Celestia Forums

I've managed to convert Robert Gaskell's very nice model of Phoebe into cmod format. The data set is available in four resolutions; this video shows the model at the second highest resolution:

http://www.youtube.com/watch?v=ifmDLKkyEhM

I've made the models available here:

http://www.celestiaproject.net/~claurel/celest ... ls/phoebe/

The highest resolution version has not been uploaded yet. I'm experimenting with using a lower resolution geometry file and a normal map derived from the high resolution model. This should give a very realistic looking rendering without the memory and performance impact of using the highest resolution model.

Some version of this model should be included in the base Celestia package.

--Chris

The model (high res) is very nice, Chris. However, we should be able to apply a texture to it. Is this possible ?

Cham wrote:The model (high res) is very nice, Chris. However, we should be able to apply a texture to it. Is this possible ?

The current model does not have uv coordinates; I am working on generating them, since they are required for a normal map.

--Chris

Also Mimas:

http://www.youtube.com/watch?v=GS3NdC_hliA

The model may be downloaded here:

http://www.celestiaproject.net/~claurel/celestia/models/mimas/

mimas256.cmod contains about 786,000 triangles, which current graphics hardware can handle without much problem. The highest resolution version has four times as many triangles.

--Chris

The technique of using normal maps for fine geometry detail in small body shape models works spectacularly well! The screenshots below show the lowest resolution versions of Gaskell's models with normal maps generated from the highest resolution versions. Each of the models has about 50,000 triangles. Other than the normal map, there is no texture.

Eros:


Phobos:


Phoebe:


Itokawa:


Models and normal maps for Phobos, Eros, Phoebe, and Itokawa are here:

http://www.celestiaproject.net/~claurel/celest ... models.zip

The are not quite perfect yet; there's still a seam visible between sections of the models. I'm working to eliminate this artifact, but it's proving persistent.

--Chris

Wow that's very nice!

Beautiful!

A video of the 433 Eros model:

http://youtu.be/_1qZc278-yA

This model is identical to the one that I uploaded a few days ago except that it uses a compressed normal map and has a uniform color (based on 'true color' images from the NEAR team.)

--Chris

chris wrote:A video of the 433 Eros model:
--Chris

Very nice. Have you solved the seams problem ?

Ready to release all the models ?

chris wrote:A video of the 433 Eros model:

http://youtu.be/_1qZc278-yA

This model is identical to the one that I uploaded a few days ago except that it uses a compressed normal map and has a uniform color (based on 'true color' images from the NEAR team.)

--Chris

Indeed, Cosmographia is getting VERY nice. Builds also fine under Linux with the same cosmographia.pro.

Fridger

t00fri wrote:Indeed, Cosmographia is getting VERY nice. Builds also fine under Linux with the same cosmographia.pro.

Beeing not a very frequent reader here anymore, I totally missed "Cosmographia" until now.

Looking at the development of Celestia, compared to the activity of the Cosmographia project - is it safe to assume that Celestia is dead by now?

(I know that this question was already answered with 'no' back in August, but that does not seem very convincing...)

Cham wrote:Very nice. Have you solved the seams problem ?

Ready to release all the models ?

The seams are still present unfortunately. I'm using a tool called xNormal to generate the normal map from the high resolution geometry, and I can't understand what's going wrong at the edges. I'll make a separate post in the modeling forum describing all the steps required to create these models.

I've repackaged all of the models (including Mimas this time) as a Celestia add-on:
http://www.celestiaproject.net/~claurel/celest ... models.zip

The bodies replaced are:
Phobos
Mimas
Phoebe
433 Eros
25143 Itokawa

They should all be in their correct orientation. This line is required in all of the definitions in order to compensate for Celestia's unconventional coordinate system:

Orientation [ 180 0 0.7071 0.7071 ]

This is the product of two rotations: 90 degrees about the x-axis followed by a 180 degree rotation about the y-axis.

--Chris

Chris,

the Itokawa model has holes, very easily visible with and without the normal map :



Also, we could see triangles in some shadow parts (low resolution of the mesh), especially around craters. I strongly suggest that you use higher resolution models.

Cham wrote:Chris,

the Itokawa model has holes, very easily visible with and without the normal map :

It's a problem with the original data set. I'll try reprocessing it to merge nearly identical vertices.

Also, we could see triangles in some shadow parts (low resolution of the mesh), especially around craters. I strongly suggest that you use higher resolution models.

Obviously, the model will look better with more geometric detail. Using the next higher level of detail would quadruple the number of triangles; in my opinion, it's not worth the modest gain in visual quality.

--Chris

steffens wrote:...
Looking at the development of Celestia, compared to the activity of the Cosmographia project - is it safe to assume that Celestia is dead by now?

(I know that this question was already answered with 'no' back in August, but that does not seem very convincing...)

Man I hope not. This is an amazing work.

I don't get why everything seems to have died down though

The Celestia developer group is quite small. They all seem to be working on other projects right now, but I know there are plans for major enhancements to Celestia. They'll just take a little longer to complete than some people might like.

In the meantime, building your own copy of Celestia from the developmental source code isn't very hard. The procedure is described in the Celestia WikiBook at http://en.wikibooks.org/wiki/Celestia#Development