Jump to content
3DCoat Forums

Percevan

Member
  • Posts

    33
  • Joined

  • Last visited

Posts posted by Percevan

  1. the answer is in the question, as they say. :)

    You're missing a little step to get your intended, final, result, yes. But first allow me to humbly point out how and why it behaves as it does... which is as it should. it's not bugged or anything don't worry.

    Quote

    When I do anything with the original object, it suddenly has the part I "cut out" restored as if I had done nothing to it.  The new piece still exists on its own layer.

    Except that as your own use of quotes imply, you actually did NOT "cut out" anything from the original layer, where everything still exist, yes. So, when you did 'Objectify Hidden', a new layer with a _copy_ of the hidden was just pasted in, rather than a cut/paste equivalent, as everybody instinctively assume when first using vox hide (I swear, everyone make the same assumption, I did too at first ;) ).

    So, after you get your objectified_new_part, mouseover the original layer, press 'h' to select it easily, and then go back to the geometry menu to use 'delete hidden'. You can then check, by using 'unhide all' that it worked as intended.

    On a side not, I'm venturing a guess here... but what may be leading user to this confusion when using VoxHide, IMO, is the fact that the new objectified object(s) get the suffix name "_hidden", as is they where  cut/pasted from the original, while they're not, just copy of some still hidden part of the original, until explicite deletion by the user.

     

  2. In my experience, it works even on complex uv islands. That being said, I'd advise to recheck the topology, if you can't stack UV island by copy/pasting, it's generally because of split components, or too differing triangulation.

    In Nivellen's case, for example, I'd bet on the later. Having the same triangulation should let the usual UV stacking works, after rotating some of those differing edges (select / edge /CW or CWW spin command). Or, since it's so similar, delete one UV island, select the other and clone it (*), position the cloned retopo parts,and easily copy paste their Island. Might be faster than picking incorrectly rotated triangle edges edges and rotating them one by one.

    (*) / edit: Still in the Retopo room of course; And 'clone' "appears" in the Commands tab if you're first in Select mode, Faces.

  3. Lumberyard is a fork from CryEngine 3.8.xxx , so it's lacking a lot of the development CryDev brought in their most recent versions. And Amazon didn't bought the "right" to benefit from any upcoming CryEngine development. On the other hand, it has a relatively decent update rhythm, not very fast but... okayish, which can even bring a couple of features not yet present in CryEngine V. For example, even though I didn't dig deep into it when I briefly tried it, Lumberyard has a component approach not unlike Unity, which might help new users and speedup things... maybe. Personally, I still prefer CryEngine, but to each its own :)

    I don't have issues exporting models from maya to CryEngine V (didn't try with Lumberyard) even though, granted, you have to be careful as the whole process involves a bit more preparatory steps that I like, so the user has to be methodical and... not in a rush :).

  4. Quote

    Je travaille en 4096*4096 et pourtant les textures du terrains sont pixélisé quand on s'approche.

    La question déborde un peu (beaucoup :) ) du cadre de 3D-Coat, et concerne surtout tes buts, méthodes et moteur de jeu utilisé.

    Mais quelque soit ce dernier, en règle générale si la taille globale du terrain est imposante, même une texture 4k peut aisément donner cet effet de flou pixéliés, surtout pour un close-up caméra. Si, dans ton moteur de jeu tu n'envisages pas de couvrir le terrain avec des éléments de végétations, roches, etc. il faut soit viser une texture 8k soit utiliser plus d'une texture, si tu as la possibilité d'utiliser plus d'un terrain - ce qui n'est pas toujours faisable facilement - ça dépends, encore une fois du moteur de jeu (unreal, unity, cryEngine, autre?), de l'aproche visé ("faux" terrain réassemblé, ou streaming de terrains?), etc.

    Et si tu envisage de couvrir au moins en partie ce terrain, ne serait-ce qu'avec un système d'herbe, généralement la texture n'a pas besoin d'être trop détaillée, elle sera alors surtout visible de loin, par LOD.

  5. The UV tab lets you edit UVs from objects present in the Paint room... it was a bit confusing to me too, at first (no so long ago really).

     

    So either from a sculpt work that you retopoed, as Tony said, and then sent to the paint room after your unwrapping there with one of the bake option in the Retopo rooms' bake menu, since such operation actually creates your low poly mesh in the paint room, its surface material(s), and its UVs.

     

    That, or import an already UV'ed mesh,  of course.

     

    (Edit: some typos)

    • Like 1
  6. first, nice models AND paint job. well done! :)

     

     

    (...) any thoughts on being able to tighten the curvature would be very helpful.

    once you have set the painting condition, there are three settings (that I know of at least), that can have a big impact to get tight or wide, soft or sharp, fuzzy or clear "edge painting":

     

    the most obvious: degree, which can be set (much) higher than 100% by the way (can be useful in some cases, and such "boost" can give interesting result when choosing a "less on..." option). first setting on the right of the condition at the top (when using nomal material painting).

     

    next on the right again: 'contrast', positive or negative... the latter can be a better choice in some cases, experiment!

     

    and third,  in the smart material editor, you also have "edge scattering", which can be filtered with a texture, or noise. this one, combined with 'contrast' and 'degree', can give interesting variations.

     

    that's the basics, I'd say, but they already offer subtle range of options, depending of what you're trying to get. After that, smart marterial layering, masking, etc, extent the possibilities even more.

  7. Hi,

    Take the following with a big grain of salt since I haven't been deeply delving into 3D-coat In's and outs  for a long time myself, but:

    1. generally yes, though If you don't fully fill a big mesh but rather paint some part(s) with other painting tools, that's not necessarily an issue. And even if you do (fill tool, big retopogroup object)... well, in most case even non-tilable textures can be useful as inputs in smart materials to fill objects, since they can easily be hidden either manually afterwards or from other layers of the same material.
    here's one of my first experiments with 3Dcoat's SM, 3 layers (at the time), 2 textures, and the sandy ones used in the wavy grooves (concave parts) of the mesh is definitely non-tileable. I didn't even have to repaint on seams thanks to edge scattering and some noise in the material settings.
     
    2015_10_29_000750.jpg
     
    2. I would recommend the opposite actually. I don't see how auto-mapping would give any benefits, except for experimenting and self-learning Smart Materials, imho.

  8. Just building up some personal experience with 3DC voxel tools, in the present case with Cut out and Vox Hide mainly. Simple but very fun to do. Here's a test concept with  rampart/battlement wall segment (and I just put a smart material on it for visualisation purposes, nothing final). This will probalby be used to make static meshes for a Skyrim Mod (Hoddminir, for anyone interested ;) )

    wall_v3.jpg

    • Like 1
×
×
  • Create New...