PDA

View Full Version : new idea on create clouds


yhloon
03-05-2006, 01:02 PM
I'm glad I own a copy of dante include clouds shader, but i also face some slow rendering problem as well.

so I came out of an idea to create clouds, alot of clouds... and render fast. in this test, I'll did not use any third party shader, hope you all like it...

1. create ubershape-sphere, check cull backface.

2. this test, I use noise factory, but i think bumpy noise or Dimmed shader (Triade) might do a better job...

3. apply edge transparency (not the RT volume Tranparency) check render until you satisfy...

4. duplicate the same sphere, reduce a bit. parent it to the first sphere.
http://i30.photobucket.com/albums/c336/yhloon/clouds%20tutorial%20image/1_2sphere.jpg

5. duplicate these sphere, move and scale to form a cloud shape.(render approx. 1.minute)
http://i30.photobucket.com/albums/c336/yhloon/clouds%20tutorial%20image/first_ganerate.jpg

6. now my idea of speed render is normal maps. apply the normal shader to camera map to ganerate a clouds normal map
http://i30.photobucket.com/albums/c336/yhloon/clouds%20tutorial%20image/NormalMap.jpg

7. you might need to touchup a bit on the aplha channel, than apply the normal map to a plane...and also the clipping map as well:) (rendertime 20 seconds)
http://i30.photobucket.com/albums/c336/yhloon/clouds%20tutorial%20image/final.jpg

the above is only a basic concept of using normal map to help create clouds, with various different form of clouds and normal maps should create some nice looking clouds in a short time....

any comments?

Loon

bronco
03-05-2006, 01:14 PM
wow, this looks very promising! never thought about using normal maps!
great! now i have to do some testing myself. thank you, loon!

Igors
03-09-2006, 04:19 PM
any comments?Loon

Hi, Loon

We also experimented with "sphere-clouds", but we added our spheres inside program :) (attachment). Let us give you an advice (of course, you can accept it or no): run away from this ASAP! There is no more giant time-eater than this one! There is an illusion that a success is near, very near, but.. the life shows it never will be achieved. We name such things as "Fata Morgana"

Giacomo_M
03-09-2006, 06:10 PM
Igor-

I'd be satisfied if I could creat clouds as good as the one you just posted. How did you do it? Details, please.

GM

Igors
03-09-2006, 06:42 PM
I'd be satisfied if I could creat clouds as good as the one you just posted. How did you do it? Details, please.

Hi, Giacomo

That's experimental shader we wrote last autumn. It's based on absolute same ideas Loon explained above:

Virtual model: let's create a few basic spheres. Then let's replace each sphere with some set (like 3-5) sub-spheres. Let's replace each sub-sphere with.. etc. - just a fractal. There are some difficulties on this way: spheres should not be located "too compact" and not "too scattered" - they should be located in "interesting" order. For sample above: 10 basic spheres, 6 fractal levels, 173K of spheres total

Render: first we used volumetric for both - clouds and their self-shadows. However, then we switched to deep shadows technique for self-shadows, it's much faster and better. For sample above: render time = 1 min 35 sec (G4 733)

bronco
03-09-2006, 07:03 PM
hi igors, any plans on releasing this experimental shader? it sure looks interesting, ...and fast.

AVTPro
03-09-2006, 07:08 PM
Hi, Loon

We also experimented with "sphere-clouds", but we added our spheres inside program :) (attachment). Let us give you an advice (of course, you can accept it or no): run away from this ASAP! There is no more giant time-eater than this one! There is an illusion that a success is near, very near, but.. the life shows it never will be achieved. We name such things as "Fata Morgana"


This is extremely realistic, Igor.

Can you post screen capture of the settings? Maybe it can be diluted a little to be faster.

Igors
03-09-2006, 07:17 PM
hi igors, any plans on releasing this experimental shader? it sure looks interesting, ...and fast.

No. It's just experiment, not a commercial product (sometimes we write progs for self-enjoying:))

Can you post screen capture of the settings? Maybe it can be diluted a little to be faster..

Alonzo, what settings you talk about? Shader's settings? or what? - not clear, please explain :)

bronco
03-09-2006, 07:23 PM
No. It's just experiment, not a commercial product (sometimes we write progs for self-enjoying:))


how self-centered! ;) make it open-source!
ok, kidding here, but i think the concept is nice and lots of people would love to play with this. i know that i would!

