PDA

View Full Version : Arrrgh, hope this doesn't happen in 8 too....



hrgiger
04-27-2004, 06:46 PM
Ok, so I'm dealing with some UV's...

I unweld my points, adjust my UV map, and then reweld and all my points don't weld back together.

This is ultra ridiculously annoying.

I know for a fact that I have moved nary a point except for in my UV map (which shouldn't affect my base model).

I've also selected two of the points that should merge and checked them in the info panel and they have the exact same x,y, and z coordiantes. Why won't these bastards merge? Now I have to waste an hour going around my model and manually merging two points at a time. I've tried using the fixed distance but no amount I come up with corrects the problem and sometimes merges points I don't want.

Anyone else ever have this problem?

prospector
04-27-2004, 06:59 PM
yep.. in 7.5c

info always say they are exactly the same in 3D space.

You are doing the only thing I have found to fix them......individually weld.:mad: :mad: :mad:

I too hope it's fixed

Chris S. (Fez)
04-27-2004, 10:43 PM
Hey HR. Do you have endomorphs? Delete any and all endomorphs then try to merge. Does it work? If it does then you probably have one or more "problem" endomorphs which will not allow you to merge. I'm curious if this is the problem or if it something else I am unfamiliar with so please let me know.

prospector
04-27-2004, 11:42 PM
Not me.

I usually do all UVing before any kind of motion stuff is added
(Morphs, bones)

But why should a morph be causing 2 points in the exact same spot not merge?

Chris S. (Fez)
04-28-2004, 07:31 AM
"But why should a morph be causing 2 points in the exact same spot not merge?"

It shouldn't but sometimes points won't weld where there is an endomorph deforming the mesh. I have had this happen a few times. I deleted the offending endomorph(s) and it merged just fine afterwards.

Dodgy
04-28-2004, 08:48 AM
I can back up the morph thing. Definitely a bug somewhere in there.

evenflcw
04-28-2004, 10:52 AM
The morph thing is definatly not a bug and makes perfect sense.

Think of endomorphs as instances of the original mesh (they pretty much are, just stored in a clever way). Consider then that you have a base mesh with two points placed at the exact same location (so they would automerge in normal cases), consider then also that you have a morph where those same two points are placed at two different location at an offset from eachother. If those two points would automerge even in the morph you would loose a unique point and it's placement.

Points will only automerge if they are in the same place across all morphs/instances. If a single morph has them at an offset, they won't merge (because that would result in loss of data).
So if you have a couple of points that refuse to merge, simply select them and reset their vmap value in all endomorphs so they loose their endomorph assignment and snap to the same location they have in the base.

Hrgiger, prospector, if you want I could make a script for you that will automate th process of manually welding points. It would pretty much do what your doing manually using a fixed distance setting like Merge Points.

EDIT:
I just tried something to sortof test the theory. If you have points placed at an offset in a morph and use MergePoints with fixed distance set to something higher than that offset, those points will merge.

Chris S. (Fez)
04-28-2004, 11:24 AM
"consider then also that you have a morph where those same two points are placed at two different location at an offset from eachother."

What I don't understand is that these points are often NOT offset in the morph. If the points are welded in both the base and the endomorph, why should they not weld with a simple merge after you unweld?

evenflcw
04-28-2004, 11:34 AM
That part I don't know :) cause I feel they should too. Luckily I've never had this happen.

Only thing I can think of is that LW/the CPU for some reason screws up a decimal or something. For example, sometimes zeros might turn up negative (with a minus infront of them). Perhaps one instance of these points have one of those minuses in the endomorph data and LW doesn't see their locations as equal... grabbing at straws here :)

If anyone has an LWO with this problem present, I'd love to take a look. No need to include a whole mesh, just a part with problems.

hrgiger
04-29-2004, 02:32 PM
Sorry, I hadn't checked on this thread in a few days....

It could be the morphs. I've been creating morphs to flatten my mesh to make UV's from that. I'll try deleting the morphs and see if that works.

hrgiger
04-29-2004, 02:41 PM
Ok, I just deleted the morphs and tried merge. It worked. Thanks Chris Fez! Today, you rock. Tomorrow, I could be disagreeing with you about something so....:D

Now that the merge problem is fixed, I can finish UV'ing...

Dodgy
04-29-2004, 03:09 PM
Originally posted by evenflcw
The morph thing is definatly not a bug and makes perfect sense.

Think of endomorphs as instances of the original mesh (they pretty much are, just stored in a clever way). Consider then that you have a base mesh with two points placed at the exact same location (so they would automerge in normal cases), consider then also that you have a morph where those same two points are placed at two different location at an offset from eachother. If those two points would automerge even in the morph you would loose a unique point and it's placement.



That doesn't make sense to me, in as much as I under stand how this happens, but not why. You want to merge these two points, so therefore you WANT to lose one's morph data as well. Shouldn't you just be able to weld them if you want to, rather than going through every morph target and removing the morph value for that point? The welded to point should be the one to retain it's morph map values. Admittedly, having them both stay discontinuous in UVs and colour maps makes sense, but not for morphs, and maybe not for weights too. Maybe weights should get the average weight of the two points, or retain the weight from the welded to point, but not have discontinuity at that point. Try it, and you'll see a discontinuous weight map for bones should perhaps be avoided. The behaviour for morphs is not justified, IMHO and maybe not for weights either.