View Issue Details Jump to Notes ] Print ]
IDProjectCategoryView StatusDate SubmittedLast Update
0006213ParaView(No Category)public2008-01-03 15:552009-05-13 13:49
ReporterDavid Karelitz 
Assigned ToKen Moreland 
PrioritynormalSeverityfeatureReproducibilityalways
StatusclosedResolutionwon't fix 
PlatformOSOS Version
Product Version 
Target VersionSomedayFixed in Version 
Summary0006213: (MARS-normal) Fix lighting when scaling by a negative factor
DescriptionWhen an actor is scaled by a negative value, the normals look wrong. Some polygons are front-facing and others are back-facing leading to lighting issues.
TagsNo tags attached.
Project
Topic Name
Type
Attached Files

 Relationships
parent of 0004786closedUtkarsh Ayachit Reflect filter does not work in parallel 
parent of 0006812closedUtkarsh Ayachit Reflect filter does not work with multiblock data 

  Notes
(0011187)
Ken Moreland (manager)
2008-04-08 17:53

Workaround: Apply the transform filter on the polygons and then re-generate normals (if necessary).

I'm not sure if there is another reasonable way around the problem.
(0011202)
David Karelitz (reporter)
2008-04-09 09:48

This was with multiblock data from a CTH dataset, which the transform filter doesn't support.
(0011264)
Ken Moreland (manager)
2008-04-10 16:26

To render something, ParaView is extracting some polygonal data from the CTH dataset. Run a filter to extract the data yourself, apply the transform filter to that, and then regenerate the normals.
(0011332)
Berk Geveci (administrator)
2008-04-15 11:59

There is no easy way to get around this. VTK would probably have to generate new normals based on the transform, I think. The best thing to do is to find another solution. What is the use case?
(0011334)
David Karelitz (reporter)
2008-04-15 12:05

The use case is that a simulation was mirrored in one axis from an experiment and we want the simulation view to match the experiment view
(0011360)
Ken Moreland (manager)
2008-04-15 16:56

Sounds like you should be using the reflect filter. However, there are known issues with the reflect filter in parallel (bug 0004786) and there are probably issues with multiblock data. If these issues were fixed, would that satisfy your requirements?
(0011362)
David Karelitz (reporter)
2008-04-15 17:01

That would work.
(0011367)
Ken Moreland (manager)
2008-04-15 17:43

There is no practical solution to this problem, so therefore we will not directly fix it. Instead, use the reflect filter instead of the transform filter and the normals will behave correctly. Once the two children bugs are fixed, the given use case will be solved.
(0011369)
Alan Scott (manager)
2008-04-15 17:46

Replaced by two child bugs.

 Issue History
Date Modified Username Field Change
2008-01-03 15:55 David Karelitz New Issue
2008-03-07 07:56 Berk Geveci Target Version => MARS
2008-03-07 08:03 Berk Geveci Category Development => 3.4
2008-04-08 17:53 Ken Moreland Note Added: 0011187
2008-04-09 09:48 David Karelitz Note Added: 0011202
2008-04-10 16:26 Ken Moreland Note Added: 0011264
2008-04-15 11:59 Berk Geveci Note Added: 0011332
2008-04-15 11:59 Berk Geveci Status backlog => @20@
2008-04-15 11:59 Berk Geveci Category 3.4 => Someday
2008-04-15 12:05 David Karelitz Note Added: 0011334
2008-04-15 16:56 Ken Moreland Note Added: 0011360
2008-04-15 17:01 David Karelitz Note Added: 0011362
2008-04-15 17:41 Ken Moreland Relationship added parent of 0004786
2008-04-15 17:42 Ken Moreland Relationship added parent of 0006812
2008-04-15 17:43 Ken Moreland Status @20@ => @80@
2008-04-15 17:43 Ken Moreland Resolution open => won't fix
2008-04-15 17:43 Ken Moreland Assigned To => Ken Moreland
2008-04-15 17:43 Ken Moreland Note Added: 0011367
2008-04-15 17:46 Alan Scott Status @80@ => closed
2008-04-15 17:46 Alan Scott Note Added: 0011369
2009-05-13 13:49 Utkarsh Ayachit Target Version MARS => Someday
2011-06-16 13:10 Zack Galbreath Category => (No Category)


Copyright © 2000 - 2018 MantisBT Team