• Announcements

    • khawk

      Download the Game Design and Indie Game Marketing Freebook   07/19/17

      GameDev.net and CRC Press have teamed up to bring a free ebook of content curated from top titles published by CRC Press. The freebook, Practices of Game Design & Indie Game Marketing, includes chapters from The Art of Game Design: A Book of Lenses, A Practical Guide to Indie Game Marketing, and An Architectural Approach to Level Design. The GameDev.net FreeBook is relevant to game designers, developers, and those interested in learning more about the challenges in game development. We know game development can be a tough discipline and business, so we picked several chapters from CRC Press titles that we thought would be of interest to you, the GameDev.net audience, in your journey to design, develop, and market your next game. The free ebook is available through CRC Press by clicking here. The Curated Books The Art of Game Design: A Book of Lenses, Second Edition, by Jesse Schell Presents 100+ sets of questions, or different lenses, for viewing a game’s design, encompassing diverse fields such as psychology, architecture, music, film, software engineering, theme park design, mathematics, anthropology, and more. Written by one of the world's top game designers, this book describes the deepest and most fundamental principles of game design, demonstrating how tactics used in board, card, and athletic games also work in video games. It provides practical instruction on creating world-class games that will be played again and again. View it here. A Practical Guide to Indie Game Marketing, by Joel Dreskin Marketing is an essential but too frequently overlooked or minimized component of the release plan for indie games. A Practical Guide to Indie Game Marketing provides you with the tools needed to build visibility and sell your indie games. With special focus on those developers with small budgets and limited staff and resources, this book is packed with tangible recommendations and techniques that you can put to use immediately. As a seasoned professional of the indie game arena, author Joel Dreskin gives you insight into practical, real-world experiences of marketing numerous successful games and also provides stories of the failures. View it here. An Architectural Approach to Level Design This is one of the first books to integrate architectural and spatial design theory with the field of level design. The book presents architectural techniques and theories for level designers to use in their own work. It connects architecture and level design in different ways that address the practical elements of how designers construct space and the experiential elements of how and why humans interact with this space. Throughout the text, readers learn skills for spatial layout, evoking emotion through gamespaces, and creating better levels through architectural theory. View it here. Learn more and download the ebook by clicking here. Did you know? GameDev.net and CRC Press also recently teamed up to bring GDNet+ Members up to a 20% discount on all CRC Press books. Learn more about this and other benefits here.

Archived

This topic is now archived and is closed to further replies.

Filip

Sky-rendering techniques

138 posts in this topic

For some new ideas on perlin noise you should look at his (Ken Perlin) new paper: http://mrl.nyu.edu/~perlin/paper445.pdf

You should never let your fears become the boundaries of your dreams.
0

Share this post


Link to post
Share on other sites
..just wanted to share a link on the Intel''s site for those of you who''s interested:

They have an article and a demo on "Generating Procedural Clouds in real time on 3D HW" as well as other interesting stuff. Check it out at:

Intel Developer Servicies/Training/Software Development/Games/(Developer Centers/Games/Graphics)

OR here''s a direct link:

http://cedar.intel.com/cgi-bin/ids.dll/topic.jsp?catCode=CLH

OR direct link to the article:

http://cedar.intel.com/cgi-bin/ids.dll/content/content.jsp?cntKey=Generic+Editorial%3a%3aclouds&cntType=IDS_EDITORIAL&catCode=CLH

They also have a Webcast presentation on Procedural 3D content generation somewhere(sorry I forgot where, just browse) where they talk about clouds as well.

...Also if you guys ever checked out Nvidia''s Effect browser - under Effects/Animation/Atmospheric/Cloud Cover - they do it with vertex shader. Although it works only on the beter video cards like Ti and GTS it is damn fast and creates shadows on an "underlying" mesh. (yet i could not figure out if they have correct shadows on clouds themselves)

0

Share this post


Link to post
Share on other sites
I know a site with some pretty nice stuff about terrain and clouds rendering. Check it out here. Only problem, it's on geocities, so i think the bandwidth is *very* limited.. the screenshots might not appear.. if it happens, try later.

I'll include a screenshot here:

[Removed due to geocities crap, see next post]

Y.

[edited by - Ysaneya on June 1, 2002 2:47:08 PM]


[edited by - Ysaneya on June 2, 2002 6:40:54 AM]
0

Share this post


Link to post
Share on other sites
Ah yes, that''s a lot better.. i''ll remove the image from my previous post, since it''s using space for nothing.