AVTPro
03-09-2006, 07:32 PM
I thought it was Kmyst setting. I don't know.. Just thought the project file would be a good study.

Igors
03-09-2006, 08:17 PM
I thought it was Kmyst setting. I don't know.. Just thought the project file would be a good study.

Ops :) Alonzo, those spheres are virtual, they are "inside shader", not in prj. And it's experimental shader, not KMyst

AVTPro
03-09-2006, 08:32 PM
Ops :) Alonzo, those spheres are virtual, they are "inside shader", not in prj. And it's experimental shader, not KMyst


Well, I'm with Bronco on this. I love you work Igors. I would be happy to have a part in creating with it. I don't have a fast Mac but I am looking for the perfect cloud solution. This looks like it. Would love to use it on this project.

I worked with Kmyst, can this do clouds formation with other object beside spheres? That is what I would like to try.

halfworld
03-09-2006, 10:27 PM
I want that shader :drool:

Igors, you rokk :)

Loon, ace trick!
:beer:
Ian

Igors
03-09-2006, 11:26 PM
Thx for good words but let's think what you would do with such "spheres-clouds"?

- you cannot fill the sky with them;

- you cannot animate them (spheres' moving isn't a serious animation);

- yes, they are "true virtual 3d spheres" and you can fly thru/around them. But.. spheres' sizes are fixed, so in close view you see big spheres instead of attractive details;

So, why do you need them? To create a single cloud? But it's much simpler to use a photo :)
That's why it's an experiment, not a practical product.

BTW: there is another one defect of this way (not so obvious). The big problem is "no focus/accent" or, in other words, an absence of auxiliary jagged clouds/pieces (spheres cannot help here). Illuminated cumulus require some "environment" (sorta noisy) to emphasize their beauty. Without it the logic/composition is missed. It's like adding a lot of light has no effect yet if there is no shadow/darkness (nothing to compare with). Hmm... sorry, maybe our explanation is not very clear :)

Vizfizz
03-09-2006, 11:51 PM
If you haven't guessed by now guys....artists don't care about those kinds of specifics. We're usually only concerned about one thing. The end result. If it looks good, we want it. lol.

You might consider this custom shader of yours a nifty little experiment in programming, but rest assured, someone would probably make good use out of it.

You guys probably have several little "nifty" experiments that you're sitting on thinking they wouldn't be of any use..... who knows what you're hiding. :bowdown:

Come on... we know you have something we want. :) lol

AVTPro
03-10-2006, 05:08 AM
Ops :) Alonzo, those spheres are virtual, they are "inside shader", not in prj. And it's experimental shader, not KMyst

This is one case where I am very careful about what I ask for because I have worked with cloud shaders before. As an artist one thing that is important to me about computer tools is interactivity, ease of use, and speed. Cloud shaders and volumetric effect is something I am even more conscious of because of the speed of my computer. I keep saying to myself, I am going to do a lot of compute-intensive render effects when I get a faster Mac, like fire, smoke and clouds. With that said, my artistic interest in a this particular cloud shader would be for a very specfic effect and not just because it's fun or look greats. ...which is normally the reason Brian :)

If the shader does only clouds for skies, right now it wouldn't be so important because i have Kmyst for that.

However if the shader can do clouds formations based on geometry. I would be very interested. If it can do a volumetric fill of any shape I model and not just spheres.

Heres a sample I did with Kmyst, but I didn't get the smoothing at that edges.

http://i47.photobucket.com/albums/f190/AVTPro/Trail320copy.jpg


Here's another sample I dea a couple of years ago. I think it would be more popular idea now with all the trapcode effect and animated strokes. I never finished this because i couldn't get Flexpath or Contortionist.

http://i47.photobucket.com/albums/f190/AVTPro/RainbowHeavenNew3_3.jpg


I understand that Igors new volumetric selfshading is hyper real, but that isn't want I would need it for.

yhloon
03-10-2006, 06:57 AM
wow, didn't ecpect so many replys over one night...

