Eros

Discussion forum for Celestia developers; topics may only be started by members of the developers group, but anyone can post replies.
Avatar
Cham M
Posts: 4324
Joined: 14.01.2004
Age: 60
With us: 20 years 9 months
Location: Montreal

Post #41by Cham » 21.08.2006, 10:38

How do you map a texture on this one ? When I try to define some texture with SSC file, I get a "flashing" asteroid and the texture isn't showing. The object gets a weird flashing color while moving around.
"Well! I've often seen a cat without a grin", thought Alice; "but a grin without a cat! It's the most curious thing I ever saw in all my life!"

Avatar
selden
Developer
Posts: 10192
Joined: 04.09.2002
With us: 22 years 1 month
Location: NY, USA

Post #42by selden » 21.08.2006, 12:01

Cham,

As Chris mentioned, these models don't have any texture coordinates, so textures can't be applied to them.

Placing a texture on a model requires the model to include texture placement information for each of the model's vertices. Those coordinates tell the graphics hardware what location within the texture should be associated with each vertex of the model.

One concern that I have is that the use of a different model architecture might require the use of a different projection for the surface texture images. The algorithm for mapping a simple cylindrical projection onto an octahedral tessellation won't be trivial. Presumably that's one reason the texture coordinates for the new Eros model aren't available yet :)
]
Selden

Topic author
chris
Site Admin
Posts: 4211
Joined: 28.01.2002
With us: 22 years 9 months
Location: Seattle, Washington, USA

Post #43by chris » 21.08.2006, 15:14

selden wrote:One concern that I have is that the use of a different model architecture might require the use of a different projection for the surface texture images. The algorithm for mapping a simple cylindrical projection onto an octahedral tessellation won't be trivial. Presumably that's one reason the texture coordinates for the new Eros model aren't available yet :)


It's actually not that difficult. I'll have a version with texture coordinates ready very soon. The tricky problem with octahedral tessellation for planets is making texture splitting work; I've got a scheme in mind, but it'll be some time before I'll have a chance to implement it.

--Chris

ElChristou
Developer
Posts: 3776
Joined: 04.02.2005
With us: 19 years 8 months

Post #44by ElChristou » 21.08.2006, 15:23

chris wrote:...I applied a wide Gaussian blur to the Eros LIDAR data, subtracted this from the original data, and applied it to a sphere...


You are talking of displacement here, right?
Image

Topic author
chris
Site Admin
Posts: 4211
Joined: 28.01.2002
With us: 22 years 9 months
Location: Seattle, Washington, USA

Post #45by chris » 21.08.2006, 15:40

ElChristou wrote:
chris wrote:...I applied a wide Gaussian blur to the Eros LIDAR data, subtracted this from the original data, and applied it to a sphere...

You are talking of displacement here, right?


Yes. Most of the asteroid shape models in raw form are essentially displacement maps--distance from center, regularly sampled over longitude and latitude.

--Chris

ElChristou
Developer
Posts: 3776
Joined: 04.02.2005
With us: 19 years 8 months

Post #46by ElChristou » 21.08.2006, 16:19

chris wrote:
ElChristou wrote:
chris wrote:...I applied a wide Gaussian blur to the Eros LIDAR data, subtracted this from the original data, and applied it to a sphere...

You are talking of displacement here, right?

Yes. Most of the asteroid shape models in raw form are essentially displacement maps--distance from center, regularly sampled over longitude and latitude.

--Chris


Have you wrote your own tool for this? (I would love to do some test with such a tool!!)
Image

Topic author
chris
Site Admin
Posts: 4211
Joined: 28.01.2002
With us: 22 years 9 months
Location: Seattle, Washington, USA

Post #47by chris » 21.08.2006, 19:36

ElChristou wrote:
chris wrote:
ElChristou wrote:
chris wrote:...I applied a wide Gaussian blur to the Eros LIDAR data, subtracted this from the original data, and applied it to a sphere...

You are talking of displacement here, right?

Yes. Most of the asteroid shape models in raw form are essentially displacement maps--distance from center, regularly sampled over longitude and latitude.

--Chris

Have you wrote your own tool for this? (I would love to do some test with such a tool!!)


Yes, I've used several custom tools, among them:
- cmodsphere takes as input a file of regularly spaced (in long and lat) height samples and generates a nicely tessellated cmod file
- another program applies a Gaussian blur of a specified standard deviation to a simple cylindrical height map
- Another program to compute the difference of two simple cylindrical height maps. I'm most interested in using the difference of the original map and a Gaussian filtered one in order to generate a normal map (using Fridger's nmtools)

cmodsphere is actually checked in to CVS, but it's not really ready for general usage yet.

