Rings through a planet

[pht]

The Aantlbzz/Zeenizzz system contains an oceanic planet Andrei, which sometimes seems to pass right through the rings of a nearby gas giant ... I guess that in reality, something like that would inevitably lead to rapid extinction of all life on the planet

[loki1950]

All the planetary rings don't work right ATM most likely drawing in the wrong order.

Enjoy the Choice

[bgaskey]

I thought that bug had been fixed prior to release, but looks like its still there in some form

[klauss]

That bug will never be fixed before Ogre (0.6 or who knows), or some major techniques work. Rings are a signle mesh that is both behind and in front of planets. It's impossible to draw correctly without z-buffer, and at the distance they're seen, they're always in the "far" queue (the one without z buffer).

Well, actually there's one way without z buffer. One mesh per segment. But noone wants that.

[ace123]

What if we added a hack to first draw the polygons of a planetary ring farther than the distance from its center to the camera (i.e. a comparison with magnitude squared, so it shouldn't be a big penalty)

Then, draw all other objects in the far queue as well as the closer half of each ring.

[pht]

Rings are a signle mesh that is both behind and in front of planets. It's impossible to draw correctly without z-buffer, and at the distance they're seen, they're always in the "far" queue (the one without z buffer).

I'm sending a screenshot of the phenomenon. Just to clarify - the rings are not around the oceanic planet, rather around a nearby gas giant.

Seems you're right, the planet always appears "behind" the rings, no matter from which side I look. This would be OK, but if I fly close enough to the planet then the rings begin to cross the planet. That shouldn't happen given the rings are not drawn via z-buffer.

[pht]

PS: I just observed the rings and they seem to rotate around the gas giant and by doing so they seem to geometrically pass through the planet (so it might not be only a rendering issue). I maneuvered to the close proximity of the rings and followed their trail... until I crashed into the planet.

BTW, if I understand this correctly then the system in question is generated dynamically, so here are the relevant parts of its xml:

Code: Select all

               <Planet name="Monterey" file="planets/gas_giant3.png" ri="-29217513472.000000" rj="-30911393792.000000" rk="-24425068544.000000" si="-26328995840.000000" sj="14662536192.000000" sk="12938689536.000000" radius="3769632.750000" x="-2964058624.000000" y="40404983808.000000" z="-6631231488.000000" year= "86033014784.000000" day="133.117340" >

                        <Ring file="rings/planet_ring.texture" innerradius="7255842.681201" outerradius="15746885.681201"  wrapx="1" wrapy="1" />

                        <Planet name="Andrei" file="planets/ocean.png" ri="40452840.000000" rj="-8287163.500000" rk="-17182194.000000" si="-187432.343750" sj="-13337928.000000" sk="5991750.000000" radius="1205707.000000" x="-53399328.000000" y="58345396.000000" z="-140282976.000000" year= "58989628.000000" day="814.326660" >

[klauss]

Well, it appears that in this case the rings actually go through the planet, and that you did get close enough for the planet to be redered on the near queue (with z-buffer), which is cool (I didn't know a gas giant could get into the near queue, but apparently it can!)

Wicked!

So, ya... fixing the autogenerator to avoid such a thing would be nice, but then again there are so many things wrong with that autogenerator...

[LukeX717]

I just want to point out a minor (hard to notice) rotation axis bug:
In VS the planet rings are rotating together with the planets. But real planet rings rotate more the way like a frisbee and not like a coin thrown up in the air. (In VS the ring rotation axis is wrong)

[pyramid]

Thanks for noticing that. I'll put this up on my agenda for improving the start systems auto generator.