GlobeMaker wrote:Rectilinear gaps without stars.
There is a possible bug in Celestia when viewing the whole star cluster shown below.
This star cluster includes all stars in the Galaxy. There are rectilinear gaps between
sectors of space. It seems the octree algorithm is putting empty gaps between
regions. The linear gaps outline solid brick shaped regions where stars are, but
with visible gaps with no stars. The empty gaps are exactly aligned with the
equatorial coordinate system. If you put the celestial grid's north pole in the star
cluster at Sol, the gaps are seen to be exactly aligned with the pole.
Your viewing position must be from 2k parsecs to 10k parsecs from
Sol. Center Sol and follow Sol. Use 5kpc and maximum star brightness.
Adjust field of view to begin to saturate a white zone. A checkerboard
pattern appears. The picture, below, shows how it looks on low
end graphics hardware. The same problem is seen on my NVIDIA hardware.
I used Windows XP and Celestia versions 1.4.0 and 1.4.1 and 1.3.2 FT 1.2.
They all showed the checkerboard of gaps between stars.
The low powered Graphics Controller was an Intel 82810 Adapter with 32 Meg of RAM
The NVIDIA adapter has 64 Meg RAM. It is a GeForce 4 MX . It is used on my
AMD Athelon at 2.21 G Hz with 2 Gig of RAM. Both showed the same checkerboard
of gaps between stars.
Versions 1.4.0 and 1.4.1 use star magnitudes up to 15.
Version 1.3.2 FT 1.2. only goes up to magnitude 12, but the gaps
are still visible as a checkerboard pattern. But this version 1.3.2 FT 1.2 needs to be
zoomed in to saturate to white. The 1.4.* versions need to be zoomed out to
saturate to white.
Use this procedure to see the bug quickly. Start Celestia. keys : h c f to
follow Sol. Display the celestia grid. Rotate to center one pole of the celestial
grid on Sol. (Display constellation borders ).
F7 q a to travel away from Sol to a distance of 4.7kpc. Use maximum star
brightness. Remove the display of the celestial grid. Adjust field of view until
the stars saturate to whiteness. Look at the checkerboard pattern. It is a bug.
Travel farther away to see octree adjustments occur in the checkerboard pattern.
The exact alignment of Sol to the pole of the celestial grid is critical and
very sensitive to deviations. Most users will not see the bug because of the
exact alignment needed. I found it because I was looking for gaps with exact
alignments, hopefully due to dust lanes. But dust lanes don't form rectilinear
grids in three dimensions.
Here is one picture with low end hardware and version 1.4.0 :