Manager. I have read the help on this many times, but lack practical examples that demonstrate these things. CBR
  24. If I had to guess I would say this looks like a very hi res plane, viewed from above, with 2 stacked displacers giving the bulgy undulations. If you used UV mapping on the plane then a very fine U-grid of lines in the luminance channel might behave like that when displaced ? Unless they've used XParticles or whatever to move a whole ton of particles across the surface and then traced them, which might also work. CBR
  • Create New...