The trouble with characters is the crease values will be adjusted over various parts of the mesh - certainly when we dealt with them. Nuts and bolts would at least have predictable values. I think when we were looking at stuff we generated arrays of edge crease values or something - I certainly remember playing with it, not sure if that got used.
Obviously, a feature is a better than no feature - you can always choose to ignore it if it doesn't fit your pipeline. Something like OpenSubDiv might be the answer as long as it's present in everywhere so deformation and UV's are handled the same in each application. On 13 April 2014 21:53, Sebastien Sterling <sebastien.sterl...@gmail.com>wrote: > Working between packages is an issue, but when it comes to characters, the > fix isn't that demanding, on the other hand if you where trying to crease > 200 bolts or screws, then you would be screwed pardon the pun. > > Yes i thought it might come bundled with OpenSubdiv, that is how it seems > to work with Modo. > > OpenSubdiv isn't the fastest, but it is really good at eradicating > imperfections you get from T poles and E poles. also tries and Ngons > > > > > On 13 April 2014 20:34, Tom Kleinenberg <zagan...@gmail.com> wrote: > >> Wasn't this another patent thing? Mental Images (and now nVidia) hold >> this strange array of patents on somewhat odd things. I know that >> MentalRay's edge-rounding shader is patented. >> >> I've worked with assets from one studio coming to us converting from Max >> to Softime for Arnold. Edge creasing was horrible because it didn't carry >> over in the models. Obviously if you have end-to-end control of the >> pipeline it's not too bad but any transfer between packages (Zbrush, >> 3DCoat, Mari) gives me the heebie jeebies. >> >> >> On 13 April 2014 21:04, Sebastien Sterling >> <sebastien.sterl...@gmail.com>wrote: >> >>> Cheers Steven, i'll definitely consider checking it out, might there be >>> a particular reason that Soft edge creasing is not supported ? >>> >>> >>> On 13 April 2014 19:35, Steven Caron <car...@gmail.com> wrote: >>> >>>> arnold supports 'hard creases' but not 'soft creases' (weighted >>>> crease)... download it and test it yourself. >>>> >>>> >>>> On Sun, Apr 13, 2014 at 11:29 AM, Sebastien Sterling < >>>> sebastien.sterl...@gmail.com> wrote: >>>> >>>>> Like many other i'm sure, i recently watched the Nvidia Pixar GPU >>>>> conference. >>>>> >>>>> http://www.ustream.tv/recorded/45386636 >>>>> >>>>> >>>>> It gave unprecedented incite into the Pixar workflow, now there was a >>>>> lot of cool stuff on display, but what interested me in this instance, was >>>>> there methodology for creasing: indicating the hardness of an edge to the >>>>> rendrerer without having to add extra geometry to hold said information, >>>>> thus optimizing the polycount, and making life easier for modelers riggers >>>>> CPU and GPU's alike. >>>>> >>>>> Creasing has always been a bit of a fuzzy area in my experience, >>>>> rarely have i seen it implemented or used in production, which is ironic >>>>> given the advantages it seems to offer. >>>>> >>>>> Now most DCC's support it in the viewport, it is available in Maya Max >>>>> Softimage Modo, and will render in mental ray as a default. however during >>>>> last years trasition to arnold, when inquiering about creasing, i was told >>>>> by a shader artist that Arnold does not support Edge creasing which i >>>>> found >>>>> perplexing. >>>>> >>>>> (not sure if it was just Mtoa) Can anyone confirm this ? >>>>> >>>>> >>>>> Personally i think Edge creasing and even vertex creasing are great, i >>>>> can't imagine how frustrating it must be for a rigger to have to deal with >>>>> 2 to 3 edge loops to control the creasing of a lip or eyelid during >>>>> blendshapes when the same result can be achieved by simply selecting 1 >>>>> edge >>>>> and indicating the crease angle. >>>>> >>>>> Have any of the beta testers for Redshift come across this ? >>>>> >>>>> I'm working on a character at the moment, and i'd like to know if i >>>>> should put in the extra detail or add creasing, given i might not be the >>>>> one to rig it. >>>>> >>>> >>>> >>> >> >
