It is currently Sat, 20-10-18, 2:55 GMT

All times are UTC




Post new topic Reply to topic  [ 68 posts ]  Go to page Previous  1, 2, 3, 4, 5  Next
Author Message
 Post subject:
PostPosted: Mon, 24-09-07, 0:55 GMT 
Offline
User avatar

Joined: Sun, 02-09-07, 20:19 GMT
Posts: 196
Location: Montreal
This would look very artificial and unsatisfying. We need sprites for this. But as you know, it's dead on the Mac ! :evil:

_________________
"Well! I've often seen a cat without a grin", thought Alice; "but a grin without a cat! It's the most curious thing I ever saw in all my life!"


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 1:12 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 20:59 GMT
Posts: 1605
Location: Norway
Cham wrote:
This would look very artificial and unsatisfying. We need sprites for this. But as you know, it's dead on the Mac ! :evil:


I wonder...
Remember our volcano, a few years ago?
If your new method is combined with a regular smoke mesh, maybe...

After all, i did manage to make decent smoke with just a regular 3ds file; the downside
was that the 3ds weighed 5 MB, and so was too impractical to use in an add-on:

Image

... Maybe it is time to have another go at it...

EDIT:
The smoke was generated by mapping one small smoke PNG onto a mesh that
consisted of several thousand polygons that were randomly folded into each other.

If one made a simpler version, and then used your technique to have it rotate
about itself...

- rthorvald

_________________
Image


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 2:33 GMT 
Offline
User avatar

Joined: Mon, 03-09-07, 23:01 GMT
Posts: 418
Location: Tuscany, Tyrrhenian Sea
Great shot, Runar. Pratically, you have placed a tree in the volcano! Cham, chromosphere tested 1.4.1; it work.

_________________
Never at rest.
Massimo


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 3:19 GMT 
Offline

Joined: Mon, 24-09-07, 3:13 GMT
Posts: 7
Wow, this is very similar to what I did with animating Jupiter's clouds years ago. Of course I didn't have the transparent sphere models back than that we do now so I never took it very far. I was toying with my new Red Dwarf texture again trying to do this, but I wasn't happy with the results. So thanks for stepping this up a notch. I can't wait to give it a try with said texture. If all goes well I may post a short video myself.

Rather than clutter up the thread, I will put my results of this here.

I dumped the videos on youtube as well as posting links to the AVI files.

You will see that I achieved a different look with my texture. It looks more like roiling flames rather than convecting plasma. Oh well it was fun to try it. I even added Runar's flares to it on the second video. So there many effects that can achieved with this. Hmmm, SOLARIS anyone?

Image

Image

Red Dwarf

Red Dwarf with Flares

Don. Edwards


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 11:05 GMT 
Offline
Site Admin
User avatar

Joined: Thu, 30-08-07, 22:52 GMT
Posts: 2726
Location: France, South, not far from Montpellier
Don, what is really great with Cham's trick is the nill momentum, cells seems to burst from the mantle to the surface, no general movement. In your example (despite being also nice), one can follow too much at least 2 layers of textures... That's what make the Cham's trick so good, at first sight no way to say how it is done, the surface become alive...


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 12:05 GMT 
Offline

Joined: Mon, 24-09-07, 3:13 GMT
Posts: 7
ElChristou,

I went back and looked at the Cham's video again and I can see were my problems are. My texture is to sharp and the contrast to high. Also my convective cells being more out of round really show up. Its a learning process like any other. You are right though, mine does need allot of work.

Don. Edwards


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 14:05 GMT 
Offline
Site Admin
User avatar

Joined: Thu, 30-08-07, 22:52 GMT
Posts: 2726
Location: France, South, not far from Montpellier
Don. Edwards wrote:
ElChristou,

I went back and looked at the Cham's video again and I can see were my problems are. My texture is to sharp and the contrast to high. Also my convective cells being more out of round really show up. Its a learning process like any other. You are right though, mine does need allot of work.

Don. Edwards


No you misunderstood me, what I point out, is that to achieve this stunning effect on stars, there is no other way than using those 6 spheres. Once one have this in mind, you only have to work on the texture and at best on the rotation speed. For example for a brown dwarf, I suppose a texture with a cell structure very dark (not too transparent) will be good to create a contrast game between the 6 layers and the color of the star lying below...

Now of course, if one break the nill momentum, then there is tons of test to do to achieve for example sand storms over Mars...


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 14:52 GMT 
Offline
User avatar

Joined: Sun, 02-09-07, 20:19 GMT
Posts: 196
Location: Montreal
There are more things to come, with the technique I exposed here : clouds motion on a gas giant (especially the "hot jupiters"). I intend to recreate (or simulate, if you prefer) Jupiter's great spot on an exoplanet (so not to interfere with our Jupiter). I'll use at least two rings of static clouds, rotating in opposite sense, plus a half-sphere rotating between them, around an orthogonal axis (it's a bit hard to explain in few words). I'll try to reproduce what is often shown on videos about Jupiter's spot in accelerated motion. I'm not sure this would be very convincing, though.

