ICE Realwave-like boat ocean simulations.

[xsiholic]

Hello ...

I am interested to know whether ICE can generate
realwave-like ocean simulations with splashing waves,
boat-wake splashes, and shoreline splashes.

Realwave can generate foam-maps and wet-maps.
Does XSI 7 have similar options?

If so, ... will XSI be able to detect Vue island shores
which are imported with the Vue xstream plugin?

In other words, can we avoid buying Realflow if we
buy and focus on XSI instead?

Thanks in advance for your replies and possible tutorials
showing splashes.

- Marco.

[farhaad_yousefi]

that would be interesting..
cause i saw an old ship and sea in demo something like pirates of carribian..
but mostly i think ice would be the whole community get hand in hand to improve and gather abilities ..and knowledge..to improve our solutions :yes:
beside this.. me to stay for some answers from professionals ..!!

[xsiholic]

The C4D plugin 'AddTheSea' creates the best examples of CG oceans I have seen.
You can really see the foam floating over the waves.

I hope the XSI experts can generate something similar with ICE.

http://www.add-the-sea.com/plugins/addthesea/galerie.htm

- Marco.

[thiago]

not exactly what you said but... http://www.kai-wolter.com/resources/xsi/ocean-in-xsi

AFAIK this is no ICE... It's a normal XSI operator.
But I have no doubt that it could be implemented in ICE.

As for "realflow type of thing".... ICE isn't exactly it.
The idea with ICE is that we gonna have (one day) a node that generate "water"... another that make "waves". Than you put it together and you have an ocean.

In other words... ICE is not a "simulation environment" that will make waves splashing around. ICE is the authoring environment where you can make all this effects happen.

[rconover]

The C4D plugin 'AddTheSea' creates the best examples of CG oceans I have seen.
You can really see the foam floating over the waves.

I hope the XSI experts can generate something similar with ICE.

http://www.add-the-sea.com/plugins/addthesea/galerie.htm

- Marco.
Hmm-there looks to be a lot of comprimizes in those examples.
Don't get me wrong-realistic water/foam is an arse to get right.
The very best is usually propriety software.

But those foam textures do not even look like part of the sim to me.
They look projected on the same geometry with a distinct fade effect
with some texture projection tracking the waves.
Looks like some cheats that weren't pushed hard enough to 'sell' to me.
So not organically realistic.

I can't say if ICE can do better. But this isn't the greatest benchmark IMHO.
Look at ILM or something.

[scaron]

its been discussed on the xsi mailing list... but these types of oceans wouldn't be doable using factory ice nodes. the key element in jerry tessendorf's wave simulation work is the fast fourier transform.

http://en.wikipedia.org/wiki/Fast_Fourier_transform

http://tessendorf.org/papers_files/coursenotes2004.pdf

you are probably going to have to program a custom ice node in c++ that does the FFT and returns the results to normal ice nodes. so you wont be seeing any oceans based on the iwave algorithm for a bit.

steven

[mocaw]

I thought that ICE was low level enough in its nodes that going to C was going to be a waste of time in a sense. This function couldn't be created with low level nodes?

[scaron]

I thought that ICE was low level enough in its nodes that going to C was going to be a waste of time in a sense. This function couldn't be created with low level nodes?

for this specific task the built in nodes aren't enough. why? i dont know cause i dont fully understand how to use FFT and use it in the graph, and trust me i have tried.

[thiago]

for this specific task the built in nodes aren't enough. why? i dont know cause i dont fully understand how to use FFT and use it in the graph, and trust me i have tried.

The FFT is just part of an Ocean Simulation tool.
FFT is a way of analyze and build the spectrum. So probably it would be best to have it as a node in the ICE tree (that just performs FFT for an array), and then be able to use it with other things.
Maybe the speed of a open FFT in the ICEtree wouldn't be the best, and a compiled node would be faster on this case because of the intense calculations performed by the algorithm.

But the nature of the FFT itself isn't anything from another world, it's heavy math but it's more of a series of multiplications (generically speaking) over Sin/Cos functions over time, this could get more complex when you need of break data into chuncks to be solved and etc. But again, ICE makes this part easier.

As far as I understand the algorithm it would be intense to have a FFT as a open compound, and not very useful since the core math need to be performed as a whole and it wouldn't help much to leave it as an open compound that people can change. But it seems very possible to me.

And yes ICE is very low level. People need to think in ICE as an interface for XSI internal arrays. What you do with the data is your problem. And it will take a while until we use half of the potential of ICE v1.0.

Right now we don't have this type of simulator, but nothing prevents it to happen... and if anything can be done to improve the framework, it will happen.

Calm down peeps, we gonna get there eventually... this is just the beginning.
Maybe we can just get a realflow plugin as an ICE node in the ICEtree?

And I bet a toolset to "fake" waves in a controllable way would be much much useful.

[zogun]

Hi all,

im new to ICE, but how about this water splashing effects? can it be done using XSI/ICE ???

http://i177.photobucket.com/albums/w208/Zogun7/RF4.jpg
*taken from RealFlow website

[aberek]

ABSOLUTELY!
It will take a little time to setup the simulation but it is very possible and prolly a bit more manageable with ICE than realflow. It's simply a matter of using the spawning nodes and some states with collision dynamics to get the desired effect.

ocean boat simulations are also possible, seeing as one was already done in ICE.


-Alex

[scaron]

http://www.xsi-blog.com/archives/273#comment-17321

glad to see patrick get the gerstner waves working... now the fft part for the full tessendorf implementation.

steven