[Paraview] Bug in Color Scale

Takuya OSHIMA oshima at eng.niigata-u.ac.jp
Tue Dec 2 09:19:40 EST 2008


I appreciate your response. I have just sent a query about the values
repoted in the Information panel to the user. Meanwhile I skimmed
through pqPipelineRepresentation and its relevant classes but haven't
found anything suspicious so far.

Takuya OSHIMA, Ph.D.
Faculty of Engineering, Niigata University
8050 Ikarashi-Ninocho, Nishi-ku, Niigata, 950-2181, JAPAN

From: "Moreland, Kenneth" <kmorel at sandia.gov>
Subject: Re: [Paraview] Bug in Color Scale
Date: Mon, 1 Dec 2008 15:44:31 -0700

> I’m running out of things to suggest.  The only problem I can think of (and
> this is a stretch) is that somewhere the range in the data array is being
> incorrectly set either by some rouge call to the information object or because
> the data in the array changes without the vtkDataArray object being aware of
> it.  Does anyone else have any ideas?  Berk?  Utkarsh?
> 
> What range of values is reported in the Information panel?  Is it giving the
> correct range, or the range used in the color scale?
> 
> -Ken
> 
> On 11/28/08 7:46 AM, "Takuya OSHIMA" <oshima at eng.niigata-u.ac.jp> wrote:
> 
>     Hi Shi, Ken, and all,
>    
>     A user of my reader reported exactly the same problem about the auto
>     scaling on ParaView 3.5.0. He told me that after applying Rescale to
>     Data Range to scalar attributes the range is always set to [0, 1],
>     even though the data actually has a range of e.g. [87200, 87250].
>    
>     He also provided me with one of the datasets that caused the problem
>     but so far I am not able to reproduce the problem (tested with the CVS
>     head on Windows and Linux 32bit). I kind of suspected an uninitialized
>     variable somewhere and tried a valgrind test but it didn't tell me
>     anything meaningful either.
>    
>     Hence, I'm interested in under what exact condition this problem would
>     occur...? Did anyone else encounter the problem, and if so, is it 100%
>     reproducible?
>    
>     Takuya OSHIMA, Ph.D.
>     Faculty of Engineering, Niigata University
>     8050 Ikarashi-Ninocho, Nishi-ku, Niigata, 950-2181, JAPAN
>    
>     From: Shi Jin <jinzishuai at yahoo.com>
>     Subject: Re: [Paraview] Bug in Color Scale
>     Date: Tue, 25 Nov 2008 12:36:55 -0800 (PST)
>    
>     > Thank you Ken.
>     > This is really strange then. I just downloaded the 32-bit version for
>     Linux
>     > and had the same problems. It sounds like a problem of my linux box.
>     >
>     > By the way, I did change the color map to HSV earlier but I am sure I
>     didn't
>     > change anything else. And the new 32bit binary, I didn't do anything to
>     it and
>     > still get the [0,1] range.  I guess I have to use some other machines to
>     see
>     > the problems.
>     >
>     > Thank you for your quick response and I will keep you updated on my
>     findings.
>     >
>     > Shi
>     >
>     > ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
>     ━━━
>     > From: "Moreland, Kenneth" <kmorel at sandia.gov>
>     > To: Shi Jin <jinzishuai at yahoo.com>; "paraview at paraview.org"
>     > <paraview at paraview.org>
>     > Sent: Tuesday, November 25, 2008 12:27:47 PM
>     > Subject: Re: [Paraview] Bug in Color Scale
>     >
>     > Nope.  I still can’t replicate the problem.  I downloaded the binary
>     you
>     > specify, loaded the data set you sent, and it set the range fine.
>     >
>     > I notice from the image you sent that the color bar does not have the
>     default
>     > colors.  Are you sure that you are actually using the version of
>     ParaView you
>     > downloaded as opposed to another version that might be in your path?
>      Earlier
>     > versions of ParaView used that rainbow scalar bar as a default.  Either
>     that
>     > or you have been playing with color scale editor dialog box.  In that
>     case,
>     > you may have accidently changed the range without realizing it.
>     >
>     > -Ken
>     >
>     > On 11/25/08 9:35 AM, "Shi Jin" <jinzishuai at yahoo.com> wrote:
>     >
>     >     Hi Ken,
>     >   
>     >     Thank you. I have attached my sample data file and a screenshot
