RESOLVING WATER PARTICLE PHYSICS IN 3D PROGRAMS
It is a pet peeve of mine that with all
of the so called 'advanced technology' available to the common
programmer today, that not a single soul has yet addressed water
particle physics within the 3D modeling community.
After all of these years, it seems they
have just barely figured out how to make water stick to its SELF and
that not in a truly accurate manner.
Hence, things like bubbles, turbulence,
and the all important, ever present issue of WHETTING remain a remote
dream it would seem.
To this date, not a single company has
come forth to challenge these gross disparagements between true
fluids and virtual fluids.
Because of this, along with poly body
fluid displacement physics flaws, 3-D rendering is still trite and
crude, with at best, limited interaction between solid bodies and
fluid bodies, at worst, they utterly fail to interact, and pass
through each other unfazed.
This makes realistic water scenes
impossible to bake.
In the end, it leads to the need of
doctoring, dubbing and mixing of live water interaction to create the
necessary final result.
Is there any hope?
Are there any companies out there with
a solution on the horizon?
Delicious Vodka DeBlair
Pentagon Recording Specialties
How do you correct this line-break/word-wrap problem now?
Am not certain about what to make of this complaint.
Having a professional background in fluvial dynamics and a smattering of particle physics, I would say that water particle physics IS constantly being addressed in the 3D world, and has been for at least the last fifteen years. Next Limit made the best breakthroughs early on by developing the idea of individual "intelligent" particles and controlling forces for them into a practical software product. (I've been working with it since Version 1.0.) (There were many earlier products and lab exercises, but all these were too limited to be practical.)
Then almost all the major software developers (Autodesk, etc.) developed their own volumetric technology products, although to achieve "fluid-like" effects rather than to replicate particle physics processes. But their purpose was to create 3D effects that looked like fluids. Next Limit added those kind of simulation solvers to their basic product, and continued to expand the abilities of their solvers along the lines of physically accurate particle physics. Lots of stuff going on in the major research labs, in terms of 3D fluvial dynamics, but all focused on individual phenomenon, such as bubble and foam formation. Some of the techniques adopted by companies such as Autodesk and Next Limit and incorporated into their products.
Then Houdini arrived with a fairly complete grid volumetric technology. Miles away from particle physics processes as we know them, but able to create large-scale effects (whirlpools, floods and smoke) in a computationally practical manner. Which is what most graphics artists want. ( Not as good as Next Limit's Realflow product in terms of ability to create a large variety of particle physics or fluid dynamic processes with lots of control features, but there's not such a large market for those features, of course.)
Anyhow, just my personal opinion from always watching the developments in the research labs, but it seems to me that people are always hammering on these problems of liquid simulations. That is, these things are always being addressed and have been under constant development for at least 15 years that I know of.
As a user of both Houdini and Realflow, I don't seem to have problems of creating and working with turbulence, wetting, absorbtion, fluid object interactions, or even bubbles. Also, I'm not sure why you claim that these things aren't done "in a truely accurate manner" - at least as far as I can tell as a professional hydrologist. Could you clarify this a bit ? Maybe give us some further detail or examples of what you mean. Just trying to understand you.
I know I am not the only human on earth who has observed water, many millions of scientists, artists, hair dressers, construction workers, farmers and pretty much everybody who has any cause whatsoever to think about how water [or any fluid really] acts in nature, and in the end, as of this date, having followed fluidics [or fluvian particles] in any of its many aspects, can honestly say that any present day fluid particle simulator comes close to the true behavior of water or similar polar fluids.
Likewise, most fire, dust and smoke particle routines fall far short of realistic as well.
You can argue that in your opinion they are real or 'real enough' and well, if so, good for you.
I have a more realistic outlook on life after spending over 4 decades working with water from multiple disciplines, including, specifically, fluidics, as they are all part of what I spent most of my life doing as an energy researcher, but beyond everything else, merely being a scientifically observant human being interacting with nature.
I could create a long list, and then spend at least a paragraph or three on each one explaining what it wrong with CG fluvian particles, but I have to think I should not need to explain the grossly evident to a person who has just declared that hydrology is a part of their own studied experiences.
When I see it done right, I'll let you know.
If you think I missed some place, feel free to cue me in on this grand new breakthrough, but likely as not, I've already seen it, and can poke holes big enough to put Lake Michigan though and not wet the sides [pun intended].
Ok, so just a sort of free-floating complaint that fluid displays are not realistic enough, yes? Was just trying to understand the point of your thread. That is clarified now, I think. Opinions are always welcome in this field.
Any chance we could see your own work in this field?
In the field of energy research you mean?
That is a bit off topic after you get away from the sheer point that its an experiential point of a person who deals with such things as whetting, hydrophobic, hydrophilic and hygroscopic materials, as well as water, solvents, surfactants, gases and oils, which in general, all have specific fluid particle behaviors.
Or, being obnoxious, were you asking me what my personal experience was with trying to wrangle low class low quality CG fluvian particle applications on a 3-D surface?
To which I can reply: I do not NEED 'experience' with the defective product to KNOW it is defective.
That is tantamount to me asking you for your physics degree if you complain its too hot outside.
No, I meant the former, your work in your professional field. You know, one Ph.D to another. Just kind of interested in the nature of your professional work and your professional interests.
Am not taking issue with your opinions about the quality and accuracy of current fluid simulation software. Your opinions are your own. You're entitled to them and entitled to air them here. We all respect that, I believe. Lord knows there are a thousand issues yet to be overcome with software development in this field. One observation I might have for this discussion is that much of the phenomenon you have mentioned here is relatively small in scale, compared to large explosions, flooding, storms at sea, raging rivers and the like. But the market for fluid sim software lies almost entirely in software that can seem to create those kind of effects. This is why Houdini dominates at the moment. So, there's just not a lot of interest in software that replicates mid and small scale fluid phenomena, and probably not a ton of R&D being poured into anything else.
Anyhow, I don't see many people raising the kind of complaints you do, and I was just wondering what kind of background led you to this, in this particular forum of all places.
What you consider 'relatively small' I consider 'of extreme importance' since for my level of attention, where almost everything is close up, all those big scene bulk project [which, by the way come across horribly cheesy in the 'fire and explosion' department] operations are pretty much null and void so far as how I produce.
For some wide area footage, its easier to drive a ship through raging seas than it is to try and recreate it.
If all of the people ON said ship are bone dry while being dunked, that makes the whole shot pointless.
When a Cg character goes into and comes out of a shower, or swimming, or getting a liquid of any form [be it blood, soda, water, oil, etc.] and it just runs right off of them, or just as bad, sticks to them like a glued on clump, that is just gross and disgusting, from a qualitative standpoint.
If I can not have water behave as water when it strikes an object, I will either find a way to film a live close-up, or barring that, cut the scene out entirely, even if somehow it is thought to be critical to the plot.
Thankfully for me, no such shots exist in my projects and if I need to I can temporarily over-texture a model for a scene where it is getting dirty [I have one by the way, 2 main characters get into a brawl in a swamp of all places right when they are 'discovered' by local denizens of the planet they have landed on, a somewhat pivotal moment in the story] and then I can simply repeat an animated scene with and without the dirt mask so that all I have to do is edit out the dirt mask on one side of a cloth when I overlay the 2 in the editing room.
As far as wet rocks, that, again, if need be, I can mask it with a layer of pseudo glass texture where the water is dense enough to have a reflective surface, and carefully back stain it in any of the places where the stone is merely damp.
Still, if you know you lack the procedural power within the application extension to make the shot work, think smart and do something to make it work, don't just leave it there looking crappy as hell because you lack the care, initiative or imagination necessary to be a good director.
If creative directors had produced a lot of these video works I have seen in the past, I might not know how rotten the product is, so in a way, its good that they lack that much sense, but still, somebody should apply those known parameters into the product.
In order to post an answer, you need to sign in.