[Bf-committers] No more subsurf and mirror modifier with scupt mode?

Nathan Vegdahl cessen at cessen.com
Fri Apr 22 08:57:55 CEST 2011


Just to toss in a quick word from a rigger's perspective.  It sounds
like most of the complaints are coming from the modeling side, where
features have apparently been lost.  But from the rigging side an
amazing new tool has been gained: being able to sculpt shape keys on a
deformed character mesh.  This is absolutely invaluable for creating
corrective shape keys, and I lament that I did not have this tool
while working on Sintel.  I cannot thank Sergey enough for this new
feature.

So there is a reason for the change in sculpt's behavior.  But indeed,
it would be cool if there was some way to have both feature sets.  Of
course, you can't _actually_ sculpt on a subdivision mesh, since most
of the vertices don't exist as such to be sculpted on.  But a decent
illusion thereof is very useful, I'm sure.

Sergey:
FYI, in production, the final modifier stack for a character generally
looks something like this:
1. [Mirror]
2. Armature
3. Subdivision
4-n. Other stuff

Usually there is no mirror modifier by the time it gets to rigging,
due to the needs of texturing.  Though there are occasional
exceptions.  I think it is generally safe to assume for
sculpt-on-armature-deform that the armature modifier comes first in
the stack.  As a rigger, I can always temporarily disable modifiers
that come before the armature deform when working, if needed.

I also assume there is no armature modifier in the stack when modelers
are doing sculpt work.

I don't know if that helps solve things or not, but just in case that
information is useful.

(Incidentally, this is all going to become quite difficult if/when we
get modifier nodes.  Already I do not expect this to be supported with
multi-modifier armature deform.  Maybe some way of selecting which
modifier you are sculpting with...)

