| Opinions on AVS versus PVWAVE? [message #1197] |
Thu, 01 July 1993 13:19 |
heirich
Messages: 1
Registered: July 1993
|
Junior Member |
|
|
I've been doing lot of scientific visualization with AVS. I have
not been happy with the experience and am frustrated to the point
where I'm about ready to pitch it into the lake. I have heard some
opinions about PVWAVE to the effect that it's much easier to use and
produces better looking output. I have no experience with PVWAVE so
I would like to solicit the net for opinions on a few points.
I'm primarily interested in making films, ie. sequences of images
which are visually coherent. (One of the frequent problems AVS has
is that it can create a sequence of images, but it resets parameters
or resizes or moves windows in the middle of a script, destroying the
visual coherence of the sequence).
Is anyone happily making films with PVWAVE? Does it support video
hardware such as Abekas equipment and tape decks? Are there any problems
with making sequences of frames?
Our biggest applications are computational fluid dynamics on unstructured
grids. This means that a visualization package must be prepared to deal
with unstructured data, ie. cells that are not distributed on a lattice.
The two main reasons we went with AVS initially was the support of an
"unstructured cell data" type and support for CFD visualization.
One nontrivial nuisance with AVS is that in order to import unstructured
data you not only have to give it the data with x,y,z coordinates for
each point, but you must also give it a definition of the vertices of the
cells which contain the data. This is a practical problem because we often
work with data from simulations where the only definition of these vertices
is implicit in the simulation code so it's not easily obtained.
Can PVWAVE handle unstructured data (eg. finite element grids)? Does
it support vector data (eg. velocity) at each point? Can it draw
streamlines, and/or advect particles through a flow field? Can you do
simple computations of a vector field, such as computing the surface normal
at every point?
A major conceptual flaw in AVS (IMHO) is that AVS has several different
data formats which it uses internally, and which are not interchangeable.
For example, I can run a simulation on a uniformly spaced grid and
visualize the results, but when I rerun it on a slightly different grid
(with a slight variation in the point spacing) I either (a) can't put
the data through the modules that I used for the uniform grid or (b) can
run them through but the results are utterly different and frequently
not useful.
Does PVWAVE hide these implementation level details of the data formats
from the user? A correctly designed package should be able to suck up the
data in whatever format you give it, and nearly all of the features should
be available to be performed on that data (barring a very small number
which would truly be nonsensical for particular formats).
We also do a lot of particle physics models. Does PVWAVE support
particle visualization very well? Can it create a sequence of frames for
a movie which show the particles moving through space?
Does PVWAVE support any video-editing features, such as combining images,
cross fades (combining images while adjusting their opacities), making titles?
Thanks for any inputs.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~
Alan Heirich, M.S., M.S. | heirich@caltech.edu
Scalable Concurrent Programming Laboratory | (818) 356 3903 - lab
256-80 California Institute of Technology | (818) 356 4600 - office
Pasadena, CA 91125 USA |
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