Hi, Loon

We also experimented with "sphere-clouds", but we added our spheres inside program :) (attachment). Let us give you an advice (of course, you can accept it or no): run away from this ASAP! There is no more giant time-eater than this one! There is an illusion that a success is near, very near, but.. the life shows it never will be achieved. We name such things as "Fata Morgana"

Igor,

That clouds you show look extremely good, the Idea that I shown is so basic, and I've no time to explore deeply, but with your shader which is a great time saver, and the end result is much-much better than mine, why not make it available for other user.


- you cannot fill the sky with them;
I think The shader can use for the clouds close in front, at the back we can use the original EI cloud shader



- you cannot animate them (spheres' moving isn't a serious animation);
I do still image more than animation, I've no worry! hehehe :)



- yes, they are "true virtual 3d spheres" and you can fly thru/around them. But.. spheres' sizes are fixed, so in close view you see big spheres instead of attractive details;
than we find other way to achive that!

I belive your cloud shader are welcome to all of us, just release it...:applause: bu not too expensive:D

regards,
Loon

bronco
03-10-2006, 09:15 AM
- you cannot fill the sky with them;


why? i guess each plugin instance can only create one cloud, right? isn't it possible to use more than one instance?

- you cannot animate them (spheres' moving isn't a serious animation);
if it is possibel to use more than one instance (see first question) and the clouds are volumetric, than it should create some very nice effects to position 2 or 3 instance inside each other and animate the position and rotation of the whole plugin. i guess there will be some limitations, because of the internal self-shadowing, right? i'd say, worth a try.

- yes, they are "true virtual 3d spheres" and you can fly thru/around them. But.. spheres' sizes are fixed, so in close view you see big spheres instead of attractive details;

see above, if more than one instance is allowed (or maybe some of the parameters, like the fractal detail, are animateable) , this is not an obstacle :)


BTW: there is another one defect of this way (not so obvious). The big problem is "no focus/accent" or, in other words, an absence of auxiliary jagged clouds/pieces (spheres cannot help here). Illuminated cumulus require some "environment" (sorta noisy) to emphasize their beauty. Without it the logic/composition is missed. It's like adding a lot of light has no effect yet if there is no shadow/darkness (nothing to compare with). Hmm... sorry, maybe our explanation is not very clear

it is indeed not very clear, sorry, at least to me.

i guess i am so excited about this whole cloud stuff, because i think eias needs fast volumetric calculation. emphasis is on fast. your experiment looks as if it has just that potential.
so here are some suggestions:

if you don't plan to investigate this concept any further, release the plugin "as is". Maybe free, or very little money. just make it clear that there will be no further development.

if you plan to go on with this, build a beta! (or alpha for that matter)

i like to name rodeo as an example. it produced some noise in the community. even if it is stoped right now, because patrick has some other, more important, things to do, it produced some noise in the community.
it showed that there are indeed some interesting things in the works for eias.
i also think that the open beta for rodeo was a good idea (hint :) ).
it is limited, sure, but it gave the users something to play with and sometimes a project needs only that limited functionality.

Igors
03-10-2006, 02:17 PM
Hi, Alonzo, Brian, Loon and Bronco

We think we understand your points of view. But please agree: "spheres-clouds" have a lot of limitations. Using of 2 and more shader's instances is not a solution cause they don't interact each with another. Ok, about this shader perspectives:

- free product: no way. Besides MrRevolver we are busy with a large free thing. Making 2 and more free things - that's too hard for us.

- little price (like $25). Hmmm.. it's possible but our enthusiasm falls down to zero when we remember a set of limitation. Imagine the product is announced with a cool cloud intro image. User sees it and imagines a cool "stormbringer" scene, he's full of bright hopes and buys the product.. Bang! no animation :sad: Bang! no multiply clouds :sad:. Looks like we foolished him, huh? Let doc explains what the shader can/cannot? Of course, but be sure - nobody is in a big hurry to read it. "I bought your product to make same nice results as the best EI artists do! Why I cannot?" It's always hard to answer for such questions, for this shader - in 10 times harder. Developers should think about all users (not about advanced only :))

AVTPro
03-10-2006, 10:03 PM
No one can cry about what it can or can't do for 25 bucks....not even me :)

So, if they need more they can buy kMyst, it does everything.

Lastly, even if you can't animate the clouds, can't you move the camera around it?


I see, nothing wrong with that. Just fix the final setting, add a few project samples files and let it go.

