[Bf-committers] Bf-committers Digest, Vol 164, Issue 1

Martin Norris martinjohnedwardnorris at gmail.com
Thu Sep 21 18:20:44 CEST 2017


Hi Zachary

​Coincident surfaces are always going to be a problem because the camera
only 'sees' the nearest facing surface, but since there are several
surfaces that are coincident, which of the surfaces is coincident is a
matter of chance.  You can imagine the problem as trying to sort a list by
value and return the first value.  If the list is {3, 3.0, 6/2} which is
the first object?  And what about for {3**(0.5), 3.0, 3}?

Just as a FYI, looking back at the blenderartists thread there is a big
difference between the left and right situations since the objects on the
left are different from the objects on the right​.  So when the camera
'looks' at the objects on the right the order can be different from the
order of the objects on the left.

To make the order of coincident faces deterministic, there is going to have
to be extra information that can be used to make the order deterministic.
The obvious thing to add is the _time_ that the surfaces become incident.
Because with a list that has {3 @ t1, 3.0 @ t2, 6/2 @ t3} there is a
deterministic way to sort the list, even if the values are all the same, by
comparing t1, t2, and t3.

I do not know freestyle internals so maybe Tamiko Kajiyama can comment on
any way of adding extra information to objects so that freestyle can
resolve coincident surfaces.  Maybe looking at a custom property for
'stacking'?

Best regards

Martin




> Date: Thu, 21 Sep 2017 06:03:30 +0000
> From: A Z <poweruserm at live.com.au>
> To: "bf-committers at blender.org" <bf-committers at blender.org>
> Subject: [Bf-committers] Question about feature update for Blender 3D.
> Message-ID:
>         <ME1PR01MB0835F1CCCB1BD1BC48A88C4B9A660 at ME1PR01MB0835.
> ausprd01.prod.outlook.com>
>
> Content-Type: text/plain; charset="iso-8859-1"
>
> To who it may concern,
>
>
> I'm not sure who should get this to start, however I've tried this email
> address
>
> as a starting point that developers do see.
>
>
> I've got a modelling issue that touches on meshes, NURBS surfaces and
>
> freestyle line edges.
>
>
> The entire thing is visible at the artists forum,
>
>
> https://blenderartists.org/forum/showthread.php?402757-
> How-do-I-solve-this-Freestyle-Lines-problem&highlight=
>
> How do I solve this Freestyle Lines problem?<https://
> blenderartists.org/forum/showthread.php?402757-How-do-
> I-solve-this-Freestyle-Lines-problem&highlight=>
> blenderartists.org
> -In this example, one lines group appears and the other doesn't, despite
> the fact that both situations are parallel identical. One of these problems
> begins at frame 380. I should certainly get a symmetrical, consistent
> display of lines in this situation. How can I prevent rendering doing my
> Freestyle Edges so inconsistently?
>
>
> I need some kind of way, be it via an enhanced layers system or otherwise,
>
> to allow a way to have surfaces, associating strictly with their freestyle
> line edges,
>
> to "stack" in the subsequent order they are introduced.  By introduced,
>
> I mean brought into surface coincidence. I then need a camera to
>
> render and interpret these coincident surfaces as "stacking",
>
> top down, even though they are coincident.
>
>
> I need this feature so that I can use blender keyframe animation to
> simulate a paper
>
> folding situation, where I may "virtualise" away any paper thickness.
>
> I need only use one surface reference on all surfaces, and while my model
>
> is only a mesh one, perhaps in wider theory it implies to NURBS surfaces
>
> too.
>
>
> Is there anything going on towards version 2.8, or anyone really, who could
>
> take a look at my feature request here and perhaps do something about
> accomplishing
>
> this?  Or even pointing me in the correct Blender direction if I'm wrong?
>


More information about the Bf-committers mailing list