Also, take note that nothing prevents us to us MORE than 6 spheres, so more complicated effects with more layers and textures. Since the sphere model is pretty simple and frame rate effective, we should be able to use more than 6, without affecting significantly the frame rate in Celestia (however, I didn't tested this yet).

I also want to simulate some kind of heat flows on "hot Jupiters", so I'll have to break the "nil angular momentum" constraint. I also have another special effect trick with rotating spheres that I'm yet unable to explain (more experiments are needed in this case).

The only real difficulty with the rotating spheres technique are the textures to be used. They must be crafted with care (and I must admit I'm not an expert with Photoshop). In brief, here are some constraints that we need to remember, while designing a new texture :

- we need at least two PNG textures (wathever the resolution), lets call them A and B, associated to the +spheres and -spheres, respectively, plus the base texture mapped on Celestia's sphere.

- Begin with texture A. If you want to create texture B from a vertical flip of A, You MUST NOT duplicate the same texture, and use the symetry operations (flipping it horizontally, vertically or 180 degrees rotation, and using the Offset operations) while creating texture A, or else you'll get some weird unatural periodic patterns in Celestia, since the same textures are applied 2*3 = 6 times on the rotating spheres. This is a major constraint while creating the textures, unless you want to create some special patterns on purpose (like what we have on Saturn's poles).

- Ideally (but not necessarily), both textures A and B should be related in some way to the base texture, to match the colors, contrasts and shapes. This is where you need lots of testing, experiments, trials and mistakes, etc. Global transparency and details transparencies (holes) of A and B are very important, especially since you'll have six layers on top of the base texture.

In many cases, I used an already existing star texture (see the astar.jpg, bstar.jpg and gstar.jpg textures available for Celestia) and erased many details using the eraser stamp tool in Photoshop. Be carefull at the left and right edges (the texture must be seamless), and remove some large bands at the poles (up and bottom edges). The six spheres will take care of the poles and there wont be any "polar pinch effect". This is actually a good way, visually, to get rid of *some* of the polar deformations occuring on the base texture, when applied to Celestia's sphere : the six rotating layers are "masking" the base sphere a bit (depending of the global transparency of A and B).

_________________
"Well! I've often seen a cat without a grin", thought Alice; "but a grin without a cat! It's the most curious thing I ever saw in all my life!"


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 20:21 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4580
Location: Hamburg, Germany
Despite, Martin's VERY interesting approach, it might nevertheless be informative to have a look at white papers from NVIDIA developer site about how developers approach similar graphics challenges:

i) RealTime atmospheric effects!
ii) RealTime volumetric smoke etc....

The following VERY impressive discussions are based on Direct3D10, but have a look at the developed strategies and presented formulae! There is a wealth of exciting material that can no doubt be transferred to OpenGL 2.0...

http://developer.download.nvidia.com/pr ... Crytek.pdf
http://developer.download.nvidia.com/pr ... Fluids.pdf

Don't miss the most impressive images!


Bye Fridger

_________________
Image


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 22:05 GMT 
Offline
Site Admin
User avatar

Joined: Thu, 30-08-07, 22:52 GMT
Posts: 2726
Location: France, South, not far from Montpellier
t00fri wrote:
Despite, Martin's VERY interesting approach, it might nevertheless be informative to have a look at white papers from NVIDIA developer site about how developers approach similar graphics challenges:

i) RealTime atmospheric effects!
ii) RealTime volumetric smoke etc....


Yep indeed very impressive those real time rendering...

Now 1) this is lecture for Chris L., 2) I understand the need for "elegant/cleaver" solutions, but all those stuff is for the very last material... I suppose it would be better to always forget the last 1 or 2 years old technology and concentrate on older stuff that (almost) everybody is able to run...

Cham's trick despite being "bricolage" is not a so bad way to render decent stars on almost all config. Perso I doubt a bit of the result on other bodies (not being fictitious), but for stars it's definitively not bad at all...


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 22:28 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4580
Location: Hamburg, Germany
ElChristou wrote:
t00fri wrote:
Despite, Martin's VERY interesting approach, it might nevertheless be informative to have a look at white papers from NVIDIA developer site about how developers approach similar graphics challenges:

i) RealTime atmospheric effects!
ii) RealTime volumetric smoke etc....


Yep indeed very impressive those real time rendering...

Now 1) this is lecture for Chris L.,

I think I can understand PERFECTLY well the formulae exposed for getting those atmospheric effects. OpenGL is not THAT difficult to learn and apply either ...

see e.g. Vincent. He started at ZERO not long ago!

There is a host of respective code examples at the NVIDIA developer site.

