Onward to glory, comrades! (1.5.1 ideas)

[Hungry4info]

Celestia 1.5.0 is finally done, and it's very nice. But it is, of course, imperfect. I now shed the dark shaddow upon all the happiness and celebration. Here's some issues I've found, but am hesitant to call bugs.

Ring shadows on Uranus -- Unrealistic as angular size of the sun is greater than that of the rings from the surface of Uranus.

Ring shadows on Neptune -- Unrealistic, as angular size of the sun is greater than that of the rings from the surface of Neptune, but not by much. I would guess the shadows on Neptune would be very hazy, rather than as defined as represented in Celestia. Perhaps a fading effect can be done for ring shadows, as is done for eclipse shadows.

Neptune's ring shadow isn't aligned right, there is a very tiny misalignment. I suspect this to be a result of Neptune not being represented as a perfect sphere, but rather as oblate, since the severity of the misalignment increases as the ring shadow moves into higher latitudes over the Neptunian year.

Ring shadows at higher latitudes for both Uranus and Neptune become exceptionally bold, rather than less visible. At the higher latitudes, the rings have even less angular diameter due to the greater distance from the rings. The sun, meanwhile, keeps pretty much the same angular diameter. Thus, the shadows should be even less pronounced at higher latitudes.

Near surface atmospheric bug near oblate planets. Oblate objects have a outer sphere, I suppose you can call it, which represents the limit to how close you can approach using the 'Home' key. Approaching this sphere results in not only your distance to approach 0, but for you to be within the atmosphere. This is, of course, reasonable. For oblate objects, however, the sphere does not conform to the shape of the planet. Thus, while over an oblate planet's higher latitudes, you close in on 0 height and yet, you're still way up above the surface. If the plane thas an atmosphere, when you descend below this sphere, the atmosphere completely vanashes. Yet, if you get close to the planet itself (the model of the planet, rather than the spherical region), the sky appears as you get closer to the model. I know this is all rather vague, if you wish for me to post images, please, just ask.

GD 66 b died X_X. It's no longer confirmed and is now in the Extrasolar Planet Encyclopaedia's 'unconfirmed' planets catalogue. It should probably be removed from Celestia.

For planets with Mie, when you are very distant and under high magnification, the mie effects jitter. The severity of the jitter increases with distance. Also, the jitter is more severe for planets close to their parent stars, even to the point that it is noticable even without magnification for hot Jupiters.

I put HD 209458 b within a comet, and (ignoring the agonizingly white cap of the comet) it appears to be somewhat glitchy. When you have HD 209458 b selected, pressing 'o' to enable orbit rendering causes it to look different. Specifically, when orbit rendering is enabled, the planet is opaque such that you cannot see the other side of the cometary cone through it. However when orbits are disabled, you can see both sides of the comet cone through the planet.

It seems that user-created stars are not gifted with the texture that Celestia's default stars are given.

But yeah, I'm not very good with describing things, so if anybody needs me to illustrate any of these with shots from within Celestia, please, tell me. I don't mind.

[BobHegwood]

Hungry,

I'm puzzled by your Neptune and Uranus Ring Shadow comments. I
have never ever seen a ring shadow for either of these planets. This
includes my old Gateway machine and my new one described in the
signature below.

Why would this be? Any ideas? I certainly see them for Saturn, but I'm
pretty sure that this is the only planet on which I can see the ring
shadows. Hmmm....

I do have one other comment though. Do you suppose that we could
ever get an ISS representation which utilizes the real colors of the
solar panels?

At any rate, thanks very much for the latest release, Chris. Very
nice when compared to the older versions, of course.

[Hungry4info]

I'm puzzled by your Neptune and Uranus Ring Shadow comments. I have never ever seen a ring shadow for either of these planets. This includes my old Gateway machine and my new one described in the signature below.

Why would this be? Any ideas? I certainly see them for Saturn, but I'm
pretty sure that this is the only planet on which I can see the ring
shadows. Hmmm....

Uranus' are difficult to see, I would guess if you searched enough, you may be able to find them. If you wish, I can upload a picture. Also, Neptune's rings aren't aligned right at the moment to cast shadows upon the planet, you'll have to run Celestia back a few decades, or foreward a few decades to see it.