No tool is perfect, long as it gets the job done.

yhloon
03-11-2006, 05:20 AM
- little price (like $25). Hmmm.. it's possible but our enthusiasm falls down to zero when we remember a set of limitation. Imagine the product is announced with a cool cloud intro image. User sees it and imagines a cool "stormbringer" scene, he's full of bright hopes and buys the product.. Bang! no animation :sad: Bang! no multiply clouds :sad:. Looks like we foolished him, huh? Let doc explains what the shader can/cannot? Of course, but be sure - nobody is in a big hurry to read it. "I bought your product to make same nice results as the best EI artists do! Why I cannot?" It's always hard to answer for such questions, for this shader - in 10 times harder. Developers should think about all users (not about advanced only :))


$25 is not bad, but in case other user complain in the future... you need beta tester...and that is me hehehe...:twisted::scream:


regards,
Loon

BJMonkey
04-05-2006, 04:02 PM
Can't you just drop the shader in here for download "as is" now... just in here? And people who download it are held to an agreement that they won't send nagging e-mails?

We'd all like to play around with it :(

Igors
04-05-2006, 11:52 PM
Can't you just drop the shader in here for download "as is" now... just in here? And people who download it are held to an agreement that they won't send nagging e-mails?

We'd all like to play around with it :(

No, we cannot. Any product (count freebie) should be documented, should have samples/demo, and (most important) should have some "guaranteed functionality". Our shader has not all this and thus we cannot upload a "something not a product"

Not any prj is finished with a product, it's normal. But ideas of unlucky prjs can be interested/usable - at least they show what's on a way was selected :)

worx3d
04-07-2006, 06:10 PM
hmm.. I've been reading this thread with growing interest. I understand that the Igors don't want to release this shader as a freebie, after all even when it was an experiment and for their own joy as they say, it's hard work. Some developers release freebies, other don't and that's their choice.

But the idea of getting a good looking cloud shader with limitations for $25 is veeeery tempting. I can think of 1 project that I will start very soon that could make good use of it, and being a non commercial project (for fun) $25 sounds reeeally tempting.

I believe, Igors, that if you guys put a web page with big warnings about the limitations of the shader, and explain that the low cost of the shader is due to those limitations and nothing more should be expected, NO ONE should/will complain. I can assure you, whenever I need a more advanced cloud shader for a commercial project, I will turn to other shaders, like kMyst, AG Mondoclouds or Psyclone (or whatever it's called, it's a nice cloud shader).

But for 3D artists that don't do commercial projects, a $25, static and limitated shader could be all it's needed. I join the other users to encourage you to make a try and release it for a low price, like $25.00, with warnings and all.

In the unlikely case that someone calls complaining, please just send him/her to me and I'll kick his/her butt to the moon!! hahaha... really, for $25 I'm sure no one will complain.

cheers and please give it a serious thought.

Joel

bronco
04-07-2006, 08:39 PM
for 25$ i would order it this second, never complain and would try to make some nice use out of it. seriously, this is much to promising to never see the light of day, even if it is limited.

Igors
04-07-2006, 10:55 PM
In the unlikely case that someone calls complaining, please just send him/her to me and I'll kick his/her butt to the moon!! hahaha... really, for $25 I'm sure no one will complain.

cheers and please give it a serious thought.
Joel

Hi, Joel

Thx for your help proposition :) The main reason of our unwilling is: we ourselves hate products that have "nice gallery" but give user nothing concrete/guaranteed. Unfortunately,
Sky4 is exactly same thing: yes, after several hours of tries you can receive (maybe) a single cloud you like, but...

But who knows, maybe you are right: $25 is $25, and, after all, developers write their programs for people. So: in first half of summer MAYBE. Before this we must to finish some other projects.

SteveW928
04-08-2006, 09:25 PM
Hi Igors,

I think as long as you are honest about the state of the product you are delivering, folks won't complain. Advertise it as a 'tool for geeky animators' or something. Maybe create a page of site devoted to selling tools at lower costs that never made it to a full public interface or usability state.

Just write up some docs and make it known that anyone wanting to purchase and use it will have to really dig into it to get results.... only for the 'advanced and adventurous' crowd.

About the only 'complaints' I think youd hear then, would be requests to add features or update it. But, if it gets popular enough, then maybe it would be worth investing the time to make it into a more full product.

-Steve

CGTalk Moderation
04-08-2006, 09:25 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.