It is currently Wed, 13-12-17, 13:04 GMT

All times are UTC




Post new topic Reply to topic  [ 32 posts ]  Go to page Previous  1, 2, 3  Next
Author Message
PostPosted: Tue, 29-08-17, 22:13 GMT 
Offline
User avatar

Joined: Sun, 05-02-17, 23:58 GMT
Posts: 2
In the screenshots for the northern hemisphere there is a noticeable misalignment between sea ice and land. I already mentioned a similar problem in the Celestia forum; I currently tend to believe that the BMNG is junk above a certain latitude (approx 80°).


Top
 Profile  
 
PostPosted: Wed, 30-08-17, 8:29 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4514
Location: Hamburg, Germany
catwrings wrote:
In the screenshots for the northern hemisphere there is a noticeable misalignment between sea ice and land. I already mentioned a similar problem in the Celestia forum; I currently tend to believe that the BMNG is junk above a certain latitude (approx 80°).


In the "olden days" (~ 2007), I have repeatedly communicated with the BMNG lead scientist Dr. Reto Stoeckli and got convinced that the BMNG surface texture was quite OK even for high latitudes, but the sea-ice was seasonally changing and thus often showed signs of apparent misalignment without also using Earth season data (as you do).

Another argument I remember is that there are many deep and narrow fjords to be filled by sea-ice. In case of low-resolution (binning) approaches to "sea-ice" these narrow ice fjords may often stay incorrectly ice-free.

I am currently exprimenting with two different GIMP-based approaches for installing ( and adapting) time-averaged 16k sea-ice data to the 16k BMNG texture of June 2004:

  • As in my previous screenshots, the sea-ice may be installed as a layer mask. This corresponds to the original method in BMNG if I correctly remember. The sea-ice mask from http://naturalearth.springercarto.com/n ... ce_16k.png of course needs to be slightly rescaled from 16200 x 8100 to 16384 x 8192 size!

  • The size-corrected 16k sea-ice mask is added as a standard layer on top of the RGB surface texture, and black color-selection is activated, followed by Colors->(black) Color to Alpha. This leaves only the whitish sea-ice, allowing for small color and geometry touch-ups.

Note that I always add the Land / Water mask (aka SpecularTexture) as alpha channel via one of my tools (tx2rgba) and then chop the 16k (RGBA + sea-ice layer) into high-quality DXT5 tiles. Note that for RGBA, the alpha channel (A) compression quality is way better in DXT5 than in DXT3 (yet DXT5 filesize = 2 * DXT3 filesize)!

Since currently I also got plenty of C++ coding to do, this level of sophistication should suffice initially in celestia.Sci. An essential highlight is the very fast performance of this 16k hires configuration.

Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Thu, 31-08-17, 17:02 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4514
Location: Hamburg, Germany
Here are some shots about the full 16k land / water mask from here
http://naturalearth.springercarto.com/n ... er_16k.png

As mentioned earlier: this layer was placed into the alpha channel, as usual.

Our traditional land / water texture by cartwright et al. stopped at about 60 degrees latitude.

The screenshots are again done from my VT tile installation in DXT5 format.

Amazon basin:
==============
[by all means click on image and then hit your browser's fullscreen key (F11 for FF)]
Attachment:
rivers1.jpg
rivers1.jpg [ 169.18 KiB | Viewed 738 times ]


Norway above 60 deg latitude
=======================
[by all means click on image and then hit your browser's fullscreen key (F11 for FF)]
Attachment:
rivers2.jpg
rivers2.jpg [ 258.54 KiB | Viewed 738 times ]


[by all means click on image and then hit your browser's fullscreen key (F11 for FF)]
Attachment:
rivers3.jpg
rivers3.jpg [ 197.68 KiB | Viewed 738 times ]


Enjoy,
Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Wed, 06-09-17, 16:58 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4514
Location: Hamburg, Germany
Arctic sea ice, touchup
=====================

After working a bit on the fitting of the northern sea ice, things are looking much better. Here is a 1920x1080 screenshot

[by all means click on image and then hit your browser's fullscreen key (F11 for FF)]
screenshot
Attachment:
arctic_touchUp.jpg
arctic_touchUp.jpg [ 263.37 KiB | Viewed 697 times ]


Next a screenshot that nicely exposes the richness of rivers and lakes
reflecting in the sun light:

The Great Lakes
===============

[by all means click on image and then hit your browser's fullscreen key (F11 for FF)]
Attachment:
great_lakes.jpg
great_lakes.jpg [ 283.08 KiB | Viewed 697 times ]


Enjoy,
Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Sat, 09-09-17, 8:27 GMT 
Offline

Joined: Fri, 13-06-14, 17:45 GMT
Posts: 59
Nice! As I remember you talked about releasing a beta ( or maybe alpha?) version soon...I think many of us would be happy to test it even if it is only an alpha.

