Star 'Points' issue

[Drydareelin]

I usually use the 'Fuzzy Points' option for stars but I prefer the 'Points' option
Whenever I apply it, I get these weird squares in the sky.

Laptop Specs:

Novatech, Windows 7 64bit
Processor: Intel(R) Pentium(R) CPU B950 @ 2.10 GHz
RAM: 4.00GB

Hope that is enough information ._.

Thanks

[selden]

That symptom normally is caused by a buggy OpenGL graphics driver.

Please update your computer's graphics driver to the version currently available from the manufacturer of your graphics hardware (usually Nvidia, AMD or Intel). Some laptops need drivers provided by the manufacturer of the laptop, but that's becoming less common.

You can determine what graphics device Celestia thinks your computer has by running Celestia
and selecting Help --> OpenGL info. The first two lines show the manufacturer and the device.

If your computer's graphics drivers are installed properly, those lines will agree with the computer's own hardware information, although perhaps with a few additional details.

You can determine the computer's own information about manufacturer and model of your graphics hardware by selecting this sequence of menu items:

Start menu --> (right mouse click) Computer --> properties

in the new window:
Device Manager --> (click on |> in front of) Display adapters

The graphics device(s) will be listed below Display adapters..

[Drydareelin]

Thank you for replying; before I posted that, I should've added that I already went onto the Intel website and checked for any latest drivers - all mine are up to date.
Anyway, I followed what you said, and all of the information co-implies with my laptop's information.
So, I guess it might still be an openGL issue, but still can't figure it out.

Rest of the openGL information:

Vendor: Intel
Renderer: Intel(R) HD Graphics Family
Version: 3.1.0 - Build 8.15.10.2559
GLSL version: 1.40 - Intel Build 8.15.10.2559
Max simultaneous textures: 8
Max texture size: 8192
Point size range: 0.500000 - 10.000000

Supported Extensions:

GL_EXT_blend_minmax
GL_EXT_blend_subtract
GL_EXT_blend_color
GL_EXT_abgr
GL_EXT_texture3D
GL_EXT_clip_volume_hint
GL_EXT_compiled_vertex_array
GL_SGIS_texture_edge_clamp
GL_SGIS_generate_mipmap
GL_EXT_draw_range_elements
GL_SGIS_texture_lod
GL_EXT_rescale_normal
GL_EXT_packed_pixels
GL_EXT_texture_edge_clamp
GL_EXT_separate_specular_color
GL_ARB_multitexture
GL_EXT_texture_env_combine
GL_EXT_bgra
GL_EXT_blend_func_separate
GL_EXT_secondary_color
GL_EXT_fog_coord
GL_EXT_texture_env_add
GL_ARB_texture_cube_map
GL_ARB_transpose_matrix
GL_ARB_texture_env_add
GL_IBM_texture_mirrored_repeat
GL_EXT_multi_draw_arrays
GL_NV_blend_square
GL_ARB_texture_compression
GL_3DFX_texture_compression_FXT1
GL_EXT_texture_filter_anisotropic
GL_ARB_texture_border_clamp
GL_ARB_point_parameters
GL_ARB_texture_env_combine
GL_ARB_texture_env_dot3
GL_ARB_texture_env_crossbar
GL_EXT_texture_compression_s3tc
GL_ARB_shadow
GL_ARB_window_pos
GL_EXT_shadow_funcs
GL_EXT_stencil_wrap
GL_ARB_vertex_program
GL_EXT_texture_rectangle
GL_ARB_fragment_program
GL_EXT_stencil_two_side
GL_ATI_separate_stencil
GL_ARB_vertex_buffer_object
GL_EXT_texture_lod_bias
GL_ARB_occlusion_query
GL_ARB_fragment_shader
GL_ARB_shader_objects
GL_ARB_shading_language_100
GL_ARB_texture_non_power_of_two
GL_ARB_vertex_shader
GL_NV_texgen_reflection
GL_ARB_point_sprite
GL_ARB_fragment_program_shadow
GL_EXT_blend_equation_separate
GL_ARB_depth_texture
GL_ARB_texture_rectangle
GL_ARB_draw_buffers
GL_ARB_color_buffer_float
GL_ARB_half_float_pixel
GL_ARB_texture_float
GL_ARB_pixel_buffer_object
GL_EXT_framebuffer_object
GL_ARB_draw_instanced
GL_ARB_half_float_vertex
GL_EXT_draw_buffers2
GL_WIN_swap_hint
GL_EXT_texture_sRGB
GL_ARB_multisample
GL_EXT_packed_float
GL_EXT_texture_shared_exponent
GL_ARB_texture_rg
GL_ARB_texture_compression_rgtc
GL_NV_conditional_render
GL_EXT_texture_swizzle
GL_ARB_sync
GL_ARB_framebuffer_sRGB
GL_EXT_packed_depth_stencil
GL_ARB_depth_buffer_float
GL_EXT_transform_feedback
GL_EXT_framebuffer_blit
GL_EXT_framebuffer_multisample
GL_ARB_framebuffer_object
GL_EXT_texture_array
GL_EXT_texture_integer
GL_ARB_map_buffer_range
GL_EXT_texture_snorm
GL_INTEL_performance_queries
GL_ARB_copy_buffer
GL_ARB_sampler_objects
GL_NV_primitive_restart
GL_ARB_seamless_cube_map
GL_ARB_uniform_buffer_object
GL_ARB_depth_clamp
GL_ARB_vertex_array_bgra
GL_ARB_draw_elements_base_vertex
GL_ARB_fragment_coord_conventions
GL_EXT_gpu_program_parameters
GL_ARB_compatibility
GL_ARB_vertex_array_object

