Just a test - 64K Moon

[Reiko]

Is it possible to make VT textures with your tools using one of my own textures? Like if i made 16k planet texture or something?

[Vincent]

praesepe,

Thanks for the links. I managed to build my own Clementine VT for the Moon in about one hour and a half (download time included).
I just had to fix the following typo in your batch file (line 11), though:

Change

Code: Select all

tx2half 1 32768 < level4\32k.bin > level3\16.bin

into

Code: Select all

tx2half 1 32768 < level4\32k.bin > level3\16k.bin

One more precision:
On Windows, one must add the bin subfolder of the gdal installation folder in the PATH environment:
...\FWTools2.3.0\bin

Now, I can enjoy views like the following one, thanks to you and to Fridger's F-TexTools!

[Andy74]

Is it possible to make VT textures with your tools using one of my own textures? Like if i made 16k planet texture or something?

Hi, Reiko.

I think this is one of the main goals of Fridger's tools -- their versatility.
Of course you can do exactly that. You just need to combine the "tx2half" and "txtiles" commands in a similar way praesepe has shown. But the first parameter <channels> has to be set correctly. For greyscale maps it is set to 1, for color maps to 3 and for color maps including an alpha channel to 4. And don't forget to convert your (png) texture to a binary file using png2bin. You needn't use the "tx2pow2" command if your texture is exactly 16K.
If you want to try I'm sure you'll find help here.

Bye
Andy

[t00fri]

Is it possible to make VT textures with your tools using one of my own textures? Like if i made 16k planet texture or something?

Reiko,

don't you think us Andorians should help each other?

It is really simple to convert any (big) texture of yours into VTs: Just take an arbitrary tool of your preference to convert your texture into PNG (e.g. GIMP). Let me know your texture's name and width. Note, the ratio width / height should be 2. Then i'll write out the commands for you.

Meanwhile, install the F-TexTools 2.0 and check that they are correctly installed by typing
e.g. 'txtiles' at the Windows command prompt (Windows-Start->Run->cmd.exe). If you see a help text displayed, everything is ready to rumble

If not, let me know what you get...

Fridger

[jogad]

Hello,

Yes, it is possible! And it is worthwhile.
Thank you very much for the links, Praesepe.

For example here are two screenshots with the hires texture that comes with Celestia 1.6.0 rc1 and with the 64k VT texture.
At last, we can see the craters near the north pole of the Moon.

sansVT.jpg

avecVT.jpg

Here is the link (for Celestia 1.6)
cel://Follow/Sol:Earth:Moon/2009-05-11T ... rc=0&ver=3

However, this texture is not perfect. There are some small black gaps in it and a bit of work is still needed.

For those who don’t want to add one more PATH to their environment or who just don’t know how to do it, you can add the following line at the beginning of your list of commands.
It is assumed that “c:\Program File\FWTools2.3.0” is the installation folder for Gdal.

Code: Select all

call "C:\Program Files\FWTools2.3.0\setfw.bat"

[t00fri]

...
Now, I can enjoy views like the following one, thanks to you and to Fridger's F-TexTools!

Wow!

Meanwhile, I am investigating a great 16k RGB color texture of the Moon as extracted by means of ISIS3, as well as some ISIS3 commands that automatically eliminate all Clementine artefacts...

Here is a tiny example, how such color textures of the Moon will look like (NO black artefacts [NULL pixels] whatsoever!):







Fridger

[praesepe]

praesepe,

Thanks for the links. I managed to build my own Clementine VT for the Moon in about one hour and a half (download time included).
I just had to fix the following typo in your batch file (line 11), though:

Change

Code: Select all

tx2half 1 32768 < level4\32k.bin > level3\16.bin

into

Code: Select all

tx2half 1 32768 < level4\32k.bin > level3\16k.bin

One more precision:
On Windows, one must add the bin subfolder of the gdal installation folder in the PATH environment:
...\FWTools2.3.0\bin

Now, I can enjoy views like the following one, thanks to you and to Fridger's F-TexTools!

Thank you for the corrections! I'm so sorry for the errors. I've done the proper corrections on the instructions and fixed the missing "k" in both code and batch files.

Praesepe,

thanks VERY much for the links and your effort of setting up the step-by-step instructions!

Fridger

Not at all! It's a pleasure to try to help (even if it's little help ) Thank you too for your tools.

I DON?T agree with you.It?s ABSURDELY COMPLICATED to generated a map with F-Tex Tools,take a huge space in HD and take many hours to end.I tried to made maps using the program,but I gave up,because it locks down or give some kind of error.I am waiting for AGES for a finished texture to download,but the only options are this DIY.Continuing this,I will have to give up Celestia entirely. Besides this,my system is Windows Vista X64 and DOS system works very badly on this system.Please find a way to become this texture available!

danielj,

You should give it another try. F-TexTools are pretty clear and straightforward to use if you read carefully the README that comes with them. Even if you type any of the commands inside a shell/cmd they will prompt the parameters you've got to provide, quite simple! As for space requirements, in the producedure given in my instructions the input map (once converted to bin) weights 4.1GB and only took 4 minutes to complete using an old computer. When finished the whole thing weights 1.1GB in DXT format which is a negligible amount of space if you consider actual hard disk sizes...

