Celestia Forums

Hi all,

today I had a go at a new precision 2k Iapetus texture. Let me illustrate the 'state of the art' below.

Again, the used raw textures are all published:

1) Lowres from Voyager:

2) Hires from Cassini:
http://photojournal.jpl.nasa.gov/jpeg/PIA06166.jpg

3) For the coloration, I used this recent Cassini 'natural color' image as a template for the computer color mapping:
http://photojournal.jpl.nasa.gov/jpeg/PIA06167.jpg

4) I took great care to precisely map the hires Cassini image to cylindrical projection using again Matthew Arcus' 'mmps' program.

5) Then followed the tedious alignment step and the coloration.

6) Finally, I also made a spec texture and a bump-map texture. Since from the spectral analysis, we know that the white stuff is largely water-ice, a spec texture makes a lot of sense! The generation of the bump-map followed my long-standing know-how including smoothing the gray-scale texture with a 5 pix gaussian blur etc...

It is worth noticing that I am now all set for implementing further hires texture fragments as they come along...Everything is neatly arranged on separate layers, ready for further additions.

When I have completed some further tests, I shall make the set available for general download in this thread.

Here is how things look so far:

First the Cassini 'natural color' photo, taken as a color reference:


Finally, 2 views from within Celestia. The lower one exhibits the lowres Voyager imaging, while the upper one displays mainly the hires Cassini one. Note the spec and bump-map effects! The ridge is nicely visible...

Enjoy,

Bye Fridger

Looking good Fridger,will you make it available for us all soon,I will have a go at modelling it then,Jestr

jestr wrote:Looking good Fridger,will you make it available for us all soon,I will have a go at modelling it then,Jestr

Certainly...

Bye Fridger

t00fri, how does your texture compare to the texture here: http://laps.fsl.noaa.gov/albers/sos/sos.html (bottom of the page)? Does it have more detail? Better color matching? Thanks.

Aaron

Yes, very nice indeed. It's strange how I always get the impression that this strange little world is somehow METALLIC !

jestr wrote:Looking good Fridger,will you make it available for us all soon,I will have a go at modelling it then,Jestr

Wow!, you can model this? What with all the craters and surface features included?

Cheers,
TERRIER

Only by hand not any kind of mathematical model,(I dont know if anyone is planning on doing this?).If I have a reasonable texture I can make the craters (and the ridge) from this,like I did with Tethys,Mimas etc.Of course there isnt any kind of data about the heights and depths of these features (though I think I remember reading the ridge extends to about 20km.It is always tempting to enhance the heights because if I dont they are hard to spot from afar-I dont think this will be the case with the ridge though on the Cassini pics it is clearly visible extending out in space from the surface.All the best Jestr

Adreitz wrote:t00fri, how does your texture compare to the texture here: http://laps.fsl.noaa.gov/albers/sos/sos.html (bottom of the page)? Does it have more detail? Better color matching? Thanks.

Aaron

Aaron,

there are many differences in detail to Steve Albers texture, yet the published raw imaging from Voyager and Cassini is essentially the same. We both use practically all the photographic material that is available for now.

In my preview images above, some lowres Cassini maps are not yet included. It is really not easy to get those projected accurately and well alligned as you may see from Steve's texture. My final texture will also include them, however.

I put particular emphasis on doing a very good job in performing the cylindical projection of the hires image. Also I carefully aligned it and put great care into smoothing the transition regions.

I also use a different appoach to coloration. To start, I switch the whole composite to grayscale and first match the brightness and contrast levels of neighboring mosaic patches. Then I computer-map the colors from the 'natural' color image shown in my post above to /the whole texture/ at once. This gives both an accurate and very uniform coloration.

Finally, I prepared both a specular texture to make the water ice regions shine and reflect, as well as a smooth bumpmap.

Last not least, Steve Albers texture involves a longitude convention offset by 180 degrees relative to the Celestia conventions. So people using it, will first have to offset his texture by a (warped) amount of width/2=1024 pixels.

Bye Fridger

Fridger,

Just wanted to say Thank You for continuing to expand our horizons
with your work on these textures. It's always easy for someone to criticize
a piece of work, but I don't really believe that many people understand the
amount of effort which goes into one of these creations.

