Page 1 of 1

Valles Marineris

Posted: 13.02.2006, 16:36
by cartrite
A model of Mars's grand canyon. A question if anyone knows.
Is there a way to make the mars globe transparent to the model above?
When I place it too close to the surface the trench disappears into the globe.
Anyhow the model will have to float above the surface for now.
This was made with Toti's Z-fold script for blender. It has 256k vertices. It is in cmod format and is 17+MB. The texture was made from a 64k black and white. ( http://astrogeology.usgs.gov/mdim-bin/d ... N&lon=000E ). Color was added to try and match John van Vliet's Mars VT. The height map was made from file megt90n000fb.img from page ( http://pds-geosciences.wustl.edu/geodat ... 0x/meg032/
). This was then converted to a USGS DEM in ASCII with 3DEM. Then it was converted to a heightmap with Wilbur.exe. 26 pixels were cut from the top which resulted in a 11520x5760 gray map. This was applied as noise in Blender. Then Toti's Z-fold script was ran with Blender 2.36. This older version of Blender can run the script in about 2 minutes with 256K vertices. The newer version of Blender failed to run the script with that many vertices.

Image

Image

Image

Image

Image

cartrite

Posted: 13.02.2006, 20:03
by ANDREA
Wow Cartrite, it's veeery nice, indeed. :D
Hope someone will reply your answer, sorry, I cannot. :oops:
But I hope that you'll be able to reach the final result.
Bye and thank you.

Andrea :D

Posted: 13.02.2006, 20:55
by selden
So far as I know, the only way to make a gouge into a planet would be to provide the entire sphere yourself, gouge and all. Unfortunately, then VTs wouldn't work on it.

Posted: 13.02.2006, 21:22
by buggs_moran
I don't think subtraction on spheres is possible within Celestia. Would be very useful though. I think you would have to join it with a sphere. Perhaps this would be better as you could add data (the Mons and Hellas Planitia) as needed. Maybe Globemaker could supply a 1x for the base model and you could put in the higher resolution bits a little at a time. A big undertaking no matter what. What was the framerate like?

Posted: 13.02.2006, 22:48
by cartrite
I figured there was no way to displace the planet underneath. I can raise the atmosphere and try to get it as close to the surface as possible.
Frame rate. Well since my computer is rather old ( A K7 ) I never really measure the frame rate. With big textures or models the frame rate is very very slow. Something I just accept. I just try to get still shots that are as realistic as possible. And mabey someday when I get a better computer I can take the frame rates into account.

I have built a model of the olympus mons area and when I run them together
the frame rate must really be slow because everything hangs, then jumps. It's almost impossible to line them up right anyhow. Seperately that is. But the models are put together from 16 different 3ds files. Blender 2.36 can run Zfold with 1 million vertices on my linux system. I going to try the western part of Mars's middle latitudes.

cartrite

Posted: 13.02.2006, 22:55
by jestr
Hi Cartrite,I tried splitting the surface of Mars (model) up into tiles the same size as the highest resolution VT texture and then starting at the lowest point (in Hellas basin),placing this on the surface of Mars (sphere).I dont know if this way Celestia would only use the model tiles on the screen at any given time,(rather than loading an entire model of Mars-which brings Celestia to a crawl on my computer).I didnt get very far with this project though as it was a nightmare trying to match the edges of each tile (in altitude in the ssc file) and anyway even when they matched perfectly you could still see a seam where the model tiles met.It would be so much easier if you could lose the part of the sphere you need to for canyons like Valles Marineris.Cheers,Jestr

Posted: 14.02.2006, 01:15
by cartrite
Hi Jestr,
That's what I'm up to. Trying to tile a Mars model. The texture I'm using is based on BW images from the beginning of this post. I scaled them down and cropped them. If you think of Mars as being divided into 8 sections A1 thru D2. Each section has 4 8k square tiles. So they should be compatable with all the other textures of Mars used in Celestia.