I do have one other comment though. Do you suppose that we could ever get an ISS representation which utilizes the real colors of the solar panels?

I haven't thought of that at all! But it's a good idea. I think there were plans to use bh's ISS model as a default ISS model, but I don't know how far that gotten.

[BobHegwood]

Well, no matter what I do, or in what render path, I do NOT see ring
shadows on any planet except for Saturn.

No big deal though. I guess that I just assumed that these rings were
so slight, that they wouldn't render an effect on the planets of Uranus
and/or Neptune. Like I said, no big deal.

By the way, bh's ISS models had the wrong colors too, until I pointed
out the discrepancies to him. Bob was then kind enough to make
the changes. It only takes about 5 seconds to do, so I just wondered
why it hadn't been done within Celestia. Again, no big deal though.

I don't use the included ISS model anyway, as I prefer to view it in its
actual state of construction.

Take care, Bob

[selden]

Hungry,

Like Bob, I see no ring shadows whatsoever when looking at Uranus. Can you provide a picture and/or URL where you see them?

Here's a screengrab of Celestia showing Uranus with its rings in front of the planetary disc. They're very fine black diagonal lines.


And here's the corresponding URL.

Rings above Uranus

You can tell that these are the rings and not their shadows because they move across the surface as you change the viewpoint using the right-mouse-button. Shadows would be cast onto the planetary surface and would not move relative to it.

[MKruer]

Speaking of rings, that is one area that could be improved.

Saturn's rings have spokes.
http://en.wikipedia.org/wiki/Saturn


And Neptune's rings are more ring arcs
http://en.wikipedia.org/wiki/Rings_of_Neptune

[Hungry4info]

The rings are not very visible until you're within a few thousand kilometres from the surface, but they become obvious when you're within 1000 or so km.



And the CEL

cel://Follow/Sol:Uranus/2008-01-28T23:3 ... lm=0&ver=2

[selden]

Yup, them's shadows alright!
(or wrong, depending on your *ahem* point of view )

[Chuft-Captain]

Is it true that OpenGL 2.0 is required to see ring, eclipse shadows etc?

... I don't see them (I assume because I haven't got OGL2).

What else won't I see?

[selden]

CC,

Basic can show eclipse shadows cast by Celestia's builtin spheres onto other builtin spheres. (OpenGL v1.1)

MultiTexture adds shadows cast by builtin spheres onto their own rings. (OpenGL v1.2)

Vertex Shaders adds shadows cast by rings onto their parent builtin spheres. (OpenGL v1.4; but not for Nvidia MX series cards.)

OpenGL v2 adds shadows cast by builtin spheres onto 3D models.

Celestia can not yet draw shadows cast by 3D models, or cast between rings and moons.

[Chuft-Captain]

Basic can show eclipse shadows cast by Celestia's builtin spheres onto other builtin spheres. (OpenGL v1.1)

Got it!

MultiTexture adds shadows cast by builtin spheres onto their own rings. (OpenGL v1.2)

Got it!

Vertex Shaders adds shadows cast by rings onto their parent builtin spheres. (OpenGL v1.4; but not for Nvidia MX series cards.)

Perhaps it's time I upgraded my drivers...

Code: Select all

Vendor: Intel
Renderer: Intel Montara-GM
Version: 1.3.0 - Build 4.14.10.3712
Max simultaneous textures: 4
Max texture size: 2048
Point size range: 0.500000 - 10.000000

Supported Extensions:
GL_ARB_multitexture
GL_ARB_texture_border_clamp
GL_ARB_texture_compression
GL_ARB_texture_cube_map
GL_ARB_texture_env_add
GL_ARB_texture_env_combine
GL_ARB_texture_env_dot3
GL_ARB_texture_env_crossbar
GL_ARB_transpose_matrix
GL_EXT_abgr
GL_EXT_bgra
GL_EXT_blend_color
GL_EXT_blend_func_separate
GL_EXT_blend_minmax
GL_EXT_blend_subtract
GL_EXT_clip_volume_hint
GL_EXT_compiled_vertex_array
GL_EXT_cull_vertex
GL_EXT_fog_coord
GL_EXT_packed_pixels
GL_EXT_packed_pixels_12
GL_EXT_rescale_normal
GL_EXT_secondary_color
GL_EXT_separate_specular_color
GL_EXT_stencil_wrap
GL_EXT_texture_compression_s3tc
GL_EXT_texture_env_add
GL_EXT_texture_env_combine
GL_EXT_texture_filter_anisotropic
GL_3DFX_texture_compression_FXT1
GL_IBM_texture_mirrored_repeat
GL_NV_texgen_reflection
GL_WIN_swap_hint

OpenGL v2 adds shadows cast by builtin spheres onto 3D models.

