[GRASS-dev] Replacement of NVIZ

Sören Gebbert soerengebbert at gmx.de
Mon Apr 23 07:19:59 EDT 2007

Sorry! More in the text below.

Sören Gebbert schrieb:
> Dear Michael,
> Michael Barton schrieb:
>> It's great that we can export GRASS data to be used in high-end
>> multi-dimensional visualization platforms like Paraview and MayVi2. I've
>> used Paraview a bit. I wouldn't call it easy or intuitive, but it is 
>> quite
>> sophisticated and powerful--though not particularly fast in my 
>> experience.
>> However, using any of these packages is not a replacement for NVIZ. NVIZ
>> will let a user quickly and seamlessly render 2, 2.5, 3, and 4 
>> dimensional
>> data within a GRASS GIS session. This is a real bonus. And NVIZ does this
>> better than any other rendering engine within a GIS that I've seen. In 
>> this
>> sense, NVIZ has a different goal from Paraview and other dedicated
>> multidemensional rendering packages.
>> So I hope that we can get NVIZ ported to wxPython and make it an even 
>> more
>> seamless part of the geospatial visualization tools for GRASS. Making
> Good news. Nice to hear that. I hope you will redesign the gui, so it 
> will be more intuitive.
> I have had a look at the NVIZ code and was afraid that this construct is 
> not maintainable.
> Well, this is hopefully related to my very little knowledge of software 
> design and C coding
> with Tcl/Tk bindings.
>> tighter connections between GRASS and other, external visualization 
>> tools is
>> also a worthy plan--similar to the integration between GRASS and R. But
>> IMHO, we should not abandon NVIZ or something like it.
> I guess you are right. An integrated 2.5d, 3d and 4d visualization tool 
> is important.
> The good thing is that ParaView is not only a visualization tool, it is
> a preprocessing and analysis tool with 4d support and powerful threaded 
> image processing.

I mean post-processing. Sorry for confusion.


> Best regards
> Soeren
>> Michael
>> On 4/22/07 4:31 PM, "Sören Gebbert" <soerengebbert at gmx.de> wrote:
>>> Syd,
>>> Syd Visser schrieb:
>>>> Soren
>>>> Paraview3 uses Qt and MayaVi2 uses Wxpython thus we are leaning more
>>>> towards MayaVi2  although we use Paraview2.6 extensively but 
>>>> strictly as
>>>> a viewer.
>>>> We find MayaVi2 is also more open to user development thus easier to
>>>> extend.
>>> I'm developing with VTK and Qt since several years and have used 
>>> ParaView1-2
>>> for several years.
>>> My experience with MayaVi is little, because the user interface was too
>>> horrible.
>>> IMHO ParaView3 is the better choice. It has a sophisticated but very
>>> intuitive user interface and is developed by well known institutes and
>>> kitware.
>>> And they are doing a great job. ParaView is designed to visualize huge
>>> datasets in parallel.
>>> MayaVi2 depends on Traits, TVTK, Envisage and wxPython. A lot of new
>>> dependencies (except wxPython).
>>> And for now i'm not able to get even the new grass wxPython gui to 
>>> run on my
>>> debian etch system
>>> because of the dependencies.
>>> ParaView3 depends only on Qt4.2. Qt is available for many, many 
>>> platforms as
>>> well as VTK.
>>> (i still don't understand why wxWidgets and python was choosen for 
>>> the new
>>> grass gui and not Qt
>>> and python ...)
>>> The only thing we need to provide is a data server and gui plugin for
>>> ParaView3:
>>> http://www.paraview.org/Wiki/Plugin_HowTo
>>> And when ParaView3 reaches a stable state and i have some spare time i'm
>>> absolutely willingly to implement them!
>>> IMHO the data server and the reader/writer to the grass database
>>> (grass data should be modified with Paraview3 and stored back into 
>>> the grass
>>> database :) if possible)
>>> should be implemented in C++ for performance reasons. I would not use 
>>> the
>>> grass python wrapper
>>> to get the data into a visualization system. The grass raster, voxel and
>>> vector functions can be
>>> accessed from C++ code directly.
>>> The data server should provide access to the grass database to read 
>>> and write
>>> raster, volume and vector data.
>>> And ParaView3 should be extended with a nice little Qt gui to access 
>>> the grass
>>> data directly from the toolbar
>>> (like Qgis). We don't need to touch the ParaView3 sources, we only 
>>> need to
>>> implement plugins.
>>> A screenshot of ParaView3 handling grass raster, volume and vector data
>>> (exported with the *.out.vtk modules)
>>> is available here:
>>> http://www-pool.math.tu-berlin.de/~soeren/grass/modules/screenshots/ParaView3_ 
>>> Screenshot_grass_data.png
>>> Just my two cents ...
>>> sorry for my English
>>> Best regards
>>> Soeren
>>>> Thanks
>>>> Syd
>>>> Sören Gebbert wrote:
>>>>> Hi,
>>>>> Syd Visser schrieb:
>>>>>> We use enthought open source tool suite http://www.enthought.com/
>>>>>> which contains Numpy, SciPi and numerous other packages.
>>>>>> our main use is with Matplotlib and Chaco for 2D graphs and MayaVi2
>>>>>> for 3D graphics
>>>>>> I think these packages would be well worth having a close look at
>>>>>> especially MayaVi2 (Python wrapped VTK) for the 3D graphics.
>>>>> I would like to prefer a C++ grass data server + grass gui plugin for
>>>>> paraview3
>>>>> to visualize 3d data. This would a nice and fast solution.
>>>>> Best regards
>>>>> Soeren
>>>>>> thanks
>>>>>> Syd
>> __________________________________________
>> Michael Barton, Professor of Anthropology
>> School of Human Evolution & Social Change
>> Center for Social Dynamics & Complexity
>> Arizona State University
>> phone: 480-965-6213
>> fax: 480-965-7671
>> www: http://www.public.asu.edu/~cmbarton
> _______________________________________________
> grass-dev mailing list
> grass-dev at grass.itc.it
> http://grass.itc.it/mailman/listinfo/grass-dev

More information about the grass-dev mailing list