Post #3by Darryl Roy » 11.01.2004, 23:42
No, there is no striping on this texture. It was originally produced for
the Orbiter community, and there, immersion is everything...
I'm a latecomer to the huge textures game, and have been filling gaps
in Orbiter's solar system - there aren't many. I originally wanted to
simply port Praesepe's moon, but to my eyes, there was way too much
mosaic striping, the albedo contrast between highlands and maria was
too great, and there was a very odd shade of green to the maria that didn't accord with most astrophotography.
Since grappling with the Clementine data myself, it becomes clear that
all the previous large texture makers had assumed 1) that the 20x5k
near-IR greyscale (which didn't pick up much from the maria at all) represented white light albedo, and 2) that the approx 10x5
color basemap for Clementine represented true color, when in fact the
near-IR filter was mapped to the green RGB channel and the mid-IR was
mapped to red. There was no filter on Clementine covering the midrange.
So, I spent a couple of weeks making my own, mapping the near-IR to
red, and faking the green with an average. Checked against Don Davis's lunar true color references, it appears to be an extremely good match.
Otherwise, the only global changes were a greyscale albedo remap to
reduce contrast between 'reddish' and 'bluish' lavaflows in the maria and the creation of fictional polar color. Otherwise, just a 5-600 local patch jobs (cut/paste/blur edges fills on coverage gaps in the greyscale,
desaturating filter-pixel dropouts on crater rims which show saturated neon, and applications of GIMPs destripe filter to eliminate or reduce all
visible vertical banding).
There are still two extant farside glitches, a discernable (but diffuse) albedo change between the east and west farsides at the seam, and one neon color dropout that I missed. I'm planning on fixing these once I
get my true color 16k Mars done...
Thanks for the pointer to hosting information.