It would build a bigger community I think...

I don't think that a limited alpha test release would ruin your reputation or give away your secrets...

On the contrary, a public release would cement your lead in this field... :clap:


Top
 Profile  
 
PostPosted: Sat, 09-09-17, 12:34 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4514
Location: Hamburg, Germany
sjohn wrote:
Nice! As I remember you talked about releasing a beta ( or maybe alpha?) version soon...I think many of us would be happy to test it even if it is only an alpha.

It would build a bigger community I think...

I don't think that a limited alpha test release would ruin your reputation or give away your secrets...

On the contrary, a public release would cement your lead in this field... :clap:


Thanks. Limited alphas are neither my style nor are they required, since celestia.Sci is VERY stable, indeed. A main problem is that I am working alone during the summer and that some features I definitely want to have initially, are simply not yet coded. Anyway I am trying hard to be fast ;-) Already now, since I have started a small communication link with people from celestiaproject, my private and public mail activities have substantially increased. Sorry, but my spare time is currently insufficient for this.

Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Mon, 11-09-17, 18:11 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4514
Location: Hamburg, Germany
Cmod package ported to 64 bit

Hi all,

a few days ago, a post by JohnVV remindet me that in celestia.Sci Chris Laurel's cmod package urgently required porting to our high building standards along with various code improvements.

This I have done meanwhile:

+++++++++++++++++++++++++++++++
Using the latest Qt and Qtcreator versions, the entire cmod project now builds throughout in 64bit without a single warning and of course without errors with these tools and settings:

latest qtcreator 4.4.0 and Qt 5.9.1
current MSVC 2017 in Windows 10 and Windows 7
Linux (openSuse 13.2, Leap 42.2): gcc 4.8.5 - gcc 6.3.x
+++++++++++++++++++++++++++++++

To achieve this result, quite a few reshufflings and additions notably of Qt headers and code were necessary. Now, the performance of the package (cmodview, 3dstocmod, cmodfix, cmodsphere, txt2cmod) is entirely stable and builds with a single click.

Notably the Qt application cmodview offers all options fast and without issues --with one exception:

As also discussed over in the celestiaproject site, the .dxt5nm normalmap format is not properly detected. Normalmaps in png format work nicely and also surface textures on non-spherical cmod bodies can be e.g. in dds or png etc.

So here I'll have to dive deeper into the original code...

Let me next display a few typical cmodview examples:

  • 4k Iapetus normalmap (png) without texture <- cf also Selden)
    [by all means click on image]
    Attachment:
    cmodview_iap.jpg
    cmodview_iap.jpg [ 97.73 KiB | Viewed 614 times ]


    and with an old 8k surface texture in dds format that I did long ago:
    [by all means click on image]
    Attachment:
    japetus_8k_texture.jpg
    japetus_8k_texture.jpg [ 92.93 KiB | Viewed 614 times ]


  • Cmod models without polar pinch from scientific data without Blender!
    Here I tested some 8 year old work about Phobos. The method is described in detail here:
    1) viewtopic.php?f=9&t=339
    2) viewtopic.php?f=9&t=337
    3) https://celestiaproject.net/forum/viewt ... pl#p124881
    [by all means click on image]
    Attachment:
    phobos_fri.jpg
    phobos_fri.jpg [ 88.09 KiB | Viewed 614 times ]

    after turning phobos around a bit, it really looks like a one-eyed smiling crocodile ;-)
    [by all means click on image]
    Attachment:
    one_eyed_croco.jpg
    one_eyed_croco.jpg [ 93.21 KiB | Viewed 614 times ]

  • Finally, I checked the gaskell cmods that were clevely enhanced by Chris Laurel: On medium-sized non-spherical cmod models (from data), he managed to superimpose hires normalmaps in the dedicated .dxt5nm format. This improves speed a lot, saves lots of polys and allows to load normalmaps of far higher resolution on top of only medium-sized cmod models! However, as ChrisL always emphasized to me: this only works without surface textures. Once I made some further progress about this important issue, I post it here...
    [by all means click on image]
    Attachment:
    eros_gaskell.jpg
    eros_gaskell.jpg [ 73.4 KiB | Viewed 614 times ]


    and this remarkble Eros rendering also happens to be an excellent approximation of the famous "Curry Wurst" as invented in Hamburg (just to test whether colors work properly in cmodview) ;-)
    [by all means click on image]
    Attachment:
    eros_curry_wurst.jpg
    eros_curry_wurst.jpg [ 80.44 KiB | Viewed 614 times ]

Enjoy,
Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Mon, 11-09-17, 20:38 GMT 
Offline
User avatar

Joined: Mon, 03-09-07, 23:01 GMT
Posts: 388
Location: Tuscany, Tyrrhenian Sea
Well done! By the way:
t00fri wrote:
Notably the Qt application cmodview offers all options fast and without issues --with one exception:...

there is another: cmodview doesn't convert binary<->ascii like 3dstocmod does. This is an useful option.