If you use the batch files is even easier, you've only got to click and wait.

Wow!

Meanwhile, I am investigating a great 16k RGB color texture of the Moon as extracted by means of ISIS3, as well as some ISIS3 commands that automatically eliminate all Clementine artefacts...

Here is a tiny example, how such color textures of the Moon will look like (NO black artefacts [NULL pixels] whatsoever!):

Excellent!! Having those missing areas removed automatically is for sure time saving! In my texture the process was done via the traditional way, by hand with the problems of managing such enormous file!

[t00fri]

Excellent!! Having those missing areas removed automatically is for sure time saving! In my texture the process was done via the traditional way, by hand with the problems of managing such enormous file!

Here is a little summary of the nice things one can do with the help of the ISIS3 routines:

As long as one works at the level of an ISIS3 *.cub file, one can exploit all ISIS3 utility routines, among them one specializing on Celementine/NIR artefacts. It is called 'clemnirclean'. It is a combination of various lowpass and noise filters. I will eliminate among other things those black blocks and stripes that represent NULL pixels. NULL pixels are special pixels that are set when faults arose somewhere in the image taking process. These pixels can be easily localized by the filter and then will be replaced by the average pixel value in the socalled 'boxcar', a testing area of a certain width and height. However, to get rid of all those black blocks, one has to iterate this 'clemnirclean' filter about 8 times! Altogether, the image quality improves considerably, since the image also sharpens and intensifies due to the noise elimination.

Moreover, a ISIS3 *.cub file consists of a stack of width x height images, one for each color band. In my colored UVVIS/NIR texture of the Moon, there are 5 color bands: R,G,B and 2 IR wavelengths. There is another ISIS3 routine that allows to extract the various color planes either singly or as a normal RGB triple e.g. in PNG format. Another option is to extract the RGB picture as a colored BIN file which then forms the input for my F-TexTools.

The colors look VERY natural!

Fridger

[Chuft-Captain]

praesepe,

FYI, there's now a later version of the GDAL windows executable than the one you directly linked to.

Perhaps for future readers of this thread, you might be better to provide this link instead of a direct line to the .EXE
Cheers
CC

[Chuft-Captain]

DanielJ,

Perhaps if you're having trouble with the cmd line tools, you need to do some more basic examples so that you understand better what it is that I-O redirection and pipes actually does.

I don't know what your level of experience with DOS is, but perhaps you need to do some really basic examples to better understand.

For example the following steps should help you understand output redirection (the ">" character)

1. Open a folder in WINDOWS which contains just a few simple text files.
2. Open a cmd prompt window and goto the same folder (using CD command)...
3. In the CMD window type the command:

Code: Select all

dir/b

and view the results.
4. Now perform the same command, but re-direct it's output to a txt file called fred.txt

Code: Select all

dir/b > fred.txt

(Notice what happens in the windows view of the folder which you already have open)
5. Type the following command in CMD window:

Code: Select all

type fred.txt

or open the file fred.txt in windows using notepad etc...
6. Now type the following command in CMD window:

Code: Select all

type fred.txt > fred2.txt

and examine fred2.txt as before in either windows or DOS.

Hope this helps and is not percieved as too patronising. As I said, I don't know what your level of understanding is, but it's always easier to understand a concept if you start with a simple example.

The next step is to understand that the dir/b and type commands in the example above are just executables like fridgers, they just do different tasks. However, the DOS output re-direction does exactly the same thing with the output of his programs -- it re-directs it to the file which follows the ">" character.

CC

[cartrite]

That color moon texture looks really impressive. Where does that come from?
I read somewhere that Japan may release a new topo and color map by years end from their spacecraft.
cartrite

[danielj]

Updated
I know NOTHING about DOS command lines,so does the majority of new users.This kind of approach (always DIY) will surely FRIGHTEN new users and enthusiats of Celestia.I only produced the normal map,because I RUN an SCRIPT.It?s time for at least a 32k texture map of the Moon to be ready to use.

I used the nm-tools and produced a Normal Map for Earth BMNG.But F-Tex Tools is more difficult to use.I don?t understand why,but every time I tried,it didn?t work.Anyway if there isn?t another way,one day I will try...

[t00fri]

I used the nm-tools and produced a Normal Map for Earth BMNG.But F-Tex Tools is more difficult to use.I don?t understand why,but every time I tried,it didn?t work.Anyway if there isn?t another way,one day I will try...

Note that you should use the latest 2.0pre1 version, install it properly with the installer F-TexTools-2.0pre1.exe ! Otherwise, the required registry entry will not be correctly done, such that the tools will not be known in every directory just by typing their name. Any previous F-TexTools versions should better be deinstalled as usual.

If you don't tell concisely WHAT did not work, noone can help you.

By design, the F-TexTools are operated EXACTLY like the nmtools.

