Map flips horizontally in wxNVIZ

[MilenaN]

Use elev_lid792_1m in 3D view. Go down with Height slider, abot 116 height the map will flip.

[annakrat]

As far as I remember nviz has always behaved this way (please correct me if I am wrong). But I agree that the flipping is strange.

Anna

[hamish]

Anna wrote:

As far as I remember nviz has always behaved this way

yes, you get the same effect in tcl/tk NVIZ and I guess m.nviz.image.

i.e. when looking up from underneath it flips it:

# NC dataset
g.region rast=elev_lid792_1m
r.univar elev_lid792_1m
...
mean: 120.769

I'm not sure how useful looking from underneath is for DEMs, but it is probably of interest when working with 3D raster volumes.

In tcl/tk nviz I notice with elev_lid792_1m if you do look:[Top], set the perspective to 15.0, zexag to 1.0, and height to 7650.01 the image disappears. At 7650.00 it reappears. In the wxnviz is does the same but it transitions out between 7610 and 7630m viewing height.

at perspective=15.0 and height=2000.0 you can see it in the process of disappearing at zexag=4.00 (actually the perspective doesn't matter), just the tops of the hills are visible. (both tcl/tk nviz and wxNviz)

for m.nviz.image I've been fighting the flip in another way while trying to get the viewing height to actually be where you ask it to be. i.e. changing the viewing height is all that is needed to flip it. (height*2.0/zexag as a correction factor seems to help avoid having the z-exag change the viewing height without you asking; correcting that for changing perspective value is probably much more math to cancel out), but that's when slightly adjusting the viewpoint in top-down view, which is a bit like balancing on a pin head. For now I added a -r rotate flag to manually flip it if needed in the d.anaglyph addon. I think in m.nviz.image the flipping I'm seeing could possibly happen with z-exag at the same place as it makes the surface disappear in nviz, but I'd have to test that.

Hamish