Note: Not trying to criticize critics either... Just wanted to thank the
good Doctor Schrempp for all of his work here.

Take care, Bob

Bob Hegwood wrote:Fridger,

Just wanted to say Thank You for continuing to expand our horizons
with your work on these textures. It's always easy for someone to criticize
a piece of work, but I don't really believe that many people understand the
amount of effort which goes into one of these creations.

Note: Not trying to criticize critics either... Just wanted to thank the
good Doctor Schrempp for all of his work here.

Take care, Bob

Bob,

many thanks for your kind appreciation! Really, what gets me going is the fun doing it

Bye Fridger

Thanks for the explanation, t00fri. I was mostly just looking for info about how your texture was different from Steve's in terms of amount of Cassini photo material, relative alignment of the photo material, and realism of colorization. Steve's texture is particularly poor in its colorization and matching of photo material, and I just wanted to know how much better yours is (or is becoming). Thanks! I've corrected some of Steve's maps for my own Celestia uses, and may put them up on Celestia Motherlode some day.

Aaron

Greetings,

Interesting to see Fridger's work from this week. I agree he has a very pleasing matching between the images as well as colorization of the Voyager low-res.

One item that may be worth thinking about for future improvements for Fridger and myself is the use of the "natural color" images. One challenge I'm facing in matching the color is that I'm actually using four color images, all supposedly in "natural color". Yet they disagree with each other for various reasons. So my question becomes is it better to try to blend in the colors from the various "natural color" images, or simply choose one of the color images as the reference and colorize the others from it? There are pros anc cons to each method. Fridger's approach is more aesthetically pleasing, though I'm holding out hope of catching some real hue variations across the satellite by retaining separate color information from multiple color images. This is a judgement call and the best procedure may vary from satellite to satellite.

scalbers wrote:Greetings,

Interesting to see Fridger's work from this week. I agree he has a very pleasing matching between the images as well as colorization of the Voyager low-res.

One item that may be worth thinking about for future improvements for Fridger and myself is the use of the "natural color" images. One challenge I'm facing in matching the color is that I'm actually using four color images, all supposedly in "natural color". Yet they disagree with each other for various reasons. So my question becomes is it better to try to blend in the colors from the various "natural color" images, or simply choose one of the color images as the reference and colorize the others from it? There are pros anc cons to each method. Fridger's approach is more aesthetically pleasing, though I'm holding out hope of catching some real hue variations across the satellite by retaining separate color information from multiple color images. This is a judgement call and the best procedure may vary from satellite to satellite.

Greetings, Steve,

I also noticed that you have been quite busy on your site during last week or so, by continuously improving your textures! Very nice, indeed.

That "infamous" 'natural color' issue is really a delicate matter, I know. So far I mostly mapped from a single 'natural color' photographic template using GIMP, but just now, I am also exploring a somewhat different technique. ImageMagic through its various command line tools allows to sample an indexed palette of colors averaged from an extended set of templates (i.e. several files). The resulting, actual template file then takes account of a number of possibly somewhat different photographic color renderings. I'll see where that will get me...

I also have a technical question for curiosity: How do you manage to get hold of the actual cylindrical projection parameters from photos, in case on can just see part of the surface but NOT a significant part of the (circular) rim?? NASA's specifications are usually too vague or nil and an incorrect choice may result in bad distortions. Certainly, in case that camera distance/object radius > 10-20 the projection ceases to be dependent significantly on perspective effects, i.e. things become considerably simpler.

Nevertheless, there are still possibly dangerous distortions lurking around

Bye Fridger

Hi Fridger,

Thanks for your comments and info about your latest explorations with GIMP, ImageMagick, etc. I'll have to learn more about those features as I am also using these tools to some extent.

I think there may yet be more that I can do in lieu of having to set some of the color images to grayscale. For example, some of the images I referred to as "natural color" are actually enhanced color (e.g. PIA 06145) so I will feel more at liberty to change their color balance now This set of images also has limited information in the blue channel so I will pay particular attention to blue related adjustements.