Fridger

[t00fri]

That color moon texture looks really impressive. Where does that come from?
I read somewhere that Japan may release a new topo and color map by years end from their spacecraft.
cartrite

Steve,

my above texture is from here:

http://ser.sese.asu.edu/MOON/clem_color.html

The previous images are from a reduced and cleaned 10k rgb sample, unfortunately with a grid. I am actually working on the bigger variant (20k, download at the bottom of that page! ) that is a combination of latest CelementineNIR and UVVIS data and the .cub file has 5 color bands RGB and 2 IR ones.
So far I was quite unsuccessful in auto-cleaning these vertical stripes and black (NULL pixel) blocks by means of the various filter routines of isis3. Apparently the above cleaned version also didn't manage very well, yet all black blocks are gone.

Fridger

[t00fri]

Updated
I know NOTHING about DOS command lines,so does the majority of new users.

Many will be curious to LEARN a little bit of DOS, if they get something nice for the effort! I am aware that this option has always be alien to you...

This kind of approach (always DIY) will surely FRIGHTEN new users and enthusiats of Celestia.

+++++++++++++++++++++++++++++++
Celestia comes with a standard set of textures, where absolutely no DIY is required!
Everything is READY to use. This fact you always ignore. This is the texture set for pure "endusers" .
+++++++++++++++++++++++++++++++

The big 64k VT sets are NOT part of the Celestia distribution!!!

They are multi GB in size and therefore require special DIY action, since otherwise the network traffic becomes excessive!

As usual, you don't only expect that someone produces these big textures for you, but also that someone entertains and PAYS for a powerful server site that could support such a heavy download traffic for your pleasure....

Fridger

[cartrite]

Did you try noseam? This runs automos and some filters to blend in the seams that may be present.
I'm downloading the 500m mosaics now. I'll see whats going on.
cartrite

[cartrite]

I'm not sure what JAX has published. They may have images out there that fill in some of the null areas.
Blending the vertical stripes will probably be difficult if not impossible though. The best time to blend them is when the mosaic is originally done. I'm having the same problem putting MARCI images together. I'm trying a program called photomet. There were some problems with that program and they are working on a patch.
cartrite

[cartrite]

I had a look at the west cube. I noticed that some null areas are not null in all bands. A large null area in band 1 is covered in band 3, 4, and 5. Cropping the area out of band 3, changing the "stretch/dn values" to match band 1 may fill that area in band 1. That would be a lot of work though for all areas. There a lot of null areas. Some are null in all bands.
Combining JAX and Clementine images would probably be the way to go though. I think they use different color detectors though so matching them will be tough.
cartrite

[t00fri]

I had a look at the west cube. I noticed that some null areas are not null in all bands. A large null area in band 1 is covered in band 3, 4, and 5. Cropping the area out of band 3, changing the "stretch/dn values" to match band 1 may fill that area in band 1. That would be a lot of work though for all areas. There a lot of null areas. Some are null in all bands.
Combining JAX and Clementine images would probably be the way to go though. I think they use different color detectors though so matching them will be tough.
cartrite

Steve,

I think it's also important to define first the aim of the task: the idea would be to attempt automatic artefact cleaning, in order to be able to provide (e.g. at CM) a cleaned 64k binary starting file for people to operate my F-TexTools on.

The only effective way I know of, where artefacts can be handled at the level of "monster" files is via ISIS3 filter routines. Right?

Meanwhile, I have practiced quite a bit in playing with boxcar sizes, low- and highpass filters etc, to get rid of quite a lot of stripes and stuff. But the Celementine data actually show sharply delimited bands that are really tough...

Fridger

PS: Incidentally, what would be the latest 85k SWBD_watermask spec texture from you, that could be just mounted with my tools on top of the RGB base texture (without an additional b<-> w inversion)? At CM, we have a link to your file. Was there some upgrade from your part, meanwhile?

If not, we can just leave everything as it is...

[cartrite]

The only effective way I know of, where artefacts can be handled at the level of "monster" files is via ISIS3 filter routines. Right?..
....

I'm not sure but ISIS3 is all that I use.ISIS2 may have some programs that clean up artifacts too.

Meanwhile, I have practiced quite a bit in playing with boxcar sizes, low- and highpass filters etc, to get rid of quite a lot of stripes and stuff. But the Celementine data actually show sharply delimited bands that are really tough...

I've ran noseam on MARCI images and saw how it worked. Judging by the results in the documentation on noseam, they run highpass and lowpass with the same size boxcar {73x73) and add them together with algebra. The example comes out pretty good. Since the mosaic is done already, you might be able to just use lowpass, highpass, both with boxcar lines and samples set to 73 and then run algebra.

Fridger

PS: Incidentally, what would be the latest 85k SWBD_watermask spec texture from you, that could be just mounted with my tools on top of the RGB base texture (without an additional b<-> w inversion)? At CM, we have a link to your file. Was there some upgrade from your part, meanwhile?

If not, we can just leave everything as it is...

I think you may already have the latest. The only thing I played around with lately was residual light. I did one with no residual light. They used to be on my websites but they are no more. I don't have a website anymore. I could upload one to CM if the one you had got lost or something.
cartrite