The same file as my other hurricane images, just a different angle. Took 5 hours to render.
...
what can i say ?!
great great render again !!!
I love it !
yea thanx for the wallpaper!...
Great image, although I would have liked less black and more amazingly realistic planet ;). You should *really* write a tutorial about how you achieve these effects. You don't even need to share your specific settings; documenting the thought process and logic in getting this would be enough.
Quote from: old_blaggard on January 29, 2008, 04:42:17 PM
Great image, although I would have liked less black and more amazingly realistic planet ;). You should *really* write a tutorial about how you achieve these effects. You don't even need to share your specific settings; documenting the thought process and logic in getting this would be enough.
Thanks, I'm actually thinking about changing the image size and shape on this one and making a very large version (although I am limited to the size of the mask). I might try to come up with some kind of explanation later.
Why is the mask rez to small or it it a matter of RAM?
Well, I actually just checked my mask again and the resolution was 5600 something, so it actually shouldn't be a problem for the size I'm thinking of.
I think you should take this to the next level - an actual animation. Of course you could render it in a much smaller resolution.
The only drawback I see is there is some lost detail on the far side of the 'cane (likely due to the image map).
Either way, incredible work.
The other drawback I see is that I don't animation. ;D
peth details :D
have you tried looking down into the eye?
There is a separate thread of trying to do hurricane and tornado clouds (http://forums.planetside.co.uk/index.php?topic=3227.0). You seem to have solved this, at least at the global level. Perhaps your techniques could help those trying to solve the spiral cloud formation problems. Willing to share?
--Dale--
He's using a image file as a mask the thread you linked to there is talking about using functions to create one procedurally.
Oh, sorry. (Isn't that almost cheating?)
eh depends on how its done, as long as its note putting a real photo on a planet though PS or something its still counts.
Quote from: Bidmaron on January 29, 2008, 07:30:51 PM
Oh, sorry. (Isn't that almost cheating?)
Hey, don't rub it in. I don't like having to do anymore than anyone else. ;)
@ Will: I havn't tried to look into the eye yet. I'll have to try it sometime.
The difference between this and just directly using a photo of a tornado is still significant. The mask is used to generate shaping for real volumetrics, so the lighting and shading is accurate no matter what the camera or sun position. And although it is limited by the resolution of the mask (unless you do some clever detail blending with fractals), you can still view it from many angles, even fairly close-up, and maintain reasonable realism. A fully procedural hurricane would be even better of course, but it's also much harder to do realistically.
- Oshyan
cheating ? what the hell ???
he managed to do a hurrican ! using TG2 tools ! image map is a terragen tool, and even if it's not a procedural hurrican, it's the best circular clouds formation i've ever seen on Terragen !!!
oh, and it's not a game... so what the point about cheating ?
the final render is really impressive to me so... either procedural or using image map to do it... who cares ? it worked !
Quote from: nvseal on January 29, 2008, 07:47:03 PM
Quote from: Bidmaron on January 29, 2008, 07:30:51 PM
Oh, sorry. (Isn't that almost cheating?)
Hey, don't rub it in. I don't like having to do anymore than anyone else. ;)
@ Will: I havn't tried to look into the eye yet. I'll have to try it sometime.
I agree with seth93 your not cheating it is one of the tool Planetside has included for our use. The object is to get the most realistic render we can.
Also if the image has been converted to grayscale and used as a mask you should be able get closer to the eye and see it well. The darker parts of the image will cause the clouds thinner. sharp lines make straight cloud edges, and feathered lines make softer cloud edges. The key (and headache) is getting the cloud Density shader set correctly in ratio to the cloud depth.
"cheating" would be adding something with another software... like adding clouds with PS or Paint.net... i'll understand the critic "you cheated"... ^^
Seth, while I'm astounded by the image nvseal created, the point of TG2 is procedural graphics...so I understand why someone might think of using an image as cheating. Nevertheless, if I understand the effects industry correctly, the only thing that matters is the end result.
Here's a question (perhaps for Planetside) - Which (image or procedural) would work better for a hurricane, if the end result needs to be an animation like in some of the movies we've seen?
mmmh i disagree... image map IS a tool of TG2... so we have (or at least we can) use it to improve our renders...
anyway... nvseal, i loooove this wallpaper !!! :D
animating hurricane with image maps would be pain in the ass....you would have to make mask for every frame and take too long...where you add couple nodes for rotation and the movement of hurricane and render away!...
I understand what he is trying to say. Personally, I don't like my images to be "tainted" by nonprocedural methods such as mask (cheating in other words). However, I also realize that for some things -- such as this hurricane -- the procedural option is just not viable (or at least not for me with my level of skills). The best option at this time is to have a mix of procedural and nonprocedural methods (the small clouds for example are procedural while the hurricane itself is made with the mask).
You image is awesome, nvseal.
I just wonder about animation. Otherwise, I don't think it matters.
yeah i don't think image map will be possible to make animation... or as stated by dhav, you'll have to make a mask for every frame...
Quote from: calico on January 30, 2008, 11:50:37 AM
Here's a question (perhaps for Planetside) - Which (image or procedural) would work better for a hurricane, if the end result needs to be an animation like in some of the movies we've seen?
I think it depends what you need the final shot(s) to look like, how much animation is needed (of camera, clouds, lighting), how close you need to get, and how much you want to be able to change the look of the shot during development. For the kind of images nvseal has created here, you only need to look at his results to see that his approach is very good. If you needed a small amount of motion to the hurricane, the image-based approach could be augmented by some small procedural warping of the cloud map (in TG if necessary).