--Nathan


On Thu, Apr 21, 2011 at 2:54 PM, Ronan Zeegers <blender at ronanzeegers.com> wrote:
> Hi Sergey,
>
> Thank to listen to my opinion! I hope it will have his weight between
> all other one ;)
> Like you said, maybe the best solution would be a slider (to give the %
> of  alpha level) or an option to be able to view the constructive
> modifiers and/or hide the base mesh.
>
> About combinations of modifiers, I remember playing with Lattice and
> Armature modifier + Sculpt in Blender 2.4x. But it was in exceptionnal
> cases.
> For me, the most commun cases are Subsuf and Subsurf+Mirror modifier.
> I'll come back to you If I find other.
>
> cheers,
>
> Ronan Zeegers
> /*Postprod 2D/3D*
> + 32 (0) 473 45 20 43
> www.ronanzeegers.com /
>
>
>
> Le 21/04/2011 13:42, Sergey I. Sharybin a écrit :
>>    Ronan Zeegers wrote:
>>> Hello Sergey,
>>>
>>> Why not keeping the old approach for the subsurf and mirror modifier?
>>> Like you said, at least supporting some simple cases.
>>> The intuitiveness of this approach seems to be subjective.
>> Current approach was introduced becaue plenty of artists missed
>> predictable sculpting on armatured mesh and to implement this i had to
>> disable old behaviour (otherwise, things can't be predictable enough and
>> both of implementation/working as user was quite difficult task)
>>> To defend the old behavior, I think that a 3D artist know that he is
>>> scuplting/moving vertex of the base mesh. Not the "virtual" vertex of
>>> the subsurf/mirrored/displaced mesh.
>>> It never disturbed me to not moving the shape because I was not clicking
>>> in an area where there was vertex.
>> It's just two different cases which can't live togeter well, but current
>> implementation could be used as "basis" for easier re-implement old
>> behaviour for constructive modifier.
>>
>> Actually, i don't think it's contructive to continue discussion like
>> "things were cool, now it's not so cool" -- it's different cases and
>> returning of (at least some of) constructive modifiers is in my
>> sculpting todo list.  I'd prefer to collect as much opinions as it's
>> possible to find out which behaviour should be used by default, which
>> additional modes should be added and so on (everything, which could help
>> to make sculpting in blender useful for wide audience of artists).
>>
>> Currently, me and Tom (a.k.a Letterrip) dicussed this things and we
>> found that re-implementing my old derived-cage patch would help a lot
>> with supporting constructive modifiers. Idea is the same as it used for
>> edit mode: draw final shape solid and mesh, which is actually editing be
>> half-transparent. It wouldn't be helpful for case of deformation
>> modifiers because things are becaming much more difficult to see in the
>> screen, but should work fine for constructive modifiers and also it'll
>> help to visualize "sculpting layer" for difficult cases.
>>
>> Personally, i don't think supporting of constructive mosidiers should be
>> enabled by default -- i'd prefer to have things enabled by default if
>> their behaviour is well predictable. Maybe i'm wrong, but it'll be
>> simple to change. Also, that half-transparent derived cage could be
>> toggleable, so it could be easily hidden.
>>
>> P.S. Maybe i forgot to mention that disabling all constructive modifiers
>> gives advantage in case of mixed constructive/deformation modifiers in
>> the stack. In this case i'll see quite final shape of object (maybe
>> without vonstructed elements as mirrored part and so on), but shape
>> itself is final.
>> P.P.S. Difference from previous implementation of derived-cage patch,
>> this half-transparent part could be crated from mesh with applying all
>> leading deformation modifiers. Maybe it'll be useful. I just not sure
>> about which combinations of modifiers are actually used by artists --
>> but you could help me with it ;)
>>> cheers,
>>>
>>> Ronan Zeegers
>>> /*Postprod 2D/3D*
>>> + 32 (0) 473 45 20 43
>>> www.ronanzeegers.com /
>>>
>>>
>>> Le 21/04/2011 09:45, Sergey I. Sharybin a écrit :
>>>>      Looks like it was implemented in 2.49 exactly in the same way as it
>>>> was before enabling sculpting on deformed mesh in 2.5 and i don't find
>>>> it intuitive.
>>>>
>>>> I'm not sure what do you mean "properly" -- i can't make strokes on
>>>> mirrored part of mesh. And what about sculpting on deformed/armatured mesh?
>>>>
>>>> Problem that we can't deal with all kinds of modifier stack content and
>>>> now we allow only that modifiers, which could be handled ~100% correct.
>>>> Of course, we could support simple cases like Bse mesh ->    mirror ->
>>>> subsurf or Base mesh ->    armature, but cases like Base mesh ->    mirror ->
>>>> armature can't be handled correct. And things could be much more
>>>> complicated here and you've got no idea where stroke happens (even in
>>>> 2.49 troke isn't happening on that point of subdivided default cube --
>>>> try to grab vertex -- it's movenment would be "delayed", it's because of
>>>> distance between dragging vertex and prush posiiton).
>>>>
>>>> That's why ide of sculpt cage was burn -- just to visualize kinda
>>>> "sculpting level" which is used for brushes just to make things more
>>>> clear about where sculpting happens. Otherwise, in a bit more
>>>> complicated modifier stack you should be making strokes far from place
>>>> you want to add some displacement. It's not intuitive at all.
>>>>
>>>> Matt Ebb wrote:
>>>>> On Thu, Apr 21, 2011 at 5:07 PM, Sergey I. Sharybin<g.ulairi at gmail.com>     wrote:
>>>>>>      Hi Ronan!
>>>>>>
>>>>>> Yep, you're right -- constructive modifiers (like array, mirror,
>>>>>> subsurf,...) were disabled when sculpting. This was made to make
>>>>>> sculpting more obvious and enable sculpting on deformed mesh.
>>>>> I forget the issues involved here, but I recall sculpting (modifying
>>>>> base level mesh, as you would in edit mode) with mirror and subsurf on
>>>>> was supported properly in 2.49 - a modeller friend I've worked with
>>>>> relied on this a lot - using the sculpt tools to tweak poly modelled
>>>>> objects. What's the difference between how it worked in 2.49 and now?
>>>>> is it possible at all to restore similar functionality as 2.49?
>>>>>
>>>>> cheers
>>>>>
>>>>> Matt
>>>>> _______________________________________________
>>>>> Bf-committers mailing list
>>>>> Bf-committers at blender.org
>>>>> http://lists.blender.org/mailman/listinfo/bf-committers
>>>>>
>>> _______________________________________________
>>> Bf-committers mailing list
>>> Bf-committers at blender.org
>>> http://lists.blender.org/mailman/listinfo/bf-committers
>>>
>>
>
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers
>


More information about the Bf-committers mailing list