Y.
0

Share this post


Link to post
Share on other sites
Dammit...

Just added this thread to my bookmarks.

By any means, Yann, write that tutorial
0

Share this post


Link to post
Share on other sites
I''ve been pretty much keeping up with this extremelly interesting post but I have a few questions. Please excuse me if these have been covered (I looked and really haven''t been answered).

So how do you animate the cloud Yann? Do you just animate the Perlin Noise and thats it? Can you just (linearly) interpolate 2 Noises together, or maybe more? How can you simulate this scenario: It starts out a bright sunny day, later it gets cloudy, and finally at dusk all hell breaks loose with violent clouds (I won''t mention the lighting :-).

How do you shoot rays through the 2D Plane. You mentioned voxels. So you break up the noise texture as a plane (when it is in 3D Space) into a bunch of Axis Aligned Bounding Boxes or something similar maybe? This step really stumps me. And you do this to simulate scattering, right? Would this be an Isotropic or Anisotropic process?

Also, maybe off topic. Have you sen any of the research using particles to simulate clouds with billboard impostors (and VFC) as performance enhancements. It creates realistic 3D volumetric clouds and they look amazing, but for a game with a ground perspective, seems a bit useless. But is it really? Have you done any research into performance wayoff''s? For something that looks better, do you think it is too expensive or would the perlin noise calculations compare?

Thanks and sorry if the questions are a little lame! ;-)
0

Share this post


Link to post
Share on other sites
Dirge, I''ll let Yann answer the other questions, but the particle->imposter cloud paper was the one he referenced at the beginning of the thread as his source for the scattering approximation formula.
0

Share this post


Link to post
Share on other sites
In order to animate perlin noise you use another dimension. So if you need a texture with noise (2D) you add another dimension, get noise in 3d, and make the third dimension time. As time moves on, the clouds animate. Perlin noise IS smooth blending of the noise. So the third dimension will slowly blend two different noises.

Regarding different weather, you slowly change the density and flufiness values.
0

Share this post


Link to post
Share on other sites
quote:

So how do you animate the cloud Yann? Do you just animate the Perlin Noise and thats it? Can you just (linearly) interpolate 2 Noises together, or maybe more? How can you simulate this scenario: It starts out a bright sunny day, later it gets cloudy, and finally at dusk all hell breaks loose with violent clouds (I won''t mention the lighting :-).


Linear interpolation between 2 Perlin noise sets is just a very basic idea. An extension would be full 3D perlin noise, as kill suggested. Besides the noise set, you can also modify the exponential ramp. You can interpolate different octaves using different functions. You also have tons of alternative interpolators than simple linear ones (even fractal ones).

I know I just barely mentioned the topic of cloud animation in this thread, but I would get some legal troubles with my company if I would go into details. Just one thing: if you are creative in the use of your noise and exponent interpolators, then you can actually create a whole weather simulator. Scenarios, such as your''s above, wouldn''t be a problem with those, you could even go much further... Just hypothetically speaking of course... (My boss reads the forums Hi Alex !)

quote:

How do you shoot rays through the 2D Plane. You mentioned voxels. So you break up the noise texture as a plane (when it is in 3D Space) into a bunch of Axis Aligned Bounding Boxes or something similar maybe?


Not really bounding boxes, more an axis aligned voxel field. Kind of like an inverted heightmap terrain.

quote:

This step really stumps me. And you do this to simulate scattering, right?


Yep. The voxel field is used to approximate the (otherwise continuous) multiple scattering integral over a discrete data field.

quote:

Would this be an Isotropic or Anisotropic process?


Which part do mean ? If applied to the tracing itself, it is isotropic in direction (assuming that your grid spacing is equal in all directions).

quote:

Also, maybe off topic. Have you sen any of the research using particles to simulate clouds with billboard impostors (and VFC) as performance enhancements. It creates realistic 3D volumetric clouds and they look amazing, but for a game with a ground perspective, seems a bit useless. But is it really? Have you done any research into performance wayoff''s? For something that looks better, do you think it is too expensive or would the perlin noise calculations compare?


As already stated, I use a similar technique (the Harris/Lastra paper) to calculate shading. Using true 3D clouds from a ground viewpoint wouldn''t be very useful: you''d wate resources and the visual results won''t be very different. Creating impostors on the fly also make animation more difficult. 3D clouds are interesting, if you do things like flight simulators though.