As for reprojecting the images, I am mostly assuming a large distance to simplify things. There are only a few cases so far where the limb was insufficiently visible as a reference. In some cases, the terminator can be of assistance. In other cases (one of the Titan images comes to mind) I can do the tedious trial and error by comparing the reprojected image to the pre-existing map. You may also be aware that you can get approximate satellite ephemeris info from Mark Showalter's Cassini Satellite Ephemeris generator (try that on Google). It's still tricky though since that ephemeris is sometimes just approximate and we do not often have exact timing info for the images. It's basically an interesting detective story at this point until more definitive info is published.

t00fri wrote:Hi all,

today I had a go at a new precision 2k Iapetus texture

Fridger, when will your new Iapetus textures be available for download? Will they appear on the Motherlode?

scalbers wrote:Hi Fridger,

Thanks for your comments and info about your latest explorations with GIMP, ImageMagick, etc. I'll have to learn more about those features as I am also using these tools to some extent.

I think there may yet be more that I can do in lieu of having to set some of the color images to grayscale. For example, some of the images I referred to as "natural color" are actually enhanced color (e.g. PIA 06145) so I will feel more at liberty to change their color balance now This set of images also has limited information in the blue channel so I will pay particular attention to blue related adjustements.

As for reprojecting the images, I am mostly assuming a large distance to simplify things. There are only a few cases so far where the limb was insufficiently visible as a reference. In some cases, the terminator can be of assistance. In other cases (one of the Titan images comes to mind) I can do the tedious trial and error by comparing the reprojected image to the pre-existing map. You may also be aware that you can get approximate satellite ephemeris info from Mark Showalter's Cassini Satellite Ephemeris generator (try that on Google). It's still tricky though since that ephemeris is sometimes just approximate and we do not often have exact timing info for the images. It's basically an interesting detective story at this point until more definitive info is published.

Hi Steeve,

thanks for your quick feedback. So it seems that our methods to do the cyl projections are basically very similar. I also use (with great preference) the large distance approximation. Then I just need to suitably fit the piece of texture under consideration within a transparent /square/ layer, the sides of which equal the object's diameter. Still a somewhat tricky affair is to estimate the tilt of the image relative to North if it was not given explicitly. Any other unknowns can be determined during the matching process of the layer overlay with existing markings in other layers.

What you describe above in the context of adjusting the coloration is also available in the GIMP coloration dialog. There a various options to manipulate the in and outbound color windows in the mapping process etc.

Bye Fridger

Hi all,

after plenty of additional 'tuning' and implementation of all available Iapetus texture pieces, my 2k texture is now completed. I am quite happy with the result. The texture is as smooth as it can be, given the wildly different resolutions that are contained in this composite texture.

Here is a 'tiny' 800x400 reduction to give you a feel:



Here are a few shots from within Celestia with spec and bumpmap textures in action:



On the Cyclops site it is stated that the reflectivity of the white ice has been measured to be >60% . So this texture definitely needs a spec file! That's how it looks in detail:



I will release the whole thing very soon... Tonight I was out in a great concert (including concerto for violin by N. Paganini , H minor. It is rarely played due to its reputation of being the most difficult violin concerto! )

Bye Fridger

Fascinating?€¦

?€¦I?€™m waiting for Fridger ?€?how to?€

Fridger,

when will your Iapetus textures (with spec and bump) ready for download ? Isn't your texture shown above finished ?

t00fri wrote:I will release the whole thing very soon... Tonight I was out in a great concert (including concerto for violin by N. Paganini , H minor. It is rarely played due to its reputation of being the most difficult violin concerto! )

I'm not surprised, considering there's no such key as "H Minor"

Evil Dr Ganymede wrote:

t00fri wrote:I will release the whole thing very soon... Tonight I was out in a great concert (including concerto for violin by N. Paganini , H minor. It is rarely played due to its reputation of being the most difficult violin concerto! )

I'm not surprised, considering there's no such key as "H Minor"

Aha, you also want to teach me music!

N. Paganini
Concerto no. 1 in D major, no. 2 in H minor

Why don't you educate yourself in the internet, for example...
Like in this violonist's repertoire:

http://www.todorov.co.uk/artists/mariorepertoire.htm