[osg-users] (VPB) osgdem not providing terrain heights with --tile-terrain-size --tile-image-size options

[Robert Osfield]

HI Scott,

The max tile sizes is in number of pixels/height field cells, neither
are in meters.

Robert.

On Thu, Mar 27, 2008 at 1:19 PM, Hulet, Scott S <scott.s.hulet at lmco.com> wrote:
>
>  Robert,
>
>         Note that --tile-image-size was set to 1024, not 111120.  I
>  basically guessed that --tile-image-size was in texel units and
>  --tile-terrain-size was in meter units.
>
>         With --tile-image-size 1024, I was attempting to get osgdem to
>  use larger image tile dimensions; i.e. 1024x1024 instead of 64x64 (to
>  improve performance).  I attempted to convert some of my .ive files to
>  .osg files to determine what image tile sizes osgdem was using as
>  default (at each level), but I only see a reference to a .dds texture
>  for the children tiles (because .dds files are baked into .ive files).
>  I could not determine the dimension of the images.  Based on the .ive
>  file sizes, however, it appears that osgdem is using fairly small image
>  tile dimensions at the finer levels.  Is there a way to determine what
>  image tile dimensions are used at each level?  Our experience has been
>  that larger image textures perform better on modern graphics cards.
>  What is the correct way to drive osgdem to use a larger size?
>
>         With  --tile-terrain-size 111120, I was attempting to drive the
>  coarsest terrain tile to 1 large terrain region (111120 meters in a
>  geocell).  It actually did have an effect that was close to what I
>  desired.  The output was 4x4 facet pairs (32 tris) for the whole geocell
>  at the coarsest level.  Basically,
>
>         So, could you provide some info on the following options?  What
>  are the units?  What is the default behavior?  I also understand that
>  terrain region size is tightly coupled to image tile dimension size.  Is
>  this a case where you set one or the other, but not both?
>
>   --tile-image-size     Set the tile maximum image size
>   --tile-terrain-size   Set the tile maximum terrain size
>
>
>  > HI Scott,
>  >
>  > The sizes of --tile-image-size is absolutely massive, a normal figure
>  > would be something like 32 or 64, not 111120. The --tile-*-size
>  > parameters controls the dimensions of the tiles, you certainly don't
>  > want a 1111120x1111120 grid size.
>  >
>  > Is there something in particular you are tyring to do with this high
>  figures?
>  >
>  > Robert.
>
>
>
> -----Original Message-----
>  From: Hulet, Scott S
>  Sent: Wednesday, March 26, 2008 8:45 PM
>  To: 'osg-users at lists.openscenegraph.org'
>  Cc: Hulet, Scott S
>  Subject: (VPB) osgdem not providing terrain heights with
>  --tile-terrain-size --tile-image-size options
>
>  Robert,
>
>         When I generate with the --tile-terrain-size --tile-image-size
>  options, I do not get any terrain relief in my output .ive files.  All
>  of the levels and triangles (mesh) are present, but all polys are in a
>  flat plane.  When I do not use these options, all is well (with the
>  exact same input source; imagery and dem).  I am using OSG 2.35 and VPB
>  0.9.6.  Note that I am using the --terrain option in both cases.
>
>         I have provided the command lines that I am using.  Excuse the
>  syntax - I'm batching osgdem with a Python system call.
>
>  NO TERRAIN RELIEF
>  -----------------
>  osgdem cmd line: os.system("\"c:/Program
>  Files/VIRTUALPLANETBUILDER/bin/osgdem\" --PagedLOD --geocentric
>  --mip-mapping-imagery--compressed --terrain --tile-terrain-size 111120
>  --tile-image-size 1024 -e " + sw_lon + ' ' + sw_lat + " 1.00 1.00 -t
>  c:/data/scratch/osg_mt_home/imagery/" + geocell + "/tifs -d
>  c:/data/scratch/osg_mt_home/dted/" + geocell + " -o " + ive_out)
>
>  EXPECTED TERRAIN RELIEF
>  -----------------------
>  osgdem cmd line: os.system("\"c:/Program
>  Files/VIRTUALPLANETBUILDER/bin/osgdem\" --PagedLOD --geocentric
>  --mip-mapping-hardware --compressed --terrain -e " + sw_lon + ' ' +
>  sw_lat + " 1.00 1.00 -t c:/data/scratch/osg_mt_home/imagery/" + geocell
>  + "/tifs -d c:/data/scratch/osg_mt_home/dted/" + geocell + " -o " +
>  ive_out)
>
>  Thanks!
>
>  Scott S. Hulet
>  Simulation Training and Support
>
>  "They never forget who they're dying for..."
>  "We can never forget who we're working for..."
>
>  Lockheed Martin Corp - STS Akron
>  1210 Massillon Road
>  Mailpoint C2E, AB-33
>  Akron, OH 44315-0001
>
>  Phone Numbers: 330.796.6616 (Desk)
>                330.796.1059 (Lab)
>                330.242.2385 (Cell)
>
>             Fax: 330.796.4050
>  _______________________________________________
>  osg-users mailing list
>  osg-users at lists.openscenegraph.org
>  http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
>