--Chris

Paolo
Posts: 502
Joined: 23.09.2002
With us: 22 years 1 month
Location: Pordenone/Italy

Post #48by Paolo » 21.08.2006, 19:45

Perhaps its time to think to a suite of Celestia editing tools or better to an Integrated Development Environment.
Kind regards
Remember: Time always flows, it is the most precious thing that we have.
My Celestia - Celui

Topic author
chris
Site Admin
Posts: 4211
Joined: 28.01.2002
With us: 22 years 9 months
Location: Seattle, Washington, USA

Post #49by chris » 21.08.2006, 22:53

Paolo wrote:Perhaps its time to think to a suite of Celestia editing tools or better to an Integrated Development Environment.
Kind regards


I'd like to, but I'm trying to spend most of my time on the core Celestia code. I will release them as a suite of robust command-line tools (like cmodfix), but a graphical interface will be up to someone else to implement.

--Chris

Paolo
Posts: 502
Joined: 23.09.2002
With us: 22 years 1 month
Location: Pordenone/Italy

Post #50by Paolo » 22.08.2006, 00:45

chris wrote:
Paolo wrote:Perhaps its time to think to a suite of Celestia editing tools or better to an Integrated Development Environment.
Kind regards

I'd like to, but I'm trying to spend most of my time on the core Celestia code. I will release them as a suite of robust command-line tools (like cmodfix), but a graphical interface will be up to someone else to implement.

--Chris


IMHO a robust suite of command-line tools with the availability of source code will be ok. So if someone will be interestend in developing an IDE will find a lot of work already done. :wink:

Kind regards
Remember: Time always flows, it is the most precious thing that we have.

My Celestia - Celui

Avatar
t00fri
Developer
Posts: 8772
Joined: 29.03.2002
Age: 22
With us: 22 years 7 months
Location: Hamburg, Germany

Post #51by t00fri » 22.08.2006, 06:19

chris wrote:
Paolo wrote:Perhaps its time to think to a suite of Celestia editing tools or better to an Integrated Development Environment.
Kind regards

I'd like to, but I'm trying to spend most of my time on the core Celestia code. I will release them as a suite of robust command-line tools (like cmodfix), but a graphical interface will be up to someone else to implement.

--Chris


Our nmtools ( for producing quickly highest quality "monster" normalmaps on normal home computers), coauthored by Robert Skuridin, will be such tools as well. They are ready since quite a while and were also used by Chris in his work above.

For the first time they generate correct normalmaps for ~ spherical bodies! This means, e.g. all normalmaps on ML are progressively incorrect in regions away from the equator.

The distribution of our commandline tools includes the Win32, Mac, Mac-Intel and Linux binaries as well as the sources. They will be for download in CelestialMatters very soon. There will also be a tutorial in CM by myself....For Win32 a setup.exe installer is included. No knowledge about image manipulation is required.

I announced this already a while ago.

Bye Fridger
Last edited by t00fri on 22.08.2006, 08:19, edited 2 times in total.
Image

Avatar
t00fri
Developer
Posts: 8772
Joined: 29.03.2002
Age: 22
With us: 22 years 7 months
Location: Hamburg, Germany

Post #52by t00fri » 22.08.2006, 08:15

I just got one more benchmark by Runar as to the performance of our new nmtools (see post above this one):

Runar, on his 1.8 GHz iMac G5 => 45 min
==================
To do all 6 levels of highly optimized 1k x 1k normalmap VT tiles, corresponding to a 64K monster normalmap, directly from the published 85K, 16bit integer srtm_ramp2.world.86400x43200.bin elevation map!

Me, on my 3.2 Ghz/3GB Ram,PC => 15-20 min
==================================
Same thing, but including the conversion of the PPM format tiles into PNG. My benchmark is ~ the same for Win32 and Linux.

Bye Fridger
Last edited by t00fri on 22.08.2006, 14:23, edited 1 time in total.
Image

Christophe
Developer
Posts: 944
Joined: 18.07.2002
With us: 22 years 3 months
Location: Lyon (France)

Post #53by Christophe » 22.08.2006, 08:30

I say vaporware :wink:

Stop teasing us Fridger, it is time to show us the code!
Christophe

Avatar
t00fri
Developer
Posts: 8772
Joined: 29.03.2002
Age: 22
With us: 22 years 7 months
Location: Hamburg, Germany

Post #54by t00fri » 22.08.2006, 09:16

Christophe wrote:I say vaporware :wink:

Stop teasing us Fridger, it is time to show us the code!


You are welcome to participate in the testing, like other developers and friends do as well.

You are just one of the developers who never showed interest in them so far ;-) .They are completely ready since > 1 month. Dirkpitt has them and kindly compiled the MAC binaries etc, Chris also has them and uses them since a while (see part of his images above ;-) ).

Let me know.

