From: Paul Argo (pargo@lanl.gov)
Date: 03/15/2001 06:43:01
I have been working with VOluumizer 1.2 for several months, and consider
my time a nearly complete success. For example, I have gotten the
monsterview demos working in perspective mode, with stereo, and it has
really impressed the visitors to our Fakespace RAVE (SGI has made it
easy for me to appear a genius!)
I have gotten the regular vogl demos working with our own colr table
interface, so we can highgrade out different density regions by changing
the opacity maps -- the conversion was nearly a no brainer. The
multipipe version (monsterview) has been much harder -- I cannot get the
screen image to respnd in any coherent manner to things I do with the
color table. something in its setup, that I have been unable to find
explained in the manuals, is affecting what is going on. The dataType,
internal and external format statements and everything seem to be the
same, and yet somehow it ends up different. One obvious difference, of
course, is that for the same data the brick sizes are several times
smaller becaus ethey are spread across several (6 in our case) pipes.
One clue is that when I rasied the debug level, in the
'voAppearanceActions::volumeOptimize' method thngs were/are different.
Where the external and interal data formats in the vogl case were
'luminance' and 'intensity8_ext' respectively, they were/are
'luminance_alpha' and 'dual_intensity8_sgis' in the monster case. I am
completely baffled, and throw myself on the mercy of some of the real
geniuses I hoipe are using this list.
Also, since we are doing stereo perspctive, and the users are going to
want to be 'inside' their data sets, I believe the best sampling mode is
going to end up 'spherical', which SGI leaves as an exercise for the
reader. I found the class and method 'voSamplinSurfaceSet' and
'generateEquation()' but am not sure what the form of the equation
should be. Has anyone solved this case, and how well does it work?
thanks ahaead of time
and hopefully will see some of you at the SGI Developers Conference next
week
paul argo
Los Alamos National Lab
pargo@lanl.gov
505.667-8355
This archive was generated by hypermail 2b29 : Thu Mar 15 2001 - 07:16:28 PST