[Celestia-developers] Location orientation

[Grant Hutchison <granthutchison@bl...>]

No, even using Orientation to correct the longitude mismatch in Goto Object,
LongLat is still 180 degrees out of orientation. To place the two named
crater on Deimos correctly, I have to use the following:
============================
Location "Swift" "Sol/Mars/Deimos"
{
    LongLat [ -178 13 0 ]
    Size 1.4
}

Location "Voltaire" "Sol/Mars/Deimos"
{
    LongLat [ -182 23 0 ]
    Size 2.9
}
===========================
Their *correct* longitudes are 2W and 358W, as can be seen from their
position on the sub-Mars hemisphere.

BTW: It seems as if Celestia uses Radius as some sort of *maximum* radius
when rendering 3ds models - is it the maximum radius overall, or the maximum
radius in some defined plane, or just the radius along a particular axis
which turns out to be the long axis of most models?
I ask particularly because it has relevance to positioning the locations
correctly - if the model isn't scaled correctly, the negative altitudes
won't work as expected if they're read off the original 3d data. (Although
even dropping the altitude of Voltaire to -0.1 makes it disappear at
present, although it's well clear of Deimos' limb with an altitude of 0.)

Grant

Re: [Celestia-developers] Location orientation

[Hank Ramsey <hramsey@qw...>]

On Sunday, June 22, 2003, at 01:58  PM, Grant Hutchison wrote:

> BTW: It seems as if Celestia uses Radius as some sort of *maximum* 
> radius
> when rendering 3ds models - is it the maximum radius overall, or the 
> maximum
> radius in some defined plane, or just the radius along a particular 
> axis
> which turns out to be the long axis of most models?

I've been wondering about this issue myself, in connection with labels 
for Eros. If I'm not mistaken, Celestia simply scales the model 
coordinates by the radius value specified. In other words, if a point 
in the model is at a distance 1.0 from the origin, it'll be scaled to 
the distance specified by the radius value. So the result will depend 
on how the model builder scaled the coordinates when creating the 
model. I would guess that in most cases the model builder would scale 
the model coordinates based on the point which is at a maximum distance 
from the origin. But I could be wrong, and in any case there's no 
guarantee, just as it's up to the model builder to get the center and 
orientation correct.

BTW, 'solarsys.ssc' currently gives a Radius of 16.5 for Eros, whereas 
the NEAR NLR190 shape model shows a maximum equatorial radius of 
17.452, with a maximum chord of 32.687 (not in the equatorial plane).

- Hank

Re: [Celestia-developers] Location orientation

[Chris Laurel <claurel@ww...>]

On Sun, 22 Jun 2003, Grant Hutchison wrote:

> No, even using Orientation to correct the longitude mismatch in Goto Object,
> LongLat is still 180 degrees out of orientation. To place the two named
> crater on Deimos correctly, I have to use the following:
> ============================
> Location "Swift" "Sol/Mars/Deimos"
> {
>     LongLat [ -178 13 0 ]
>     Size 1.4
> }
> 
> Location "Voltaire" "Sol/Mars/Deimos"
> {
>     LongLat [ -182 23 0 ]
>     Size 2.9
> }
> ===========================
> Their *correct* longitudes are 2W and 358W, as can be seen from their
> position on the sub-Mars hemisphere.
> 
> BTW: It seems as if Celestia uses Radius as some sort of *maximum* radius
> when rendering 3ds models - is it the maximum radius overall, or the maximum
> radius in some defined plane, or just the radius along a particular axis
> which turns out to be the long axis of most models?

Celestia normalizes 3DS models by finding an axis aligned bounding box of 
the model, then using the inverse of the long axis as a scaling factor.

> I ask particularly because it has relevance to positioning the locations
> correctly - if the model isn't scaled correctly, the negative altitudes
> won't work as expected if they're read off the original 3d data. (Although
> even dropping the altitude of Voltaire to -0.1 makes it disappear at
> present, although it's well clear of Deimos' limb with an altitude of 
0.)

I'm investigating this problem right now; I'll fix the problem with 
negative altitudes, though visibility determination for labels on 
irregular bodies still might not be ideal.

--Chris

Re: [Celestia-developers] Location orientation

[Hank Ramsey <hramsey@qw...>]

On Sunday, June 22, 2003, at 09:00  PM, Chris Laurel wrote:

> Celestia normalizes 3DS models by finding an axis aligned bounding box 
> of
> the model, then using the inverse of the long axis as a scaling factor.

So if I understand correctly, the model is centered at the center of 
the axis-aligned bounding box? And the scale factor is such that half 
the maximum dimension of the axis-aligned bounding box scales to the 
radius value? Is that right?

- Hank

Re: [Celestia-developers] Location orientation

[Grant Hutchison <granthutchison@bl...>]

Chris:
> Celestia normalizes 3DS models by finding an axis aligned bounding box of
> the model, then using the inverse of the long axis as a scaling factor.
Thanks. So all I need to do is find the maximum dimension of the
axis-aligned bounding box for a particular object, set *that* as the radius
of the object in Celestia, and things will scale correctly.
One final detail - are we talking the distance from the centre of the object
to the farthest face of the bounding box, or half the long axis of the
bounding box? Or are these things identical because the bounding box is
always symmetrical about the origin?

Grant