/ Yann
0

Share this post


Link to post
Share on other sites
Yann - can you give some links to some papers on the techniques that you use ( multiple scattering - etc ), just on generally what''s going on... Cheers

Death of one is a tragedy, death of a million is just a statistic.
0

Share this post


Link to post
Share on other sites
quote:

I know I just barely mentioned the topic of cloud animation in this thread, but I would get some legal troubles with my company if I would go into details. Just one thing: if you are creative in the use of your noise and exponent interpolators, then you can actually create a whole weather simulator. Scenarios, such as your''s above, wouldn''t be a problem with those, you could even go much further... Just hypothetically speaking of course... (My boss reads the forums Hi Alex !)



I love answers that are riddles! I think I know what you''re saying but anyone wanna else take a crack at what he''s implying (so as to not get him sued helping us out)? I suppose I''ll need to get a prototype going before I can play with those values but thanks to everything shared in this forum I know mostly what to do! BTW, Yann, which game are you working on!? I wanna play.

quote:

Not really bounding boxes, more an axis aligned voxel field. Kind of like an inverted heightmap terrain.



/:-( <- Still stumped.
I can visualize the upside down heightmap but I''ll admit I don''t know what a voxel field is (I use the world voxel to describe the boxes in an octree, only definition I know). I''m sorry but can you explain this step a little more (if not I understand, anybody got some idea''s)? Luckily I don''t need any cloud scattering yet (still need the clouds), but I''ll definetly want to liven up my sim with clouds with silver linings (that''s what it''s for, right).

Also, I''m sorry if you had covered this too (I can''t remember and don''t feel like reading the many posts already on this thread, although I''ll have to eventually) but minus the clouds, have you talked any about the colors associated with the sky? You''re using a sky plane right (or dome, doesn''t matter)? How do you create such beautiful gradients and color contrasts (your pics were gorgeous). Day/Night cycles rock, and I''d probably get that going before clouds anyways. I suppose you just modulate the pixel colors according to time of day (so if dawn interpolate red/yellow/blue across sky to either ends (east-west). So would you blend at most 3 colors (at dawn or dusk) and at like noon, there would be equilibrium (where the sky is completely light blue)?

Thanks a bunch, I can''t wait for your tut. When I got some free time maybe I''ll see what I can accomplish and post some code on my site (CodeFortress.com). It''s amazing how many game makers still think a skybox will suffice or just blend a cloud layer (in some cases) when you can make such realistic scenes using this method.
0

Share this post


Link to post
Share on other sites
Concerning animating clouds, I don''t know what Yann''s doing but here''s what I did

1. I used perlin noise in three dimensions and used the third dimension as time. Depending on how quickly the third parameter to the noise function changed the clouds will animate slower/quicker.
2. I used a wind vector and used the texture matrix to translate the cloud texture. Using the wind vector you can modify how quickly and in which direction the clouds are moving.
3. What you have left are the density/fluffiness parameters to the exponential function, the frequencies of the noise and cloud color. You can modify the fluffiness and the density parameters together to achieve denser clouds. During the storm the clouds will obviously take up the whole sky. You can modify the cloud color to make them darker or lighter. I didn''t really modify the frequencies procedurally because they''re very hard to control, it''s something u just play around with and set for life (IMO).

Now, with those values you can do some pretty cool tricks. If you want a storm, modify the third value to the noise function more quickly and increase the wind speed. Also, make the clouds darker and mode dense. If you want a bright sunny day, the clouds will move slower, will be lighter and less dense. In some cases you might want the clouds to move and animate really fast but be light (as if you filmed them for days and show the film in 2 minutes). You really have to play around with the values to figure it out.

I found modifying the values procedurally is very hard to do, so I found some really good values for different weather conditions, and when I want the weather to change I just interpolate between them. By changing the time period of this interpolation you can achive really quick change of weather or a graduate one.

Concerning the voxel fields, it''s really beyond the scope of this thread. You should look up rendering voxel based landscapes to get some idea. There are a lot of tutorials on it. The fraze "The voxel field is used to approximate the (otherwise continuous) multiple scattering integral over a discrete data field." is really a fancy way to explain something (more or less) simple. I''ll try to explain it in simple terms. It was already mentioned above in this thread, so you might wanna look above for a more advanced discussion (with pics!!!).