We are working hard to get CM opened. That's the bottleneck, not the nmtools. The tutorial for nmtools is still not finished, though. But you certainly wouldn't need one ;-)

I would be most interested to learn about further benchmarks.

Bye Fridger
Image

Christophe
Developer
Posts: 944
Joined: 18.07.2002
With us: 22 years 3 months
Location: Lyon (France)

Post #55by Christophe » 22.08.2006, 09:37

It's true that I've never played much with textures and models, I don't have much expertise in the field. I'd however be happy to run some tests and give you some feedback, you know my address :wink:

I, for one, am very impatient to see CM launched and discover what you've prepared for us!
Christophe

Avatar
t00fri
Developer
Posts: 8772
Joined: 29.03.2002
Age: 22
With us: 22 years 7 months
Location: Hamburg, Germany

Post #56by t00fri » 22.08.2006, 09:43

Christophe wrote:It's true that I've never played much with textures and models, I don't have much expertise in the field. I'd however be happy to run some tests and give you some feedback, you know my address :wink:

I, for one, am very impatient to see CM launched and discover what you've prepared for us!


Good...

But I warn you: NO i18n internationalization YET ;-)

Cheers,
Fridger
Image

Topic author
chris
Site Admin
Posts: 4211
Joined: 28.01.2002
With us: 22 years 9 months
Location: Seattle, Washington, USA

Post #57by chris » 24.08.2006, 06:46

I'm now satisfied with the work I've done with the NLR data set.

This image is a screen capture of an Eros mesh with one vertex per degree at the equator and a 2880x1440 normal map:
Image

You can grab a zip file with the mesh and model from http://www.celestiaproject.net/~claurel/celestia/asteroids

There's a version with the full-resolution normal map and another with the normal map resized to 2048x1024--use this one if your graphics hardware doesn't support non-power-of-two textures (all GeForce 6 and 7 series cards do.) I've also uploaded two higher resolution versions of the mesh. The 4 vertices per degree version doesn't have texture coordinates or tangents (but if someone needs this, let me know . . .)

Remember: you will only see the normal map if you're running a recent CVS version of Celestia and have the OpenGL 2.0 path enabled.

That's enough of Eros for the time being . . . I've proven to myself that the new OpenGL 2.0 mesh rendering code is working well enough.

--Chris

phoenix
Posts: 214
Joined: 18.06.2002
With us: 22 years 4 months
Location: Germany - Berlin

Post #58by phoenix » 24.08.2006, 09:23

chris wrote:There's a version with the full-resolution normal map and another with the normal map resized to 2048x1024--use this one if your graphics hardware doesn't support non-power-of-two textures (all GeForce 6 and 7 series cards do.)


looks really cool!
but how do i know if my card does support non-pow-of-two textures? (GF 5900 FX).
most recent celestia win32-SVN-build - use at your own risk (copy over existing 1.5.1 release)

Avatar
Chuft-Captain
Posts: 1779
Joined: 18.12.2005
With us: 18 years 10 months

Post #59by Chuft-Captain » 24.08.2006, 10:14

phoenix wrote:
chris wrote:There's a version with the full-resolution normal map and another with the normal map resized to 2048x1024--use this one if your graphics hardware doesn't support non-power-of-two textures (all GeForce 6 and 7 series cards do.)

looks really cool!
but how do i know if my card does support non-pow-of-two textures? (GF 5900 FX).


Just try it. If your card doesn't support non-power-of-two, then the texture won't display. Eros will have a white surface.
"Is a planetary surface the right place for an expanding technological civilization?"
-- Gerard K. O'Neill (1969)

CATALOG SYNTAX HIGHLIGHTING TOOLS LAGRANGE POINTS

Fightspit
Posts: 510
Joined: 15.05.2005
With us: 19 years 5 months

Post #60by Fightspit » 24.08.2006, 10:20

phoenix wrote:
chris wrote:There's a version with the full-resolution normal map and another with the normal map resized to 2048x1024--use this one if your graphics hardware doesn't support non-power-of-two textures (all GeForce 6 and 7 series cards do.)

looks really cool!
but how do i know if my card does support non-pow-of-two textures? (GF 5900 FX).


You graphic card doesn't support non-power of two:
http://www.nvidia.com/object/LO_20021117_4942.html

For Geforce6 and Geforce7:
Geforce6: http://download.nvidia.com/ndemand/prod ... ies_24.pdf
Geforce7: http://www.nvidia.com/object/IO_30459.html
Motherboard: Intel D975XBX2
Processor: Intel Core2 E6700 @ 3Ghz
Ram: Corsair 2 x 1GB DDR2 PC6400
Video Card: Nvidia GeForce 8800 GTX 768MB GDDR3 384 bits PCI-Express 16x
HDD: Western Digital Raptor 150GB 10000 rpm
OS: Windows Vista Business 32 bits


Return to “Ideas & News”