[GRASSLIST:5483] Re: NVIZ and GTOPO30 on osx

[Glynn Clements]

Ian Macmillan wrote:

> Hi all, I have been trying to visualize a gtopo30 tile in NVIZ, but
> rather unsuccessfully. When I load the tile into NVIZ the vertical
> exxagerration seems to be incredibly high. When I try to draw the
> image with anything but zero vertical exaggeration, the mountain peaks
> become incredibly exaggerated or the image simply does not appear. 
> With zero exaggeration, the only wire mesh that shows is a single line
> that seems to be several thousand meters above the flat raster. In
> addition, in some views with a z exag of 0.001 (the peaks are still
> really pointy), there are some "phantom" E-W trending dark bars that
> appear at several thousand meter (>10000m?) elevations. Querying the
> data in NVIZ reveals correct elevations, it is just the image that is
> really distorted.
> 
> The DEM shows up in GRASS as a raster and a shaded relief (using an
> aspect map). When I brought in the dataset I changed all of the -9999
> values to null using r.in.bin (as signed integers). r.info tells me
> that the elevation values are -79 to 4328 (as they should be), and
> everything looks right with this data in GRASS. I have tried changing
> the region size (g.region) without any real success so far. I have
> also messed with the polygon resolution with no real success.
>  
> I have gotten nviz to work in other locations with other DEMs, so I
> know it works on my machine, so I can't figure why I can't image this
> data. Any ideas out there?

NVIZ doesn't like lat-lon locations; it acts as if the horizontal
units are metres. So, for a GTOPO30 DEM, you have a 40m x 50m region,
with 4000m high peaks (100 times the width of the region).

Importing into an X-Y or Plate Carée (Equidistant Cylindrical)
location should produce more sensible results.

-- 
Glynn Clements <glynn.clements at virgin.net>