Picture standing in the middle of a plain. You look up and see clouds with the sun being directly overhead. Directly overhead there is a point on the cloud that''s really dense. That means there are a lot of cloud particles higher then that point. Picture the sun ray go through all these particles in a straight line to you. Because there are a lot of particles on its path the cloud will appear dark to at this point since every little particle takes a way a little bit of brightness. Now the point right next to it is not so dense (because there are fewer particles above it) so the cloud at that point seems much brighter (there are less particles to take away brighness from the ray at that point).

These particles are continious. That means that if you take two points really really really close to each other, you can always find another point in between. But we can''t represent that in limited texture memory, so we use descrete values. We store how dense the cloud at a certain point is in the alpha component of a texel at that point. When we render the clouds the alpha component defines transparency. We can''t do true voxels in realtime, so we just have a 2D texture.

Now we need to shade the clouds (find their brightness). How do we do that? Obviously the higher the alpha, the more particles are above the cloud, so the darker it should be. So at every point on the texture we look at alpha and for alpha == 255 we make clouds real dark and for alpha == 1 we make them real light. At alpha == 0 there are no clouds

Now if you look at somewhere in the middle of this thread you''ll see that this approach will lead you to incorrect results. The clouds far away will look flat. Why? The reason for that is that if you calculate your shading this way you''ll act as if the sun is directly overhead each particle. That''s obviously not true. This is where ray tracing comes in. You create a voxel space (you defined voxels correctly) from the alpha component in the texture. This way you have a 3D representation of the particles. Then you trace a ray from each lowest voxel to the sun. On the way of a ray you calculate how many particles intersect the ray. Remember, each particle takes away a little bit of brightness, so depending on how many particles are on the way of the ray you calculate the shading. This will give you correct results.

If you have any more questions about this explanation (you''ll probably have a lot, I''m not very good at explaining things), post here If you have questions about voxels and tracing a ray throught them you should probably start another thread.

Yann: I really want to know what you''re working on too. So if the NDA doesn''t stop you, let us know about as many details as you can. When''s the game coming out? BTW, how did your metting with the publishers go?
0

Share this post


Link to post
Share on other sites
No that was a great explanation, I think I get it a lot more now! Thanks! I'm still trying to picture ways to do the conversion from the 2D Texture (Texels) to 3D Voxels, but it'll come to me.

On a side and strange note, as I searched for more info on realistic simulated sky techniques a few days ago, I came upon a Voxel tutorial on Flipcode, here's the link for the voxelly inclined, such as me: http://flipcode.com/voxtut/

Also this link has probably been thrown around too (I REALLY need to read this whole thread again or at least copy it to a nice long text file), but here is an excellent resource on everything that's been talked about here: http://freespace.virgin.net/hugo.elias/models/m_clouds.htm

There was a decent one up on the Intel website too but it would never let me download the demo or PDF. Lots of info (pictures did not look a 1/100th as good as the one's posted here though).

EDIT: I finally found DOS4GW.EXE (on the net: http://www.gameprogrammer.com/2-frust.html, download the demo, it comes with it) and I ran the demo from hugh elias' website and it is amazing! Everybody should check this out. The resolution sucks but it looks like a fast-motion captured video of a sky.

[edited by - Dirge on June 7, 2002 1:29:02 AM]
0

Share this post


Link to post
Share on other sites
Brinkster killed my account, so all my images in this thread went down. I just relocated everything onto a (hopefully) better webhost, I hope they'll be up a bit longer now. *sigh*

/ Yann

[edited by - Yann L on June 7, 2002 8:15:33 PM]
0

Share this post


Link to post
Share on other sites
quote:
Original post by python_regious
Yann - can you give some links to some papers on the techniques that you use ( multiple scattering - etc ), just on generally what''s going on... Cheers


Hmm, most papers about atmospheric scattering unfortunately contain very heavy math and tend to overcomplexify the problem. They are mostly useless to realtime programming. Here is a small collection of papers that focus more on the practical aspects:

* "Realtime cloud rendering" by Harris/Lastra (the one I linked at the beginning of the thread).
* "A Practical Analytic Model for Daylight" by Preetham/Shirley/Smits (this one is a very good mathematical explanation of the basic effects in the atmosphere, and how to program them: aerial perspective, atmospheric spectral radiance, Rayleigh scattering, etc).
* "Modeling and rendering of various natural phenomena consisting of particles" by Nishita/Dobashi (Extends the idea to various volumetric natural effects, and rendering them)
* "Display method of the sky color taking into account multiple scattering" by Nishita/Dobashi (Everything up from the basics. 100% pure math).