[selden]

I wouldn't expect them to, but do the squares change if you change Celestia's "Render Path"?
I'm wondering if it does something different for Basic when compared to OpenGL 2.0.

Type a Ctrl-V several times to step through the paths Celestia thinks are available for your hardware.

[Drydareelin]

Ah it worked It changed from 'openGL 2.0' to 'Basic' and the squares dissapeared

OpenGL Vertex Progam produced the squares.
Multitexture removes the squares.

Thank you so much for the help Do you know what the issue was in the end?

[t00fri]

Ah it worked It changed from 'openGL 2.0' to 'Basic' and the squares dissapeared

OpenGL Vertex Progam produced the squares.
Multitexture removes the squares.

Thank you so much for the help Do you know what the issue was in the end?

But the 'Basic' path is not what you should be aiming at. A number of Celestia features are only integrated for the OGL 2.0 path.

Could you make this further test, please:
Return to the OGL 2.0 rendering path and try all star styles successively (via CTRL+S).
Which ones work correctly and which ones give squares or other failures?

Since each star style is rendered in a different manner, this test might allow further conclusions about your problem.

I note that you have a fairly recent version 3.1.0 of OGL and the GLSL shader version is also 1.4.0.
The "classical" one being 1.2.0 and OGL 2.1.x. There can always be bugs in the driver despite (or just because of) it's recent version. Did you try installing a considerably older driver version that e.g. still has
OGL 2.x and GLSL = 1.2.0?

Also did you make sure that this latest driver version actually supports your graphics chip fully?

With OGL 2.1.x and GLSL =1.2.0 you certainly should not be seing squares instead of point stars.
We have a very large amount of OGL testing in this case.

Fridger

[Drydareelin]

The OGL2.0 works o all except the 'Points' star option.
My laptop is fairly new (I got it for xmas) and had needed no new updates since then... basically, I have it the way I got it.
I am worried downgrading my graphics driver would affect other programs though?

[W0RLDBUILDER]

I've had this same issue. I use fuzzy points or scaled disks.

[t00fri]

The OGL2.0 works o all except the 'Points' star option.
My laptop is fairly new (I got it for xmas) and had needed no new updates since then... basically, I have it the way I got it.
I am worried downgrading my graphics driver would affect other programs though?

Of course, downgrading would only be temporary, to find out whether there is a bug in your driver.
You can find some general instruction on how to change graphics drivers properly in this forum

viewtopic.php?f=15&t=12440
or by googling.

The genreal rules also apply for non-NVIDIA drivers.

