More problems with synchronous CustomOrbits

[granthutchison]

I posted about non-rotating Miranda a few days ago.
Several other planetary satellites with CustomOrbits are not quite maintaining synchronization between spin and orbit. While most synchronous satellites in Celestia maintain good longitude orientation for many hundreds of years, a few drift noticeably after only a few years (thereby messing up my RotationOffset posting!)
As a fix, I've been able to get them all to behave properly by adding accuration RotationPeriods to their definitions.
The affected satellites and the corresponding RotationPeriods are:

Europa RotationPeriod 85.22834590
Titan RotationPeriod 382.69074184
Miranda RotationPeriod 33.92350159
Ariel RotationPeriod 60.48909292
Umbriel RotationPeriod 99.46022991
Titania RotationPeriod 208.94077099
Oberon RotationPeriod 323.117567539

Grant

[Calculus]

Grant, you are doing a wonderful job!
There are so many fixes you've already done. I'm not sure I got all of them. It would be nice to gather them in one post or provide a .scc with all fixes with some comments or any other practical way one can come up.

[granthutchison]

Grant, you are doing a wonderful job!
There are so many fixes you've already done. I'm not sure I got all of them. It would be nice to gather them in one post or provide a .scc with all fixes with some comments or any other practical way one can come up.

Real soon now I hope to post an "alternative" annotated solarsys.ssc file in the web space Selden has kindly offered me.
This should include:
1) Corrections to some planetary axis orientations
2) RotationOffsets to give correct prime meridian positions for the planets
3) Satellite prime meridians orientated correctly
4) Basic texture maps revised to a zero-in-the-middle standard
5) N-S mirrored maps of Phobos and Diemos, to fix the mapping problem there
6) Correctly orientated texture maps of the Uranian satellites, placing the original map south at Celestia north
7) A couple of gridded maps that can be laid on to allow close observation of the behaviour of the various bodies (at present my entire Celestia solar system consists of white objects marked with numbered grids - latitude increasing westwards for direct rotators, latitude increasing eastwards for retrograde rotators)

Grant

[chris]

Real soon now I hope to post an "alternative" annotated solarsys.ssc file in the web space Selden has kindly offered me.
This should include:
1) Corrections to some planetary axis orientations
2) RotationOffsets to give correct prime meridian positions for the planets
3) Satellite prime meridians orientated correctly
4) Basic texture maps revised to a zero-in-the-middle standard
5) N-S mirrored maps of Phobos and Diemos, to fix the mapping problem there
6) Correctly orientated texture maps of the Uranian satellites, placing the original map south at Celestia north
7) A couple of gridded maps that can be laid on to allow close observation of the behaviour of the various bodies (at present my entire Celestia solar system consists of white objects marked with numbered grids - latitude increasing westwards for direct rotators, latitude increasing eastwards for retrograde rotators)

I'll integrate your changes--well, the ones I haven't gotten already--into the main Celestia tree. I'll probably make versions of the textures maps myself, however. I want to start with high-resolution source images to avoid artifacts from decompressing, modifying, and recompressing the standard images.

--Chris

[granthutchison]

I'll probably make versions of the textures maps myself, however. I want to start with high-resolution source images to avoid artifacts from decompressing, modifying, and recompressing the standard images.

Great. The maps that need attention are:

Phobos and Diemos
A north-to-south mirror to bring them into register with the slightly odd behaviour of the 3ds shape models when texture maps are pasted on.

Mars, Io, Europa, Ganymede
A 180-degree shift in longitude, to bring the prime meridian from the edge to the centre.

Callisto
A 150-degree (eastwards) shift in longitude, to bring the prime meridian to the centre.

Ariel, Umbriel, Titania, Oberon (+ Venus surface map, Pluto & Charon albedo maps)
These bodies are generally mapped with the convention that the north pole is the one north of the ecliptic plane, making their rotation retrograde. Celestia north poles lie south of the ecliptic for these bodies (direct rotation with inclination greater than 90 degrees).
So all these texture maps need to be rotated 180 degrees to swap north for south, bringing surface features back into the correct orientation.

Miranda
Requires two manoeuvres: a 90-degree (eastward) shift in longitude to bring the prime meridian to the centre, and then a 180-degree rotation to swap north for south, as for the other Uranian satellites.

BTW: When playing with the Goto Object function, I noticed that it counts longitude eastwards from the prime meridian. But the usual astronomical mapping convention numbers longitude westwards from the prime meridian for all directly rotating bodies apart from the Earth, and uses eastwards numbering to flag maps of retrograde rotators (Venus, the Uranians, Triton, Pluto and Charon). So what with the differing poles conventions and the differing longitude conventions, there's a deal of confusion to be had by anyone trying to navigate Celestia while armed with a USGS/NASA/JPL map!
I did consider reversing Celestia's planetary axes to match the mapping, and defining negative rotation periods to create retrograde rotation - which would solve the north/south confusion in the maps and in Goto Object - but I was stymied when I realised I couldn't reverse the CustomOrbits of the Uranian satellites.

Grant