They are all available online (although some of them might require an ACM account, not sure). Search for them through google or at GraphicsPapers.com.

quote:

BTW, Yann, which game are you working on!? I wanna play.

Yann: I really want to know what you''re working on too. So if the NDA doesn''t stop you, let us know about as many details as you can. When''s the game coming out? BTW, how did your metting with the publishers go?


Thanks for asking Unfortunately, I can''t comment. Just this: it will be in realtime 3D (doh, obviously) and it will feature some extremely advanced graphics

quote:

Also, I''m sorry if you had covered this too (I can''t remember and don''t feel like reading the many posts already on this thread, although I''ll have to eventually) but minus the clouds, have you talked any about the colors associated with the sky? You''re using a sky plane right (or dome, doesn''t matter)? How do you create such beautiful gradients and color contrasts (your pics were gorgeous).


Oh yeah, right, I forgot about this one.

For the colours, I use a skydome. Just a simple hemisphere. The skycolour gradient is nothing but gouraud colours interpolated over the triangles of that dome. Now, it''s not only a simple interpolation between colour 1 and colour 2. You have to supply at least 4 to 5 colours to get a good result, since in reality the gradient isn''t linear and changes through different hues, depending on elevation.

For the day/night cycles, I simply interpolate (linear) between a large collection of good looking sky gradient sets, that were precalculated before. You can eg. use TerraGen, Bryce and co. for that task. You can also use real photographs of the sky.

There is one detail to be aware of: If you look at the real sky, you''ll notice that the gradient is not uniform over the whole ''hemisphere''. The colour changes are almost non-existant if you look straight above (at zenith), but the further down to the horizon you look, the larger and more compressed colour changes will be. This effect is caused by Rayleigh scattering.

To simulate this effect in a skydome, you need to tesselate it rather much, since otherwise you can''t reproduce the subtile and thin colour gradients at the horizon. But on the other hand, a high tesselation will waste faces for nothing towards the top of your dome. Colour changes are very smooth here, so a lower tesselation would be enough.

The best way to deal with that problem, is an anisotropic tesselation. This simply means, that you tesselate your dome stronger, as you approach the horizon. Your faces do not cover the whole hemisphere in a uniform way, but they ''compress'' towards the horizon. The compression ratio is a function of the spherical angle between the viewer (assumed fix in the middle of the dome), and a given position on the skydome: Cratio = f( elevation_angle ). f(x) is a compression function. I use an exponential one, but you could also use a sinusoid or a spline. Depends on the visual results you want. And since it''s a precalc process, performance is not an issue.

By using that anisotropic tesselation, you can get detailed high quality gradients, while using a minimum of faces.

/ Yann
0

Share this post


Link to post
Share on other sites
I was looking up on stuff Perlin noise and saw this article. Download the program and run it (you''ll need DOS4GW). It looks amazing! It''s some of the most realistic sky rendering I''ve seen.

http://freespace.virgin.net/hugo.elias/models/m_clouds.htm

Enjoy! =)

- Jay

"Strictly speaking, there is no need to teach the student, because the student himself is Buddha, even though he may not be aware of it." - Shunryu Suzuki

Get Tranced!
0

Share this post


Link to post
Share on other sites
I finally came up with a way to have good framerates with realtime weather changes on systems that don''t have shaders (so you can''t implement perlin noise in hardware).

Let''s say our texture for clouds is 512x512. Create two of these textures. Set a delta D, let''s say to 2 seconds. Now, generate perlin noise every 2 seconds. Every 2 seconds flip the textures, so you always have a new version of the noise and old. Throughout these 2 seconds blend both textures (can be done in one pass even on older cards) and interpolate alpha value throughout the two seconds. This way you won''t see a sudden change in the way clouds look.

With this method you''ll see a sudden FPS drop every 2 seconds because the noise is generated. So instead set up your loops so they break off after some noise is generated and start from where they ended on the next function call. This way you can stretch the noise creation across the 2 seconds (or whatever your D is) and you won''t see an FPS drop at all and will have great procedural clouds.