These are the boundries I'm using to define the model tiles. The one shown is MC-17 and MC-18 with the added pixels at the bottom to make 2 8k squares. Trouble is when I process tiles MC-9 and MC-10 they come up with a different center, different radius, etc. So they don't match. I'm sure there is a formula to do this but so far I have not found it. That's got to be the answer. A formula. Or mabey it's the axis tilt of Mars??? Or a way to define a center for all the tiles?

One thing I found is when I split up the model to get the 3ds conversion to work I over lap them. Once they are converted to cmod I copy the meshes to one file and run MODFIX with -u -o -b. No seams. I havent tried to do this with the different model tiles yet tho. I'll probally get the seams then.
cartrite

Mars 1x model donation

Posted: 16.02.2006, 23:32
by GlobeMaker
Hello Cartrite and associates,

I can donate a 1x Mars model for you. We only need to agree on some details.
To understand the file sizes versus resolution and frame rate, see the
discussion on the various Venus models that I benchmarked. :
http://www.celestiaproject.net/forum/viewtopic ... 404d6bd2c7
I will prepare Mars models for you, after Cartrite replies to this questionaire :
Can you convert .stl files to .3ds ?
If yes, then a 100 megabyte model in .stl file format , with 1x exaggeration and 1/4 degree resolution
will be made available. If No, then 1/4 degree grid may not be convertable
to .3ds on my Accutrans3D software. Then only a 1/2 degree grid would be donated as a .3ds file.

Do you have an FTP website I can send the files to directly?
Then I do not need to clog up my website with your files.
Do you want tiles at 1/128 degree grid of Valles Marineris?
I have them ready. All my files will be seamlessly joinable.
If all of Mars were done at 128 points per degree, that would be 100 Gbytes
in .stl file format. Here is a preview at 7x exaggeration and 1/128 degree grid of Argyre Planitia :
http://www.reliefglobe.com/mars_price15.html
And here is Valles Marineris at 1x (no exaggeration)
http://www.reliefglobe.com/mars/mars5/m1x_valles-b.jpg
Please note that when looking at all of Mars at once when not exaggerated,
will look like a ball, even with the model you are interested in.
If Valles Marineris is 8 miles deep and Mars is 4210 miles wide, only 2 pixels will be
gouged out of the spheroid.

Do you have 3D mesh editing software? If you do, then I could make a
rough Mars using a 1 degree grid, and you can edit it to place your
detailed areas on the coarse Mars mesh. I looked at Autodesk 3D Max and
it costs $3500, so I did not buy it. Maybe Blender can make the edits. If you
can use .stl files, you will have the original files that are not sectioned into many .3ds parts. Then you can convert to .3ds after the coarse planet is merged with the detailed Valles.

Posted: 17.02.2006, 19:28
by cartrite
Hi GlobeMaker,
Thanks for your Kind Offer.
But my computer can't handel file sizes like the ones you showed in your
links. I've already made a low resolution model of mars but I'm not sure if I'm going to
continue with that approach. I'm happy to place them on the Mars that is there. To place them
on another model will just require more memory.

To answer your questions:
1. Can you convert .stl files to .3ds ? NO
2. Do you have an FTP website I can send the files to directly? NO
3. Do you want tiles at 1/128 degree grid of Valles Marineris? Not really. This is a project
I'm doing in my spare time. The GOAL is to LEARN about how to make models and what my computer can do.
So I'm pushing my computer to the edge to see what I can do with what I have.
The posts were to share what I've learned in case someone else wants to try it.

3. Do you have 3D mesh editing software? Blender 1.4.1 and 1.36

You have models that are of much better quality and resolution. Very Impressive.

This is my latest attempt. I used the 128 pixels/degree image to apply the noise. megt00n270hb.img

Image

Image

Image

The subdivision is 1.05 million vertices. I didn't figure out how many data points makes to a degree.
Toti's Z-fold script was used with a radius of 3416 km and a height of 29.4 km.
I'm not sure what that means as far as exaggeration.

This is the limit that my machine can do. Blender will not subdivide further and the text editors can't
combine the files any further. This only half the model that I made. I did learn alot though.

Thanks Again,

cartrite