Jump to content


Photo

axial sweep issue


  • Please log in to reply
5 replies to this topic

#1 3dworks

3dworks

    Advanced Member

  • Members
  • PipPipPip
  • 388 posts
  • LocationBerlin, Germany

Posted 08 February 2018 - 06:44 PM

sometimes it's like climbing up a glass wall, FZ seems to refuse to work as expected. not sure if it is a bug or if i am doing something wrong. in that case, please make it 'artist friendly'.

 

see 2 screenshots. both are surfaces, using axial sweep with pretty standard settings and the source set as centered. result is...not as expected. as you can see, the right side of the path poly has 2 closer points, but i don't understand why the result of the sweep is unusable. i used the object doctor on the path object before sweeping.

 

any hints? is it a bug? can this behaviour be fixed? in my long experience with FZ it is not the only case where construction and result of simple sweeping operations just don't match...

Attached Thumbnails

  • screenshot_20180209_axial_sweep_issue.fmz - 3.png
  • screenshot_20180209_axial_sweep_issue.fmz - 2.png

Attached Files


Edited by 3dworks, 08 February 2018 - 06:45 PM.

3dworks visual computing + photography


#2 pylon

pylon

    Moderator

  • Moderators
  • 795 posts
  • LocationLos Angeles

Posted 08 February 2018 - 08:57 PM

Hi Markus,

 

You have some duplicate points, slightly ajar:

 

Attached Thumbnails

  • axial_sweep_issue_fmz_-_1.png


#3 3dworks

3dworks

    Advanced Member

  • Members
  • PipPipPip
  • 388 posts
  • LocationBerlin, Germany

Posted 09 February 2018 - 04:39 AM

ok, then the question is: why doesn't object doctor find them? why does the tool not include some tolerance settings? since we cannot construct anymore precisely in polygonal / facetted way (this mode was killed with version 7) it is difficult to control point placement. since then, i am always plagued with smooth to facetted conversion problems. honestly, i'm getting into this kind of troubles quite frequently and imo it shouldn't be such an issue. additionally, when you switch the view mode to show points, FZ slows down to a crawl when you are working with larger models. all in all i wish FZ would get beck to be a program which enables people to model fast, precisely and reliably (without so many tolerance induced errors and without so many crashes), which was more in version 6.x than in 8.x - unfortunately.


3dworks visual computing + photography


#4 3dworks

3dworks

    Advanced Member

  • Members
  • PipPipPip
  • 388 posts
  • LocationBerlin, Germany

Posted 09 February 2018 - 05:16 AM

just as an explanation to how the 'dirty' shape result was achieved: it is part of an imported 2D DWG which i got from a producer. i traced those 2D lines and obvioulsy there was some precision issues when snapping to the existing drawing. this is a standard situation when dealing with reconstructions based on 2D plans. i forgot to say that i also run the 'reduce mesh' tool on this shape, with angles of 2 degrees set for both edge and face. which we could get better analysis / fix tools which can spot such problematic near points.


Edited by 3dworks, 09 February 2018 - 05:16 AM.

3dworks visual computing + photography


#5 Bo Atkinson

Bo Atkinson

    Advanced Member

  • Members
  • PipPipPip
  • 175 posts
  • LocationMaine USA

Posted 12 February 2018 - 07:31 AM

I too have had similar problems,  through many formZ versions. Despite explanations given, my understanding was often not achieved. Instead, my solution was always to try every other conceivable method, to build from scratch… In a case like this, usually first try ‘facetting’ the curve with equally sized segments, (sometimes toggling odd vs even number of segments for effect). This used to be achieved with the “zero radius helix”, but now the Polygonize tool does this well. If sweeping did not work, sometimes NURBS  lofting could do it right, sometimes rearranging the positions of the “multiple entities” could fix bad results.  Sometimes even lofting “derivatives of the paths” rather than the cross sections, (kinda). Additional ideas would arise if all else failed and if time permitted… 

 
Personally i take it the math has a peculiar dependence on startigng and ending points, instead of simply disregarding the concepts of starts and ends, of points. Just let closed paths enjoy all points as equals, so to speak.


#6 3dworks

3dworks

    Advanced Member

  • Members
  • PipPipPip
  • 388 posts
  • LocationBerlin, Germany

Posted 12 February 2018 - 03:32 PM

also useful would be to have an integrated 'tolerance' slider which would make the tool ignore points which are closer than a set distance.


  • Des likes this

3dworks visual computing + photography





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users