If you stretch your noise generation across the 2 seconds you have to create a texture in system memory and just write to it. This way you won''t slow down the hardware at all, it doesn''t have to wait ''till you unlock anything! Then at the end of the 2 seconds you just quickly lock it and send whatever it is you need to send down the bus. This should be really quick, I''m working on perfecting this technique right now.

Also, this technique can be used for most procedural textures that take advantage of perlin noise so it might make sense to generalize it somehow. I''m not exactly sure yet how to make this a general utility class, but I''m getting there

0

Share this post


Link to post
Share on other sites
kill: Sounds good but I believe you can''t leave a texture unlocked for 2 seconds. I believe that would severly degrade performance and (don''t quote me on this) would create memory errors and a possibe crash. Streaming this cloud info seems like the best way to do it though. Personally I don''t want to use vertex and pixel shaders until I have a version that is compatible with the wealth of hardware already out there. I would definetly support those options, but I agree with you, let''s get it working on as many comps as possible first.

Tell us what you''re able to implement with this method. It has much promise.

coderx75: I mentioned that demo as well as where to get DOS4GW about 2 posts above yours on the same page. Just thought I would mention that (seems humorous to me).
0

Share this post


Link to post
Share on other sites
Who said the texture had to be locked for 2 seconds? Just create an offscreen buffer, fill it, and when it''s time, upload it to the texture.

Y.
0

Share this post


Link to post
Share on other sites
This is starting to get a bit off topic (we should start another thread on this subject alone) but if you were to "upload" the offscreen buffer to the texture, I believe you still have to lock it to gain access to it''s pixels. Maybe I''m missing the point here? What do you mean by upload? I believe we could just do what Yann said though and render to a p-buffer, then apply it slowely to a texture (as kill said), but either way, in D3D at least, you have to ask for permission to modify a texture (lock it, then unlock or else). Oh wait, then again, if you render to a texture you solve this problem (no lock/unlock), but you still need to create the base perlin noise texture octaves.
0

Share this post


Link to post
Share on other sites
Simply do, what Ysaneya said: Fill up your Perlin noise (frame by frame) into a standard memory buffer. Then, if it''s done, lock a texture, copy your offscreen buffer to it, unlock it (in one frame). In OpenGL it''s even easier, just a glTexSubImage2D() call. If you make sure that your memory buffer pixel format matches the destination texture format 1:1, then this copy will be very fast.

/ Yann
0

Share this post


Link to post
Share on other sites
Dirge: you still need to lock and unlock the texture, but this way the hardware doesn''t wait for you. It has nothing to do with D3D or OpenGL, it''s a hardware thing.

When you write directly to video memory you lock the surface and then all rendering that involves this texture has to wait until you unlock. With the same texture in system memory plus two hardware textures you just write the noise to system memory, so eventhough you lock it, the hardware doesn''t have to wait since you don''t use it for regular rendering.
0

Share this post


Link to post
Share on other sites
A little more on cirrocumulus. No code changes, but fixed the parameter controlling cooling upon condensation (updated in above post). I believe Kaneko''s too-stringy clouds (on his convection web page) came from ignoring this parameter. Here''s a screenshot of the water droplet densities (cross-sections at two times).



Giving the sun perspective definitely helped give the clouds more depth. I also noticed that the exponential shading was blurring detail in darker cloud regions (while nicely accentuating the detail at cloud edges); I''ve tweaked this a little to maintain some detail but it could certainly be greatly improved.



A bubble model à la Neyret (Qualitative simulation of convective cloud formation and evolution) gives a useful hierarchy (for level-of-detail modeling, shading and rendering, for example?). Stable Bénard convection cells can be generated using the coupled map lattice model. I''m curious about how efficient his scheme is. It seems like it could be much more efficient than the full coupled map lattice model for large skies. But I''d like to see more complete results to see if it really gives good looks. Here''s a side-by-side comparison of a single rising Neyret bubble and an MYDN cumulonimbus cloud. Is there potential here?


0

Share this post


Link to post
Share on other sites
Oh..... my.... GOD!!!!! I managed to the post I posted in march this year, and look how many clever/interesting (some even brilliant) replys I got... I just have to give everybody that made a post in here a hand! Specially Yann L for you great replys! Gee, almost have to give myself a hand for starting it =)

Anyway, I think everybody that want to learn how to render good clouds in real-time should read this thread. VERY, VERY educating!!

So, thanks a lot... again!!!
// Filip
0

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.