A purely procedural approach, while academically a very interesting area of study, is probably not the ideal way to create this kind of shot if the end result is what matters most. It would be extremely difficult to make a purely procedural hurricane look this good. The way nvseal has done this would allow for a good amount of camera movement (we could probably get really close if the images were supplemented by extra procedural or image-based details), the lighting can probably be changed quite dramatically in a way that a 2D background would not allow, and if you wanted to change any part of the hurricane you could easily paint onto the texture map however you wanted. TG is being used to render the perspective and volumetric lighting, which is a perfectly valid use of the software whether or not procedural textures are needed.
Having said all that, I like to do things procedurally because that's what interests me :)
Matt
As the one behind version 2 of the procedural hurricane I have to say that I certainly do not consider the image mask method cheating; I just prefer procedural :)
There are a couple of points to remember:
The image mask method has been developed and improved over a long period where as the work on the procedural versions is only around a week old. Give us a chance, we don't even have a unified method yet ???
This is probably the main thing in favour of procedural: If you are using an image mask you need to create a new mask for each new hurricane where as with the procedural approach you can have a new hurricane by adjusting a couple of parameters and or seed values.
Ultimately with the much talked about macro wrapper for node networks including a hurricane will be as simple as adding a normal cloud layer. Insert the macro / clip file, set the parameters; size, position, twist etc. Job done ;D ;D ;D
Firstly... great image, as usual :)
As for image maps being cheating ???... TER's are essentially image maps.... leaf/bark textures.... flow maps from WM... satellite imagery-derived masks... OMG I'm a cheat ;) Hands up all those who have used images for stone textures/displacement maps?
Yes I like to do things using procedurals as much as possible, but at the end of the day there are some things you either can't do, or shouldn't do (for performance reasons) using only procedures.
You could animate an image mapped hurricane if you really tried hard enough...
- Generate procedural masks to control the spread of cloud, contrast/black level/ white level of mask image
- Modify mask image between frames with ImageMagick and/or commandline morphing utility to generate variations of the image on the fly
- Add fractal components to cloud mask and animate those
- .. and of course your standard cloud animations
Put them all together and you could generate a reasonable animation. Changing a deep low pressure system into a full blown hurricane using only procedures would also be technically quite difficult as the function describing the hurricane will be different to a mask defining the early stages.
I certainly doubt that there is any competition between the two hurricane models as to which one is better... they're just two different ways of trying to get the same result. I'm watching both with great interest ;D
Quote from: bigben on January 30, 2008, 06:20:27 PM
I certainly doubt that there is any competition between the two hurricane models as to which one is better... they're just two different ways of trying to get the same result. I'm watching both with great interest ;D
I hope I wasn't implying that there was any kind of competition :o
The procedural version; at least the one I am working on, is basically generating an image mask procedurally on the fly so has definite prominence implications. :(
I am going to keep at the purely procedural version to see how far it will go but a better approach may be to use procedural methods to produce an image mask using an external application. POV-Ray because I know how to use that :D
As the one who made the comment:
1) As I said, I was confused and thought he had procedurally done it. I'm still not sure I understand what he did. What I think is that he took a relatively featureless planet, used a hurricane image and an alpha map derived from it, and applied this result to the terragen planet. Certainly not trivial.
2) My original words were "Oh, sorry (isn't that almost [emphasis added] cheating?)"
3) My apologies to the artist.
--Dale--
Quote from: Bidmaron on January 30, 2008, 07:55:53 PM
As the one who made the comment:
1) As I said, I was confused and thought he had procedurally done it. I'm still not sure I understand what he did. What I think is that he took a relatively featureless planet, used a hurricane image and an alpha map derived from it, and applied this result to the terragen planet. Certainly not trivial.
2) My original words were "Oh, sorry (isn't that almost [emphasis added] cheating?)"
3) My apologies to the artist.
--Dale--
No apologies needed. What I did was get a satalite image of a hurricane off the internet from nasa. I then made the image black and white and imported it into TG 2 with an image map shader. Using a new camera, I projected the image down towards the planet, then I attached the image map shader to a color adjust shader which then went into the cloud shader taking the place of the cloud density fractal. That's pretty much it.
Well, thanks for the explanation. It's a beautiful image no matter how it was done. Thanks for sharing. Sorry for the confusion.
--Dale--
Thanks for explaining your method, nvseal.
I was looking around this morning to see how the clouds held up from closer angles and let this render during classes. This is very low quality but it does give an idea. There some parts that don't look exactly correct but it does look pretty interesting I think. ;D
Yep. Matt's answer is what I wanted to hear.
Nice to know how you did it, more or less what I thought you had done.
A perfectly valid technique. Total cheat ;D but what the Hell it works ;D ;) ;D
I am not going to give up on the "Proper method", one day we WILL HAVE computers powerful enough ::)
We already do, the flaw is simply I would like to live long enough to see the final render.
LOL
Quote from: Will on February 01, 2008, 07:28:28 PM
We already do, the flaw is simply I would like to live long enough to see the final render.
Quote from: Will on February 01, 2008, 07:28:28 PM
We already do, the flaw is simply I would like to live long enough to see the final render.
Science is working on that (Allegedly).
I was hoping for real time "I want a whole electronically synthesised universe in my office" ;D
here:
http://www.nbos.com/products/astro/astro.htm
http://www.lemoneyes-radio.com/stargen/
Nvseal, I think the close-up version also looks quite good and deserving of a higher detail render. Quite honestly I think this thing could be animated. You're not going to see rotation on a hurricane in realtime from orbit - it would have to be rotating at 1000's of miles per hour for that to happen. So all you need to worry about is camera movement. Could be quite cool, or maybe it would just expose the flaws in the model. ;)
- Oshyan