[Paraview] Bug or feature?

Wed May 10 01:40:04 EDT 2017

Hello,

Sorry for not posting to the list: this was by mistake, not my intention!
(But there is also this "feature" of the list that it never sends me my own
postings back - with the effect that my mail program (Gmail) moves threads
in the mailing list that I have initiated to my personal mail folder, no
way to bring it to the forums.

I am thus posting the stripped down example once more for all, together
with screen shots that illustrate the problem:

example_0.zip - contains the vtu file with the unstructured grid.
Displaying Col2Comp::Comp2 shows the problem best

screenshot1.jpg - shows how it looks with "Surface" rendering on my computer
screenshot2.jpg - shows the same with "Volume" rendering --> WRONG!
screenshot3.jpg - shows the same again, but now after extracting
Col2Comp::Comp2 into a separate column using the "Calculator" filter -->
CORRECT!

Regarding the color map editor I can only say so far that I did not change
anything between the 3 screenshots - and since screenshot3 shows the
correct display I am not really sure that this is the issue.

The cell connectivity is indeed not as it is typically expected: The blocks
are touching each other, but only vertically adjacent blocks also share
edges and nodes. This is on purpose because it reflects the kind of problem
I am dealing with. Of course it is possible that also this influences the
volume rendering, but again the fact that screenshot 3 is correct while 2
is wrong lets me doubt this track.

My guess without further investigation is that it might be some caching
issue: Assuming that a rendered volume is somehow kept in memory for a
possible further use, there might be some problem with managing the cached
data for "columns" and "components" properly.

In other words: My current guess is that there is a bug in the volume
rendering.

With kind regards - and thanks for any helpful hints!
Cornelis

2017-05-10 2:43 GMT+02:00 Andy Bauer <andy.bauer at kitware.com>:

> Hi,
>
> Please keep the discussions on the mailing list so that everyone that
> wants to can follow along.
>
> I'm able to volume render by components as well as magnitude. It may be
> that you need to play around with the Color Map Editor to get things as
> you'd like them to look.
>
> I'm not an expert on volume rendering but it seems to me like something is
> wrong with your cell connectivity when I use the Extract Surface filter.
> It's possible that this is also causing issues with the volume rendering
> but I couldn't say for sure.
>
> Andy
>
> On Sun, May 7, 2017 at 5:20 AM, Cornelis Bockemühl <
> cornelis.bockemuehl at gmail.com> wrote:
>
>> This is my example, mostly stripped down to the core of the problem: see
>> the Col2comp data array that has 2 components (see attached zipped vtu
>> file). The model has something more than 600 cells, and in both components
>> (comp1 and comp2) many of them are 0. In comp2 it is even all except 18 -
>> and this can easily be seen in a "surface" view: almost everything is blue
>> == 0. It can also be analyzed with cut and slice filters, so it is also the
>> case "inside" the model.
>>
>> Now with a volume view this is completely different: comp2 looks as if
>> indeed many cells have values "above blue"!
>>
>> But I must admit that initially I did not look well enough: also the
>> volume view changes a little if I change between "magnitude", "comp1" and
>> "comp2" - only that it does not look like matching the surface views at
>> all: all cells except one actually have a 0 in one of the two components -
>> which should be clearly visible also in the volume view I suppose!
>>
>> I could reproduce this behaviour not only on my self-compiled version 5.3
>> paraview, but also in a downloaded and installed binary version 5.1.2
>> (64-bit).
>>
>> In other words: I am a bit confused now! ;-)
>>
>> Regards,
>> Cornelis
>>
>>
>>
>> 2017-05-07 0:49 GMT+02:00 Andy Bauer <andy.bauer at kitware.com>:
>>
>>> Hi,
>>>
>>> I am able to do a volume visualization for each component of a 3
>>> component array (i.e. a vector) of both point and cell data in PV 5.3. This
>>> was done with a both a vtkImageData and a vtkUnstructuredGrid. Once I
>>> choose my 3 component array for volume rendering in the upper left corner
>>> of the GUI, the default is the Magnitude but I can switch to each component
>>> as well with the drop down dialog that is just to the right of where I
>>> choose which array I want to pseudo-color by.
>>>
>>> Any chance you could share your dataset and specific instructions on how
>>> you're doing things? Alternatively, you can use the Calculator filter to
>>> extract one component of your multi-component array but this is a bit
>>> computationally wasteful.
>>>
>>> Cheers,
>>> Andy
>>>
>>> On Sat, May 6, 2017 at 12:15 PM, Cornelis Bockemühl <
>>> cornelis.bockemuehl at gmail.com> wrote:
>>>
>>>> Dear Paraview users and experts,
>>>>
>>>> In some project I am generating a dataset that has a multi-component
>>>> column as a result, with a variable number of columns depending on the
>>>> problem. These data are "cell data" of some "unstructured grid", and I can
>>>> nicely visualize it in Paraview, component by component.
>>>>
>>>> Only with the visualization as a "volume" I have problems: It seems
>>>> that with this you can only visualize the "magnitude", while switching
>>>> between the components does not work! This works only in other viewing
>>>> modes, but the "volume" is sometimes very useful because you can look
>>>> "inside" the volume.
>>>>
>>>> Now I am asking myself if this is a bug (with a certain chance to have
>>>> it fixed or even fix it myself...) or a feature?
>>>>
>>>> I mean: the fact that "magnitude" is the default for displaying
>>>> multi-component arrays already gives some hint in the direction: this type
>>>> of columns is mostly intended to be used as coordinates! Which is actually
>>>> not my case.
>>>>
>>>> Technically I could of course switch to just generating a set of
>>>> single-component arrays instead of one multi-component array. I only did it
>>>> otherwise because the columns are very closely related, so it looked like a
>>>> good idea initially...
>>>>
>>>> Any hint or experience?
>>>>
>>>> Regards,
>>>> Cornelis
>>>>
>>>> PS: I am using Paraview 5.3 on Windows 64bit
>>>>
>>>> --
>>>> Cornelis Bockemühl
>>>> Basel, Schweiz
>>>>
>>>> _______________________________________________
>>>> Powered by www.kitware.com
>>>>
>>>> Visit other Kitware open-source projects at
>>>> http://www.kitware.com/opensource/opensource.html
>>>>
>>>> Please keep messages on-topic and check the ParaView Wiki at:
>>>> http://paraview.org/Wiki/ParaView
>>>>
>>>> Search the list archives at: http://markmail.org/search/?q=ParaView
>>>>
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://public.kitware.com/mailman/listinfo/paraview
>>>>
>>>>
>>>
>>
>>
>> --
>> Cornelis Bockemühl
>> Basel, Schweiz
>>
>
>

-- 
Cornelis Bockemühl
Basel, Schweiz
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview/attachments/20170510/7bc61325/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: example_0.zip
Type: application/zip
Size: 33195 bytes
Desc: not available
URL: <http://public.kitware.com/pipermail/paraview/attachments/20170510/7bc61325/attachment.zip>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: screenshot1.jpg
Type: image/jpeg
Size: 23175 bytes
Desc: not available
URL: <http://public.kitware.com/pipermail/paraview/attachments/20170510/7bc61325/attachment.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: screenshot2.jpg
Type: image/jpeg
Size: 14634 bytes
Desc: not available
URL: <http://public.kitware.com/pipermail/paraview/attachments/20170510/7bc61325/attachment-0001.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: screenshot3.jpg
Type: image/jpeg
Size: 11227 bytes
Desc: not available
URL: <http://public.kitware.com/pipermail/paraview/attachments/20170510/7bc61325/attachment-0002.jpg>