PDA

View Full Version : Multishift slides some edges exaggeratedly



probiner
07-17-2013, 04:59 AM
I think I kind of nailed the Multishift bug that has been bothering me many times. Fogged (https://fogbugz.newtek.com/default.asp?63286_v0o549qp190rigma), this is just a community notification.


https://www.youtube.com/watch?v=uCEnKw2CSak

Cheers

Tobian
07-17-2013, 07:55 AM
Yup, it's often completely unavoidable to get micro non-flatnes, when using things like boolean + merge threshold to fix the points, and so I get problems like this all the time. Thanks for hitting the nail on the head as to what was causing it!

probiner
07-18-2013, 07:51 AM
It's strange because having them all non-planar along the cut doesn't seem to break. I think it's the mix that screws it.

Object attached.

This issue is for Shifts only. You and me will still have the issues with the Insets not respecting both the inner edges and an offset contour distance and not just one.

Cheers

JoePoe
07-18-2013, 08:29 AM
So, are you saying this is new behavior for multishift in 11.5.1? (disclaimer.... which I am not in).

Wouldn't the primary suspect be the newer tool... Flatten. Perhaps it is doing something..... funky.

Do you still have the same multishift behavior when manually flattening with the stretch tool? (I know, possibly not as precise... but just for experiment sake).

probiner
07-18-2013, 08:34 AM
So, are you saying this is new behavior for multishift in 11.5.1? (disclaimer.... which I am not in).
Not 11.5.1 related


Wouldn't the primary suspect be the newer tool... Flatten. Perhaps it is doing something..... funky.

Do you still have the same multishift behavior when manually flattening with the stretch tool? (I know, possibly not as precise... but just for experiment sake).
Stretch tool does the same.

Make a sphere, stretch a group of polgons on the X axis. Select all polygons but one. Multhishift and Shift the thing. Sliding hell :)

Cheers

JoePoe
07-18-2013, 08:43 AM
Hmmm, I've only tried your box example in 9.6.1 using stretch to flatten. No problems here.

Will try the sphere as soon as I get back to computer.

Tobian
07-18-2013, 09:17 AM
I've noticed this issue since they introduced Multishift, it's just really hard to replicate and only very rarely does it. It's good that probiner is finally nailing down the cause!

probiner
07-18-2013, 09:27 AM
Hmmm, I've only tried your box example in 9.6.1 using stretch to flatten. No problems here.

Will try the sphere as soon as I get back to computer.

Just followed the steps in the video and replaced Flatten with Stretch same issue.

Cheers

JoePoe
07-18-2013, 10:08 AM
Interesting!
So I guess we are dealing with a change somewhere between 10-11.5.

For what it's worth, 9.6.1 and all, the sphere example checks out okay also.

probiner
07-18-2013, 04:21 PM
Tried the stuff I did in the video in 9.6 and at least for me it happens.

cheers

JoePoe
07-18-2013, 06:57 PM
Very VERY interesting.
I know you believe me.... just posting anyway.

I must be lucky. Time to play the lottery :).

probiner
07-19-2013, 01:39 AM
I believe you :D Try the attached file.

I must admit the cube some times gives me no issues in 9.6 or it can even be "fixed" by changing the 'Shift' to 'Average Point Normal' and then back to 'Contour'. But the sphere, no solution.

Cheers

Lewis
07-19-2013, 07:26 AM
Very VERY interesting.
I know you believe me.... just posting anyway.

I must be lucky. Time to play the lottery :).

You changed Shift settings to Contour2, try to match probiner settings.

BTW yes this is bugging me most of time and i really use multishift a lot.

JoePoe
07-19-2013, 08:54 AM
@ Lewis :) I knew someone was going to point that out. The box seems to work, .... for me, with ALL inset and shift types. (posted Probiner's box with Contour below).

@ Probiner .... Re: the sphere....Now that is some funky geometry for a test. Results below. Contour 2 does give a sort of decent "result", Contour (1) is worse, but neither cause the super fly away action we are discussing. Point Normal shift (which I guess is, for all intents and purposes, pretty close to Smoothshift) works great.
BUT, all shifts happen smoothly with the mouse action. The "results" aren't pretty, but they are (again for me, for whatever reason) doing what, I think, they are supposed to do (i.e some points do start to get a little unruly as the shift gets bigger)... After all, that is some crazy geometry to mutishift.

And for everyone out there. I am NOT denying that there is a problem. I'm simply trying to help trouble shoot. :)

probiner
09-25-2013, 03:40 PM
Just filled one for Thicken too: https://fogbugz.newtek.com/default.asp?65877_li32g6vp

https://dl.dropboxusercontent.com/u/143766132/Forums/NewTek/Thicken_Bug.png

Cheers

Tobian
09-25-2013, 04:17 PM
Yup, even breaks if you subdivide the mesh once, which should sort out all the non-planars...

probiner
10-11-2013, 02:14 PM
Ah! this one is a head-scratcher: https://dl.dropboxusercontent.com/u/143766132/Forums/NewTek/Thicken_Bug_02.lwo

So simple and Thicken>Contour fails :O

JoePoe
10-11-2013, 02:39 PM
Funny that this thread pops up again.
This just happened to me for the first time...
0 non-planars in this case.

117619

probiner
10-31-2013, 04:41 PM
Sad to say that none of these were fixed.

MS case was closed, due to supposed fix, but checked this latest release and still there.
https://fogbugz.newtek.com/default.asp?pg=pgPublicView&sTicket=66850_kh25qa0h

Cheers

Tobian
10-31-2013, 04:46 PM
*sigh*

Lewis
11-01-2013, 06:19 AM
Repoent the case (just reply to case and it'll upodate and open it automatically)