Fridger

[John Van Vliet]

--- edit ---

[t00fri]

you might want to grab a copy from Intel and replace the one MS installed

While the MS habits are well known, in this case I doubt it, since we know his OGL specs (cf above)
It's a quite recent version (OGL 3.1.0 and shader language version GLSL = 1.4.0) which I don't think smells of MS...

Fridger

[Chuft-Captain]

I wouldn't expect them to, but do the squares change if you change Celestia's "Render Path"?
I'm wondering if it does something different for Basic when compared to OpenGL 2.0.

Type a Ctrl-V several times to step through the paths Celestia thinks are available for your hardware.

FYI,

I've just replicated this issue on my new hardware (as it's only now with the new hardware that I'm able to use Open-GL 2.0 render paths for the first time! )

For me, the exact circumstances are:

1. All Render paths OK when using Fuzzy, or Scaled Disks.
2. When using "Points"... Basic and Multi-Texture are OK. The problem only occurs in Open-GL Vertex, and Open-GL 2.0

NOTE: that I have just finished building this system and installing OS, so still haven't got latest drivers and Windows Updates, however the Graphics I'm using at present is the new integrated Intel 4000 iGPU in the latest Ivy-Bridge chips (i7-3770K in my case), so there may not be any driver updates yet, as it's only just been released.

CC

[t00fri]

Are you guys certain that your Celestia version does NOT (accidentally?) include ChrisL's 'shader star' patch that was also circulated here some time ago?? Some users compiled and offered celestia 1.6.1 including that patch.

The patch hooks precisely at the 'point stars' style but renders the stars in a completely different manner. It showed square stars for some graphics cards.

Note that this patch is not officially supported.

Fridger

[Chuft-Captain]

I'm using a pretty standard 1.6.0 release ... Executable size: 2,835,968 bytes; Dated: 9/7/2009 (That's 7/9/2009 for the Yanks).
I haven't downloaded any special offerings for a long time.

FWIW, here's the contents of my shaders folder:

shaders.JPG

[t00fri]

I'm using a pretty standard 1.6.0 release ... Executable size: 2,835,968 bytes; Dated: 9/7/2009 (That's 7/9/2009 for the Yanks).
I haven't downloaded any special offerings for a long time.

FWIW, here's the contents of my shaders folder:

shaders.JPG

CC,

the shaders/ folder is not used anymore for recent shader code. It only contains older shaders. The more recent GLSL star shaders reside directly near the top of render.cpp.

Code: Select all

// Shader for drawing star PSFs
// A star is drawn as the sum of a Gaussian PSF and a glare function
// G(x). Currently, G(x) = 1/(kx^3 + 1), where k is the glare decay constant.
#define STRINGIFY(x) #x

static const string StarVertexShader(STRINGIFY(

uniform vec2 viewportDimensions;\n
uniform float magRange;\n
uniform float psfSigma2;\n
uniform float glare;\n
uniform float glareDecay;\n
uniform float exposure;\n
const float minVisiblePixelValue = 1.0 / 500.0;\n
attribute float magnitude;
varying vec2 starCenter;\n
varying vec4 color;\n
varying float brightness;\n
\n
void main()\n
{\n
    vec4 projectedPos = gl_ModelViewProjectionMatrix * gl_Vertex;\n
    vec2 ndc = projectedPos.xy / projectedPos.w;\n
    float b = pow(2.512, -magnitude * magRange);\n
\n
    float r2 = -log(minVisiblePixelValue / (exposure * b)) * 2.0 * psfSigma2;\n
    float rg2 = (exposure * glare * b / minVisiblePixelValue - 1.0) / glareDecay;\n
\n
    starCenter = (ndc + 1.0) * 0.5 * viewportDimensions;\n
    color = gl_Color;\n
    brightness = b;\n
    gl_PointSize = 2.0 * sqrt(max(r2, rg2));\n
    gl_Position = projectedPos;\n
}\n

));


The corresponding Fragment shader follows. Both Vertex and Fragment shaders are encoded as strings, using the STRINGYFY macro. If you can't locate this shader, ChrisL's patch is absent.

Why don't you install the 1.6.1 release which lacks MANY bugs compared to 1.6.0?
It only takes 1 minute or so

Fridger