>     taken
>     >     under paraview-3.4.0 under 64bit Linux.
>     >     Just now I tried the Windows version and it seems to be working
>     fine. So I
>     >     guess this is a problem only found in the Linux version. By looking
>     at the
>     >     range for the variable Rank in the information panel, we can see it
>     is
>     >     [3,3] however the color displays only for range [0,1].
>     >     Please let me know if you can reproduce this problem on Linux. The
>     binary
>     >     I used is paraview-3.4.0-Linux-x86_64.
>     >   
>     >     Thanks a lot.
>     >     Shi
>     >
>     >     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
>     ━━━
>     >     From: "Moreland, Kenneth" <kmorel at sandia.gov>
>     >     To: Shi Jin <jinzishuai at yahoo.com>; "paraview at paraview.org" <
>     >     paraview at paraview.org>
>     >     Sent: Tuesday, November 25, 2008 7:50:42 AM
>     >     Subject: Re: [Paraview] Bug in Color Scale
>     >   
>     >     Re: [Paraview] Bug in Color Scale Shi,
>     >   
>     >     I have not run into this, nor have I heard of anyone else who has.
>      Could
>     >     you tell us more about your data or, better yet, send us an example?
>     >   
>     >     One thing that has changed with the rescale range feature recently
>     is that
>     >     the range is now computed based on the values that appear throughout
>     an
>     >     entire volume, not just what appears on the surface.  Thus, if you
>     are
>     >     rendering a volume of data with 1 in the middle but with the surface
>     only
>     >     having up to 0.01, the range is still going to be up to 1.  Check
>     the
>     >     ranges in the Information panel.  The scalar bar range should agree
>     with
>     >     that.  If you don’t know where the upper values are located, try
>     using the
>     >     threshold filter to find them.
>     >   
>     >     -Ken
>     >
>     >     On 11/24/08 8:24 PM, "Shi Jin" <jinzishuai at yahoo.com> wrote:
>     >
>     >         Hi there,
>     >       
>     >         I just upgraded my paraview from 3.10 to 3.40. Immediately I
>     found
>     >         that the same data was not colored correctly with the new
>     version. It
>     >         turned out that in the Display Tab, Rescale to Data Range button
>     >         actually does nothing. If I clicked on Edit Color Map.., I found
>     that
>     >         the maximum of data range is always 1 (minimum always 0), no
>     matter
>     >         what the real range is. I can of course uncheck Automatically
>     Rescale
>     >         to Fit Data Range, and set the range using Rescale Range button
>     by
>     >         hand. However, I think this is not what it is supposed to be.  I
>     tried
>     >         both the Linux-64bit binary found online and my own build on an
>     ubuntu
>     >         box and they have the same problem.  Please let me know if it is
>     >         indeed a bug or just something I didn't get right. Thanks a lot.
>     >       
>     >         Thanks a lot.
>     >         Shi
>     >
>     >         _______________________________________________
>     >         ParaView mailing list
>     >         ParaView at paraview.org
>     >         http://www.paraview.org/mailman/listinfo/paraview
>     >
>     >        ****      Kenneth Moreland
>     >         ***      Sandia National Laboratories
>     >     ***********
>     >     *** *** ***  email: kmorel at sandia.gov
>     >     **  ***  **  phone: (505) 844-8919
>     >         ***      web:   http://www.cs.unm.edu/~kmorel <http://
>     www.cs.unm.edu/
>     >     %7Ekmorel>
>     >
>     >    ****      Kenneth Moreland
>     >     ***      Sandia National Laboratories
>     > ***********
>     > *** *** ***  email: kmorel at sandia.gov
>     > **  ***  **  phone: (505) 844-8919
>     >     ***      web:   http://www.cs.unm.edu/~kmorel
>     >
>     > *
>     >
> 
>    ****      Kenneth Moreland
>     ***      Sandia National Laboratories
> ***********  
> *** *** ***  email: kmorel at sandia.gov
> **  ***  **  phone: (505) 844-8919
>     ***      web:   http://www.cs.unm.edu/~kmorel
> 
> *
> 


More information about the ParaView mailing list