_________________
Never at rest.


Top
 Profile  
 
PostPosted: Mon, 11-09-17, 20:44 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4514
Location: Hamburg, Germany
fenerit wrote:
Well done! By the way:
t00fri wrote:
Notably the Qt application cmodview offers all options fast and without issues --with one exception:...

there is another: cmodview doesn't convert binary<->ascii like 3dstocmod does. This is an useful option.


Massimo,

sorry if my sentence was somewhat misleading: I did not mean to imply that cmodview now performs all tasks that the entire former cmod package did. I meant that cmodview does properly what it was supposed to to since its existence. The other package members like 3dstocmod, cmodfix etc deal with remaining (conversion) tasks. New code improvements, will come surely once I find some time for this.

Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Mon, 11-09-17, 20:52 GMT 
Offline
User avatar

Joined: Mon, 03-09-07, 23:01 GMT
Posts: 388
Location: Tuscany, Tyrrhenian Sea
That feature was postponed by Chris, if I remember well, along with the .OBJ material library (.MTL) support...

_________________
Never at rest.


Top
 Profile  
 
PostPosted: Tue, 12-09-17, 20:50 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 775
Location: N 42.38846 W 83.45456
one of the tools built in making cmodview is "cmodfix"

cmodfix converts a binary to a ascii and ascii to binary
Code:
Usage: cmodfix [options] [input cmod file [output cmod file]]
   --binary (or -b)      : output a binary .cmod file
   --ascii (or -a)       : output an ASCII .cmod file
   --uniquify (or -u)    : eliminate duplicate vertices
   --tangents (or -t)    : generate tangents
   --normals (or -n)     : generate normals
   --smooth (or -s) <angle> : smoothing angle for normal generation
   --weld (or -w)        : join identical vertices before normal generation
   --merge (or -m)       : merge submeshes to improve rendering performance

_________________
"I don't pitch Linux to my friends, I let Microsoft do that for me."
Using OpenSUSE 42.1 & Scientific Linux 6.7


Top
 Profile  
 
PostPosted: Tue, 12-09-17, 21:14 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4514
Location: Hamburg, Germany
Right, John. I actually used cmodfix a lot in my extensive testing of the new 64 bit cmod package. Whenever a binary .cmod model had to be modiified, cmodfix -a ... was used to first convert the binary to the ascii .cmod format. After the changes were done, "cmodfix --normals --binary..." was used again to convert back into a binary .cmod format.

Cheers,
Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Wed, 13-09-17, 5:27 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 8:49 GMT
Posts: 121
Location: Oxford UK
t00fri wrote:
Arctic sea ice, touchup
=====================

After working a bit on the fitting of the northern sea ice, things are looking much better. Here is a 1920x1080 screenshot

[by all means click on image and then hit your browser's fullscreen key (F11 for FF)]
screenshot
Attachment:
arctic_touchUp.jpg


Next a screenshot that nicely exposes the richness of rivers and lakes
reflecting in the sun light:

The Great Lakes
===============

[by all means click on image and then hit your browser's fullscreen key (F11 for FF)]
Attachment:
great_lakes.jpg


Enjoy,
Fridger


As a retired cartographer, I enjoyed these images very much!


Top
 Profile  
 
PostPosted: Wed, 13-09-17, 7:32 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4514
Location: Hamburg, Germany
Many thanks bh! I hope things are going well over there ...

Cheers,
Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Tue, 07-11-17, 13:24 GMT 
Offline
User avatar

Joined: Sun, 09-03-14, 13:09 GMT
Posts: 53
Location: Hangzhou, China
Wow... This is just amazing. I've been waiting for this for quite a few years (since I was a naive high school student). Thank you for all those hard work!
Have you considered hosting the project on GitHub or something... Just can't wait to get a glimpse of the code. ;)

_________________
Hanghang


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 32 posts ]  Go to page Previous  1, 2, 3  Next

All times are UTC


Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group