Ain't got it! I suspect my hardware is not capable of OpenGL 2.0

Celestia can not yet draw shadows cast by 3D models, or cast between rings and moons.

I'm looking forward to this....maybe in 1.5.1

Cheers
CC

[BobHegwood]

The rings are not very visible until you're within a few thousand kilometres from the surface, but they become obvious when you're within 1000 or so km.

Well I'll be damned...

Yeah, I see them on my machine too, now that you have pointed them
out via the CEL URL.

Never saw them before though. I guess I never got close enough to
notice. Learn something new every day.

THANKS, Bob

[BobHegwood]

If I understand this thread correctly, Selden, Hungry is saying that the
rendered shadow images simply do NOT change their appearance as
they would realistically as you move your viewpoint around the planet.

Is this correct Hungry? I have also since noticed that when I place a
barely visible ring around Jupiter, the shadow, by far, is rendered much
too darkly, and with NO regard whatsoever to the viewer's location
and/or perspective.

Just some more information from that Grumpy Old Brain-Dead Geezer.

[chris]

If I understand this thread correctly, Selden, Hungry is saying that the
rendered shadow images simply do NOT change their appearance as
they would realistically as you move your viewpoint around the planet.

No, the shadow images shouldn't change their appearance as the viewer moves. The appearance of shadows depends on the position of the light source and shadow casting objects, not the viewer position.

Is this correct Hungry? I have also since noticed that when I place a
barely visible ring around Jupiter, the shadow, by far, is rendered much
too darkly, and with NO regard whatsoever to the viewer's location
and/or perspective.

It's the darkness that's the problem, though that may be a problem with the Jupiter ring texture you're using. What Hungry is saying is that the Uranus and Neptune ring shadows should appear much fuzzier than they do because the apparent size of the Sun is greater than the apparent width of the rings as seen from the surface of Uranus. I haven't verified this myself, but it's likely he's right. And it's true that Celestia doesn't try to render fuzzy ring shadows, so there's at least the potential for a problem. A fairly simple way to make fuzzy shadows is to adjust the mipmap level of detail by an amount determined by the apparent size of the sun, the resolution of the ring texture, and the width of the rings.

--Chris

[selden]

Bob,

Shadows cast on a flat, dull surface shouldn't change much depending on your viewpoint. They do change if the surface is bumpy so that parts of the surface are hidden depending on your viewpoint. I can only guess at how shadows cast into a deep hazy atmosphere would vary depending on your viewpoint.

I think in this case Celestia is essentially emulating the first situation: shadows cast on a smooth, dull surface. Maybe someday it'll be able to do shadows cast into a deep atmosphere.

However, I think that Hungry's point is slightly different.

The sun is a large sphere. When you look at it through the rings of Uranus from the "surface" of Uranus, the rings can be seen as lines across the Sun's circle -- there's sunlight coming directly to you past the rings. Your viewpoint is in the penumbra of their shadows. Since the rings are very narrow, there's no region on the surface of Uranus which where the sunlight is completely blocked. In other words, the rings cast no umbra. As a result, the shadows should be grey, not black, and broader than they're currently being drawn.

My impression from what I vaguely recall Chris writing some time ago is that Celestia's shadow-casting is a compromise. To it right would require what is called "ray tracing", which would need a lot more computation than Celestia can afford to do in the little time it has between frames. However, Celestia does manage to represent both umbra and penumbra when casting circular eclipse shadows by the Moon onto the Earth, so one can hope that an improved technique can be devised for casting ring shadows in a future version of Celestia.

[symaski62]

Basic can show eclipse shadows cast by Celestia's builtin spheres onto other builtin spheres. (OpenGL v1.1)

Got it!