Quote:
2) I understand the need for "elegant/clever" solutions, but all those stuff is for the very last material... I suppose it would be better to always forget the last 1 or 2 years old technology and concentrate on older stuff that (almost) everybody is able to run...

Cham's trick despite being "bricolage" is not a so bad way to render decent stars on almost all config. Perso I doubt a bit of the result on other bodies (not being fictitious), but for stars it's definitively not bad at all...


As I said, Martin's trick is great, but one should not forget about the real way to do this eventually...

Cheers,
F.

_________________
Image


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 22:35 GMT 
Offline
User avatar

Joined: Sun, 02-09-07, 20:19 GMT
Posts: 196
Location: Montreal
I agree. The star rendering in Celestia currently sucks, and someday, something should be done about it (at the code level), so we can have a more realistic rendering, for ALL stars. The trick I'm exposing is really just a "bricolage" (or a kind of "hack") which is applied to a few stars only, so it's not very satisfying.

_________________
"Well! I've often seen a cat without a grin", thought Alice; "but a grin without a cat! It's the most curious thing I ever saw in all my life!"


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 22:49 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4580
Location: Hamburg, Germany
Cham wrote:
I agree. The star rendering in Celestia currently sucks, and someday, something should be done about it (at the code level), so we can have a more realistic rendering, for ALL stars. The trick I'm exposing is really a "bricolage" (or a kind of "hack") which is applied to a few stars only, so it's not very satisfying.


Absolutely, Martin.

I just wanted to recall the target, which is feasible with the technology of today. In order to
really render stars naturally we DO need

High Dynamic Range Rendering. There is NO way around that.
http://en.wikipedia.org/wiki/High_dynam ... _rendering

Bye Fridger

_________________
Image


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 22:55 GMT 
Offline
Site Admin
User avatar

Joined: Thu, 30-08-07, 22:52 GMT
Posts: 2726
Location: France, South, not far from Montpellier
t00fri wrote:
ElChristou wrote:
t00fri wrote:
Despite, Martin's VERY interesting approach, it might nevertheless be informative to have a look at white papers from NVIDIA developer site about how developers approach similar graphics challenges:

i) RealTime atmospheric effects!
ii) RealTime volumetric smoke etc....


Yep indeed very impressive those real time rendering...

Now 1) this is lecture for Chris L.,

I think I can understand PERFECTLY well the formulae exposed for getting those atmospheric effects. OpenGL is not THAT difficult to learn and apply either ...

see e.g. Vincent. He started at ZERO not long ago!

There is a host of respective code examples at the NVIDIA developer site.


What's up tonight? ya un truc qui passe mal? :wink:

What I mean is that as far as I know, Chris L. is the one in charge for OGL stuff. Nothing more, nothing less.

(Remember Vincent already answered that what he learn was not much, just the necessary to be able to display a few lines/stills. I'm not denigrating here Vincent's efforts (I would be unable to do what he done), I'm just saying that actually one must have a huge background in math and now physic to be able to code those stuff. I'm sure you can understand and learn that, but please, you won't makes me believe that Mr everybody can learn and apply all this in from zero in... let's say... 1 year full time...)


Quote:
Quote:
2) I understand the need for "elegant/clever" solutions, but all those stuff is for the very last material... I suppose it would be better to always forget the last 1 or 2 years old technology and concentrate on older stuff that (almost) everybody is able to run...

Cham's trick despite being "bricolage" is not a so bad way to render decent stars on almost all config. Perso I doubt a bit of the result on other bodies (not being fictitious), but for stars it's definitively not bad at all...


As I said, Martin's trick is great, but one should not forget about the real way to do this eventually...


We do agree, but the only problem is that "eventually" right now, sounds to me not before 2 or 3 (or more) years... :cry:


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 24-09-07, 23:07 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4580
Location: Hamburg, Germany
ElChristou wrote:
What I mean is that as far as I know, Chris L. is the one in charge for OGL stuff. Nothing more, nothing less.


I am NOT thinking in terms of Celestia (only) ;-) . Also my Cosmological Visualization project will need
good stars and better atmospheres. Chris L. is NOT in charge of this. Forgot?

Quote:
you won't makes me believe that Mr everybody can learn and apply all this in from zero in... let's
say... 1 year full time...

Also I did not imply that 'everyone' should start coding that stuff.

I just did not agree that these new concepts displayed at the NVIDIA site for example are ONLY a
lesson for Chris L. ;-) . Not more not less.

Quote:

We do agree, but the only problem is that "eventually" right now, sounds to me not before 2 or 3 (or
more) years... :cry:



My own time scale is much better defined and certainly less than 2 or 3 years...


Bye Fridger

_________________
Image


Last edited by t00fri on Mon, 24-09-07, 23:13 GMT, edited 3 times in total.

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 68 posts ]  Go to page Previous  1, 2, 3, 4, 5  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