Eclipses of Jupiter Satellites

[piellepi]

Hi Celestians!
Yesterday I was traveling near Jupiter and its Galilean satellites. I was watching "Io" (Enter, Io, Enter, G) and I accelerated the time (L) up to 1000x. Then Io disappeared into Jupiter's shadow, so I went back (J) just to witness the eclipse, but something was wrong: I did a snapshot of the problem.

The eclipse shadow is completely wrong!
The same happens for Europa and Ganimede:






I'm using Celestia 1.4.0pre6.

This is my graphic card (with the newest drivers dowloaded from ATI)
Vendor: ATI Technologies Inc.
Renderer: RADEON 9600 x86/MMX/3DNow!/SSE
Version: 2.0.5014 WinXP Release
Max simultaneous textures: 8
Max texture size: 2048

Supported Extensions:
GL_ARB_multitexture
GL_EXT_texture_env_add
GL_EXT_compiled_vertex_array
GL_S3_s3tc
GL_ARB_depth_texture
GL_ARB_fragment_program
GL_ARB_fragment_program_shadow
GL_ARB_fragment_shader
GL_ARB_multisample
GL_ARB_occlusion_query
GL_ARB_point_parameters
GL_ARB_point_sprite
GL_ARB_shader_objects
GL_ARB_shading_language_100
GL_ARB_shadow
GL_ARB_shadow_ambient
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_crossbar
GL_ARB_texture_env_dot3
GL_ARB_texture_mirrored_repeat
GL_ARB_transpose_matrix
GL_ARB_vertex_blend
GL_ARB_vertex_buffer_object
GL_ARB_vertex_program
GL_ARB_vertex_shader
GL_ARB_window_pos
GL_ATI_draw_buffers
GL_ATI_element_array
GL_ATI_envmap_bumpmap
GL_ATI_fragment_shader
GL_ATI_map_object_buffer
GL_ATI_separate_stencil
GL_ATI_texture_env_combine3
GL_ATI_texture_float
GL_ATI_texture_mirror_once
GL_ATI_vertex_array_object
GL_ATI_vertex_attrib_array_object
GL_ATI_vertex_streams
GL_ATIX_texture_env_combine3
GL_ATIX_texture_env_route
GL_ATIX_vertex_shader_output_point_size
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_draw_range_elements
GL_EXT_fog_coord
GL_EXT_multi_draw_arrays
GL_EXT_packed_pixels
GL_EXT_point_parameters
GL_EXT_rescale_normal
GL_EXT_secondary_color
GL_EXT_separate_specular_color
GL_EXT_shadow_funcs
GL_EXT_stencil_wrap
GL_EXT_texgen_reflection
GL_EXT_texture3D
GL_EXT_texture_compression_s3tc
GL_EXT_texture_cube_map
GL_EXT_texture_edge_clamp
GL_EXT_texture_env_combine
GL_EXT_texture_env_dot3
GL_EXT_texture_filter_anisotropic
GL_EXT_texture_lod_bias
GL_EXT_texture_mirror_clamp
GL_EXT_texture_object
GL_EXT_texture_rectangle
GL_EXT_vertex_array
GL_EXT_vertex_shader
GL_HP_occlusion_test
GL_NV_blend_square
GL_NV_occlusion_query
GL_NV_texgen_reflection
GL_SGI_color_matrix
GL_SGIS_generate_mipmap
GL_SGIS_multitexture
GL_SGIS_texture_border_clamp
GL_SGIS_texture_edge_clamp
GL_SGIS_texture_lod
GL_SUN_multi_draw_arrays
GL_WIN_swap_hint
WGL_EXT_extensions_string
WGL_EXT_swap_control

Somewhere I read that eclipses for "cmod-ed" satellites don't work: in my "solarsys.ssc" there is no "3ds" nor "cmod" mesh for those satellites, so they would be eclipsed well.
Even cycling with ctrl-V gives no better result: the "OpenGL v2.0" (sorry, but I don't remember its real name! ) gives a completely black satellite, while the other values ("basic", "multitextures" and so on) give the same bug as the screenshot.
Do you have any idea?
Thanks a lot

Ciao!
Pierluigi

[selden]

Unfortunately, your screenshots are not visible.

It would be very helpful if you could provide Cel:// URLs from the failing viewpoints. On my system I don't see anything wrong with the eclipses of Io when going forward or backward in time at high speed.

[symaski62]

piellepi image !

copie/colle

http://piellepi.web-gratis.net/bug/io.jpg

http://piellepi.web-gratis.net/bug/europa.jpg

http://piellepi.web-gratis.net/bug/ganimede.jpg

[selden]

Errore 403
NON SI POSSIEDONO I PERMESSI PER ACCEDERE AL FILE.

CONOSCI GIA IL MONDO DI WEBGRATIS?

SCOPRILO CLICCANDO QUI

[symaski62]

piellepi

[piellepi]

OOPS!
Sorry but I don't know why my pictures coudn't be seen...
Perhaps the free site has restricted the access.
I think symaski62 was able to dowload them!
If you can see the pictures, what do you think about the problem?

For Selden
Which kind of shadow do you see at the time of my snapshot?
When I made the snapshot I just paused Celestia with "space", but time was running at 1x, not at high speed.
In fact, if time is at 1000x, the only thing you see is a disappearing and a reappearing satellite!
If necessary I will try to create a cel URL script , but I have never tried it

Thanks a lot
Pierluigi

[selden]

The shadows look normal on my system. Not like the ones above.

To record a Cel:// Url, type Ctrl-C (or Ctrl-Insert) to Celestia.

To insert that Cel:// URL into a document, type a Ctrl-V.

[piellepi]

Selden,
I tried at office and I had the same problem with the shadow on Io (I had no more time to try the other satellites! ).
Here I have the same Celestia version (1.4.0pre6) but with a completely different video card (a poor Intel card ) so I suspect that the problem is the version.
Which version of Celestia do you use?

Ciao e Grazie!
Pierluigi

[Alessandro]

Hi,

I have the same problem as Pierluigi

cel://Follow/Sol:Jupiter:Io/2005-04-27T ... 9&lm=49158


Ciao
Alessandro

[Hannibalector]

the moons looks like cookies. i got the same problem adn just install celestia over my old. MAKE A BACKUP OF THE DATA-FOLDER!

[hjw]

Well, with certain render paths I see the same.
My system: Linux GF 6600GT, Driver 66.29 (the new one has bugs with other software), celestia cvs

I see the same with render paths
- Basic
- Multitexture
- OpenGL vertex program
- OpenGL vertex program/NVidia combiners

Io is black with
- NVidia GeForce FX
- OpenGL 2.0
because it the eclipse begins earlier and ends later (and is displayed without
problems) with GeForce FX / OpenGL 2.0


Horst.

[selden]

Using your URL, I do see the cookie-cutter edges. It's due to the low resolution of the shadows being cast in the older render paths. Shadow edges are smooth in the OpenGL 2.0 render path on my FX5200 at work and on my GF6600GT at home.

You can improve the edges of eclipse shadows by editing celestia.cfg, but your system will run more slowly. Change the line

Code: Select all

  EclipseTextureSize     128

Increase 128 to a larger power-of-two.

There are other parameters there which improve orbits and shadows cast on or by rings. Be sure to read the comments in celestia.cfg which describe them.

[piellepi]

Hi Selden!
Thanks for your suggestions, but unfortunately even a value of 1024 gives exactly the same result, at least here at office: I'll try later at home with my ATI 9600 card.
I made more investigations about the bug and I think the problem is related to shadow rendering routines: I captured 3 more snapshots of the event, each time zooming towards Io by means of the mouse wheel.
http://piellepi.web-gratis.net/bug/zoom1.jpg
http://piellepi.web-gratis.net/bug/zoom2.jpg
http://piellepi.web-gratis.net/bug/zoom3.jpg
I'm sorry if links don't work

If you try zooming yourself, you can see (like in the third picture) that Jupiter's shadow is created programmatically: in fact you can see that the shadow has different levels of darkness, from transparency to black.
I am perplexed by the shape of the shadow itself: what would be an ellipse becomes a "cookie-cutter tool"! But I'm not using OpenGL 2.0: If I press ctrl-v to enter "OpenGL 2.0", Io becomes completely black.

For other ATI-Celestians, I also tried to change the properties of OpenGL programming by means of the ATI Control Panel: passing from "Highest Performance/Low quality" to "Highest quality/low performance" there is no significant change in Celestia. Has anyone obtained better results?

Ciao
Pierluigi

[piellepi]

OOPS
link don't work!
I'm sorry

try zooming towards Io...

Pierluigi

[bdm]

I have seen the same problem with Celestia version 1.3.2 with eclipse shadows. I have also seen a similar problem with the shadow of Saturn on its rings.

[selden]

If you have an Nvidia 6200 or better or an ATI 9500 or better, you should consider installing one of the Celestia v1.4.0 prereleases. Smooth shadows are created using OpenGL v2.0 routines. The rough-edged shadows won't be improved for other cards. Sorry.

[bdm]

If you have an Nvidia 6200 or better or an ATI 9500 or better, you should consider installing one of the Celestia v1.4.0 prereleases. Smooth shadows are created using OpenGL v2.0 routines. The rough-edged shadows won't be improved for other cards. Sorry.

I have found that typing Ctrl-V once selects the correct driver, and this makes the shadows nice and smooth. Is there a way to specify the default rendering to use without having to type Ctrl-V?

[selden]

The OpenGL 2.0 code is still being developed and has some bugs. It'll probably be the default for those cards which can support it in the next prerelease or maybe the one after that.