MultiTexture adds shadows cast by builtin spheres onto their own rings. (OpenGL v1.2)

Got it!

Vertex Shaders adds shadows cast by rings onto their parent builtin spheres. (OpenGL v1.4; but not for Nvidia MX series cards.)

Perhaps it's time I upgraded my drivers...

Code: Select all

Vendor: Intel
Renderer: Intel Montara-GM
Version: 1.3.0 - Build 4.14.10.3712
Max simultaneous textures: 4
Max texture size: 2048
Point size range: 0.500000 - 10.000000

Supported Extensions:
GL_ARB_multitexture
GL_ARB_texture_border_clamp
GL_ARB_texture_compression
GL_ARB_texture_cube_map
GL_ARB_texture_env_add
GL_ARB_texture_env_combine
GL_ARB_texture_env_dot3
GL_ARB_texture_env_crossbar
GL_ARB_transpose_matrix
GL_EXT_abgr
GL_EXT_bgra
GL_EXT_blend_color
GL_EXT_blend_func_separate
GL_EXT_blend_minmax
GL_EXT_blend_subtract
GL_EXT_clip_volume_hint
GL_EXT_compiled_vertex_array
GL_EXT_cull_vertex
GL_EXT_fog_coord
GL_EXT_packed_pixels
GL_EXT_packed_pixels_12
GL_EXT_rescale_normal
GL_EXT_secondary_color
GL_EXT_separate_specular_color
GL_EXT_stencil_wrap
GL_EXT_texture_compression_s3tc
GL_EXT_texture_env_add
GL_EXT_texture_env_combine
GL_EXT_texture_filter_anisotropic
GL_3DFX_texture_compression_FXT1
GL_IBM_texture_mirrored_repeat
GL_NV_texgen_reflection
GL_WIN_swap_hint

OpenGL v2 adds shadows cast by builtin spheres onto 3D models.

Ain't got it! I suspect my hardware is not capable of OpenGL 2.0

Celestia can not yet draw shadows cast by 3D models, or cast between rings and moons.

I'm looking forward to this....maybe in 1.5.1

Cheers
CC

http://downloadcenter.intel.com/default ... 2_download

[BobHegwood]

Okay, thanks Selden, Chris...

As usual, I simply just don't know what the Hell I'm talking about.
BIG surprise there, I'll bet.

Look at how much you're teaching me though. When I croak, I'll be
much smarter.

[danielj]

To me,Neptune rings seem arc like:
http://img238.imageshack.us/my.php?imag ... ngsjn8.jpg
But I don??t know if is it a solar ilumination (direction) question...

[Hungry4info]

The sun is a large sphere. When you look at it through the rings of Uranus from the "surface" of Uranus, the rings can be seen as lines across the Sun's circle -- there's sunlight coming directly to you past the rings. Your viewpoint is in the penumbra of their shadows. Since the rings are very narrow, there's no region on the surface of Uranus which where the sunlight is completely blocked. In other words, the rings cast no umbra. As a result, the shadows should be grey, not black, and broader than they're currently being drawn.

This is exactly what I am trying to say. Thank-you, Selden.

As usual, I simply just don't know what the Hell I'm talking about. BIG surprise there, I'll bet.

We all make mistakes, you're trying to contribute, and it's the thought that counts.

Look at how much you're teaching me though. When I croak, I'll be much smarter .

Aww... that's so depressing and fatalistic... but halarious at the same time. You have a very unique personality that I believe adds some life to this forum.

[fsgregs]

Thanks to Chris and to all the devleopers for 1.5.0. The program is just spectacular.

If this is a thread regarding what should be done for 1.5.1, I have only begun evaluating 1.5.0, but three things come to mind quickly:

1. Near galaxies such as Andromeda, Triangulum and most prominently, Sagittarius, are much too bright when viewed from our solar system. This has been commented before in other threads. We need to drop their apparent magnitude at least 50%.

2. Dwarf planets should be given their own designation (orbit lines, colors, etc.) so that we have Planets, Dwarf Planets, and Asteroids and comets.

3. The zenith is still much darker than it is on Earth, when looking up at mid-day. Here is a screenshot:


By fsgregs

As can be seen, it is almost blue-black as seen from the ground. It needs to be lightened up at least 30% or so.