PDA

View Full Version : bevel kills symmetry ?



madno
10-17-2015, 03:24 AM
Hi,
have an object -> made a bevel -> symmetry does not work anymore:

130377

Tried with lw bevel and lwcad mass offset.
LW 2015.3

It works with a box, but making only boxes gets boring.

Is this normal? Can I do someting to get symmetry back (tried increasing the sym value in General Options -> did not help).

:help: please

ernpchan
10-17-2015, 07:56 AM
There are a bunch of plugins that fix symmetry on here. I'm sure all, if not at least one of them, will fix your model.
https://www.lightwave3d.com/assets/plugins/contains/Symmetry%20/

JoePoe
10-17-2015, 09:02 AM
Are you sure everything is copacetic with this model? Can you post it?
Bevel keeps symmetry for me (2015.3). As does Multishift....

To fix use what Ernest posted..... or do the 'ol manual delete-half-and-mirror routine.
(if going the manual route, make sure that the centerline is zeroed out before the mirror!!)

madno
10-17-2015, 09:19 AM
Many thanks ernpchan,

px_Symmetry_Stuff did the trick.
Seems so symmetry has been an issue before I encountered it.

Joe,

normally I am sure that I did something wrong with the geometry (see my forum status "geomessy").
Layer 1 orig before bevel
Layer 2 after bevel without fix
Layer 3 after bevel and fix with the plugin (need to merge the polygons).

JoePoe
10-17-2015, 09:40 AM
That's really really weird.
I symmetrized the model first and then did the bevel and symmetry remained intact post bevel.

But I can't figure out why the original won't accept it....

(same thing in 11.6.3 fwiw)

Edit: AND still happens when it's just those two polys..... and Info statistics shows exactly the correct coordinates for perfect symmetry. WEIRD.

madno
10-17-2015, 10:10 AM
Maybe a Modeler miracle.
For a non pro like me this is really a challange. I try something -> get strange results -> start to test -> loose time -> think I am too stupid -> most of the time that is true. But sometimes LW just seem to be made to make people become insane. But to not let this sound like LW baching; I read a Modo 901 thread - seems so its users' life is hard as well :-)

JoePoe
10-17-2015, 10:21 AM
I'm gonna lose sleep over this one... I can feel it :D.

Everything is "fine" if you triple the original, so I thought it had to do with non-planers. But nooooooo. 8~

jeric_synergy
10-17-2015, 01:30 PM
:: popcorn ::

spherical
10-17-2015, 01:53 PM
I LOVE threads like this! :D

rustythe1
10-17-2015, 01:59 PM
when this happens to me (a lot) I just use the knife tool straight down on the middle, delete one side, mirror then band glue it back

Oedo 808
10-17-2015, 02:03 PM
Seems the polygon on the left is degenerate in some way, running Normal Corrector (https://www.lightwave3d.com/assets/plugins/entry/normal-corrector/) fixes it up. I seem to get this quite a lot, even on planar polygons, an object might not round properly or some such thing, running this says it fixes x amount of polygons and then it seems ok. Sometimes you might see the shading of a polygon change even though there are no obvious defects like coexistent vertexes, run this and it'll turn back to normal and behave normally.

stiff paper
10-17-2015, 03:19 PM
Seems the polygon on the left is degenerate in some way, running Normal Corrector fixes it up.
You could probably just select the funky poly, do a SHIFT T then a SHIFT Z to achieve the same result without a plugin.

Maaaaaaaybe it's a point order problem?

Oedo 808
10-17-2015, 10:47 PM
Yeah that did work, but if you're going to fix one you might as well clean up all potential problem polygons, but certainly you could triple and merge the tris, it might work for all.

It could be a point order problem but I wouldn't have a clue... although now you've mentioned it using mesh reknitter should fix it, might try that. Does welding and unwelding change the point order? I just can't remember how things work when I'm not in Modeler, in fact I probably wouldn't even have to test it, if I were sat in front of it now I'd probably say either it does or it doesn't, but when I'm away I forget how half the things work. :stumped:

madno
10-18-2015, 12:51 AM
Oedo 808,
Normal Corrector said it fixed 138 polygons (model has changed since I posted it).
Seems so Symmetry Stuff and Normal Corrector are the plugins to go. I just which LW3DG will at some point in time add such features as native tools.

pinkmouse
10-18-2015, 05:17 AM
I've never had much luck with Symmetry. It hates me.

JoePoe
10-18-2015, 07:41 AM
So the normals of polys showing in statistics as planar need to be fixed?? Super.... :screwy:...:grumpy:.

stiff paper
10-18-2015, 07:45 AM
Does welding and unwelding change the point order?
Tripling then merging forces Modeler to reconstruct/make a new quad, so I don't think unweld/weld will work. The polys will still all be the same. Mesh Reknitter might work. That is, if Mesh Reknitter itself works in newer versions of modeler... and it's 4874 years old now, so it's bound to be x32 only...

If I remember rightly, all this kind of wonkiness was introduced at the same time in LW version... nnnng... hmm... I'm going to say... 6. But I can't really remember. It was whichever version it was where suddenly you noticed your models had unexpected 2 point polys when it had never done that before. I think. Or maybe I just imagined it.


I've never had much luck with Symmetry. It hates me.
It hates everybody. Don't go thinking you're a special snowflake.

Edit: I just tried messing about with point order on otherwise symmetrical polys. Nope. It isn't a point order issue. Somebody should submit this as a bug, because it's definitely a strange thing that's happening.

madno
10-18-2015, 08:52 AM
Ok,
report sent.

Oedo 808
10-18-2015, 08:55 AM
Yeah I tried that 3DD MeshReknitter plug-in and it didn't work. You're right it is 4874 years old, Fujio Ichikawa has a ton of plugins and a good portion are x64, alas not that one however. Interestingly enough Polygon Corrector still assess that a fair few normals need correcting even after tripling and re-merging with MergeTrigons_X.

And yes symmetry hates everyone, but look on the bright side, instead of dreaming of multi-axis radial symmetry with painful longing, you're just dreaming of x-axis symmetry that works... that's a good thing right? :stumped: