Celestia shade off; color problem

Report bugs, bug fixes and workarounds here.
Topic author
Psykotik
Posts: 233
Joined: 02.11.2002
With us: 21 years 8 months
Location: Geneva
Contact:

Celestia shade off; color problem

Post #1by Psykotik » 14.08.2003, 03:32

Celestia has been the first program to force me to buy a new graphic card. I'm not a gamer, so I thought I was immune to this kind of expenses. What a fool :)

So that's 4 weeks I own a new graphic card, a MSI GeForce FX 5600, with the latest brand drivers from MSI (nvidia 44.03).

Exploring ALL the space of Celestia days by days, I was surprised to discover a bug I never seen before (I can't remember exactly, but I think with the first card, a GeForce 2 MX 200, celestia worked correctly).

The bug is I can't see anymore the shade off. The colours change in a unvarnished way, and the result... look at result :

The first exemple, with a Rass' planet :
Image

The second exemple, with a Rass' (again ?) pulsar
Image

I've tried disabling anti-alias or anisotropic options, with no luck.

Has every anyone seen this kind of problem ?

Avatar
John Van Vliet
Posts: 2940
Joined: 28.08.2002
With us: 21 years 10 months

re

Post #2by John Van Vliet » 14.08.2003, 10:12

Did you get the driver from Microsoft using the auto update or check for update on the Microsoft web page

or

did you get the driver from nVidia 's web page

because MicroSoft's driver SUCKES!!!-----DONT USE IT

Topic author
Psykotik
Posts: 233
Joined: 02.11.2002
With us: 21 years 8 months
Location: Geneva
Contact:

Post #3by Psykotik » 14.08.2003, 11:06

Hello John,

As I said, I'm not using the driving from Microsoft, nor from Nvidia. I got the drivers from MSI, my card constructor.

danielj

Re: Celestia shade off; color problem

Post #4by danielj » 14.08.2003, 15:30

I would like to know how can you get the pulsar textures working.I tried,but still didn?t work.For example,PSR 1257

Topic author
Psykotik
Posts: 233
Joined: 02.11.2002
With us: 21 years 8 months
Location: Geneva
Contact:

Post #5by Psykotik » 15.08.2003, 19:11

Nobody has ever had this kind of problem ?

Daniel, be sure to modify the file psr1257.stc, adding "PSR1257" after "300002". It'll define the name of the Star.

Topic author
Psykotik
Posts: 233
Joined: 02.11.2002
With us: 21 years 8 months
Location: Geneva
Contact:

Post #6by Psykotik » 19.02.2004, 13:57

6 months later, no one has ever had this problem ? Because I'm still having this bug :evil:

Avatar
selden
Developer
Posts: 10190
Joined: 04.09.2002
With us: 21 years 9 months
Location: NY, USA

Post #7by selden » 19.02.2004, 14:25

Psykotik,

I think maybe once in a while I sometimes see something like what you describe: objects look like "Emissive true" is enabled, even though it isn't.

I do not see it consistantly, and it goes away if I set "Ambient light" to off.
Selden

Topic author
Psykotik
Posts: 233
Joined: 02.11.2002
With us: 21 years 8 months
Location: Geneva
Contact:

Post #8by Psykotik » 19.02.2004, 15:01

Hi Selden,

I don't know anything about "emissive true", but I can assure you I constantly have the textures displayed as shown. Moreover, ambiant light is always switched to off...

I was hoping that new nvidia drivers will resolve the problem, but that's 6 months of drivers I've installed :lol:

If I remember right, Jeam and other forum's user are having exactly the same trouble.

Avatar
selden
Developer
Posts: 10190
Joined: 04.09.2002
With us: 21 years 9 months
Location: NY, USA

Post #9by selden » 19.02.2004, 15:11

Can you provide a "good" picture that can be compared with your "bad" picture?

Unfortunately, since you only provided the "bad" picture, I don't know what that particular object is supposed to look like.

You might try turning Ambient Light on (low or medium) and then off again. Does that make a difference?
Selden

Darkmiss
Posts: 1059
Joined: 20.08.2002
With us: 21 years 10 months
Location: London, England

Post #10by Darkmiss » 19.02.2004, 15:36

I think he is talking about the bands appearing within the emitted light
rather then the light shading out smoothly.
CPU- Intel Pentium Core 2 Quad ,2.40GHz
RAM- 2Gb 1066MHz DDR2
Motherboard- Gigabyte P35 DQ6
Video Card- Nvidia GeForce 8800 GTS + 640Mb
Hard Drives- 2 SATA Raptor 10000rpm 150GB
OS- Windows Vista Home Premium 32

Avatar
selden
Developer
Posts: 10190
Joined: 04.09.2002
With us: 21 years 9 months
Location: NY, USA

Post #11by selden » 19.02.2004, 15:47

One of the causes of bands of equal shades of color is configuring the Display Properties "Settings" to "medium (16bit)" colors or fewer. You must set your display to full color (24 or 32 bit) to prevent them.

If your system is set for 32bit color and you're still seeing banding, then something else is going on -- maybe another program is interfering?
Selden

Topic author
Psykotik
Posts: 233
Joined: 02.11.2002
With us: 21 years 8 months
Location: Geneva
Contact:

Post #12by Psykotik » 19.02.2004, 15:56

Darkmiss has clarified my query.

Here is a good picture, taken from the Jeam's site (before he had the same problem).

Image Image
Last edited by Psykotik on 30.07.2004, 18:40, edited 2 times in total.

Topic author
Psykotik
Posts: 233
Joined: 02.11.2002
With us: 21 years 8 months
Location: Geneva
Contact:

Post #13by Psykotik » 19.02.2004, 16:29

Selden,

That's what I've always thought : a color problem.

I've checked the Display Properties : set to 32 bits, and I've checked again with another program (Sandra 2003) to be totally sure.

There is no interfering program, since I've tried closing all tasks before launching Celestia.

I'm really thinking about a bug between Celestia and my video card (not so new, nowadays...). Today, I use nvidia driver 52.14.

Anyway, thanks for your time Selden.

Edit : if Chris read by chance this topic

Code: Select all

Vendor: NVIDIA Corporation

Renderer: GeForce FX 5600/AGP/3DNOW!

Version: 1.4.1

Max simultaneous textures: 4

Max texture size: 4096



Supported Extensions:

GL_ARB_depth_texture

GL_ARB_fragment_program

GL_ARB_imaging

GL_ARB_multisample

GL_ARB_multitexture

GL_ARB_occlusion_query

GL_ARB_point_parameters

GL_ARB_point_sprite

GL_ARB_shadow

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_mirrored_repeat

GL_ARB_transpose_matrix

GL_ARB_vertex_buffer_object

GL_ARB_vertex_program

GL_ARB_window_pos

GL_S3_s3tc

GL_EXT_texture_env_add

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_compiled_vertex_array

GL_EXT_draw_range_elements

GL_EXT_fog_coord

GL_EXT_multi_draw_arrays

GL_EXT_packed_pixels

GL_EXT_paletted_texture

GL_EXT_point_parameters

GL_EXT_rescale_normal

GL_EXT_secondary_color

GL_EXT_separate_specular_color

GL_EXT_shadow_funcs

GL_EXT_shared_texture_palette

GL_EXT_stencil_two_side

GL_EXT_stencil_wrap

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

GL_EXT_texture_lod_bias

GL_EXT_texture_object

GL_EXT_vertex_array

GL_HP_occlusion_test

GL_IBM_rasterpos_clip

GL_IBM_texture_mirrored_repeat

GL_KTX_buffer_region

GL_NV_blend_square

GL_NV_copy_depth_to_color

GL_NV_depth_clamp

GL_NV_fence

GL_NV_float_buffer

GL_NV_fog_distance

GL_NV_fragment_program

GL_NV_half_float

GL_NV_light_max_exponent

GL_NV_multisample_filter_hint

GL_NV_occlusion_query

GL_NV_packed_depth_stencil

GL_NV_pixel_data_range

GL_NV_point_sprite

GL_NV_primitive_restart

GL_NV_register_combiners

GL_NV_register_combiners2

GL_NV_texgen_reflection

GL_NV_texture_compression_vtc

GL_NV_texture_env_combine4

GL_NV_texture_expand_normal

GL_NV_texture_rectangle

GL_NV_texture_shader

GL_NV_texture_shader2

GL_NV_texture_shader3

GL_NV_vertex_array_range

GL_NV_vertex_array_range2

GL_NV_vertex_program

GL_NV_vertex_program1_1

GL_NV_vertex_program2

GL_NVX_ycrcb

GL_SGIS_generate_mipmap

GL_SGIS_texture_lod

GL_SGIX_depth_texture

GL_SGIX_shadow

GL_SUN_slice_accum

GL_WIN_swap_hint

WGL_EXT_swap_control

Topic author
Psykotik
Posts: 233
Joined: 02.11.2002
With us: 21 years 8 months
Location: Geneva
Contact:

Post #14by Psykotik » 04.03.2004, 13:07

I'm pretty sure that's a color problem : switching from 32-bits to 16-bits in Celestia will render EXACTLY the same image. No change; I'm stuck with 16-bits :-(

I've searched around the web, but nothing found... Chris, do you have any idea ?

Topic author
Psykotik
Posts: 233
Joined: 02.11.2002
With us: 21 years 8 months
Location: Geneva
Contact:

Post #15by Psykotik » 30.03.2004, 19:13

I don't know if it does really matters, since only Selden was interested in, but I've found the source of the problem : Celestia it-self.

Going back to 1.2.5pre6 worked like a charm; the colors were correctly displayed. Every version since 1.3 doesn't work. It seems like something is broken.

don
Posts: 1709
Joined: 12.07.2003
With us: 20 years 11 months
Location: Colorado, USA (7000 ft)

Post #16by don » 31.03.2004, 00:33

Hopefully Chris will have some idea as to why you have problems with your graphics card and other folks don't. And then be able to make a modification of some kind also.
-Don G.
My Celestia Scripting Resources page

Avatar: Total Lunar Eclipse from our back yard, Oct 2004. Panasonic FZ1 digital camera (no telescope), 36X digital zoom, 8 second exposure at f6.5.

chris
Site Admin
Posts: 4211
Joined: 28.01.2002
With us: 22 years 5 months
Location: Seattle, Washington, USA

Post #17by chris » 01.04.2004, 11:21

I don't observe any problems with the blue planet, but I do see the banding in the pulsar beam on my system. However, I don't see how you could not get banding with this model. The beam is composed of segments which decrease in opacity the further they are from the star. The important thing to observe is that the opacity is constant over each segment--it's not interpolated, and beam isn't subdivided finely enough for the transitions between segments to be undetectable.

--Chris

Topic author
Psykotik
Posts: 233
Joined: 02.11.2002
With us: 21 years 8 months
Location: Geneva
Contact:

Post #18by Psykotik » 02.04.2004, 01:07

Hi Chris,

If you look carefully at the two pictures on february 19th, you'll see Celestia render differently the pulsar (BTW, Mira too, but not so evident).

And I can affirm older celestia versions didn't manage these textures in this way... it was smoother !

I can provide a bigger picture of what these pictures should look, if needed.

Thanks for your time.

DBrady
Posts: 66
Joined: 14.07.2003
With us: 20 years 11 months
Location: Sydney

Post #19by DBrady » 27.07.2004, 23:32

I recently noticed that i was suffering from this problem(probably for quite a while). I think there may be alot of other people suffering from it too but they just dont know. How can you tell?

Well,
In the standard distribution go to earth with its default texture and look at south america. If it looks like this...

Image

...then you've got a problem. It should look like this...

Image

Notice that there is no banding. The cause of the problem seems to be the nvidia drivers. Defining the colour depth as equal to the desktop colour depth appears to just render everything as 16-bit even though your desktop might be 32-bit. Furthermore the nvidia drivers dont give much control over many opengl features in their drivers so you'll have to download geforce tweak from guru3d...

http://www.guru3d.com/geforcetweakutility/

Download, install and run.
Go to the opengl tab,
navigate to the bottom,
uncheck the box marked def. opengl render quality,
and finally quit.
Dont mess with anything else unless you know what your doing.

Image

This should fix it.

If you find you are suffering from this problem could you post a reply. If alot of people respond then it might be worth making it a sticky.
Slan

Avatar
selden
Developer
Posts: 10190
Joined: 04.09.2002
With us: 21 years 9 months
Location: NY, USA

Post #20by selden » 27.07.2004, 23:46

DBrady,

Well, I hate to say this, but it seems that I'm not very observant. I don't see any obvious differences between the two pictures, even though my desktop is set to 32bits.

What should I be looking for?

Added slightly later:

After staring for a while, I see that the northern parts of South America in the upper picture tend to have more light green than in the lower picture. Also, the regions of different colors of green tend to be more distinct in the upper picture than in the lower one.

I don't see those symptoms on my home system: it looks like the lower picture.

System specs:
256MB 500MHz P3, WinXP Pro SP1
128MB GF4 Ti4200, drivers v61.76

Under Performance and Quality settings:
Image settings: Quality
Antialiasing, Anisotropic and Vertical sync are all "application controlled"
Color correction: "standard mode" with all sliders centered, except that Vibrance is off.

Celestia v1.3.1-1 final
Celestia v1.3.2pre10
Celestia v1.3.2pre (cvs)
Selden


Return to “Bugs”