#Waterfall
34 messages in this thread
Your feedback on WATRFALL.ZIP is just what I need.
Telling me specifically where I can do better is
very useful. I'm glad that once you
<figured out … the dialog box, everything went
really quickly.>
<the little box that comes up says "this may take
awhile… " so of course I wait and I wait, … >
Oops! A button labeled "Continue" on that box is
almost invisible (same color as the box).
<it says use the Camera Script_gen… >
<The script talks about seeing spheres …>
I should try getting more sleep before deadlines <g>!
Hmm, I wonder if the script itself can assign
one view to the camera.
<… its really neat! Those little boxes have a life of
their own. They don't look anything like what you would
expect keyframed motion to look like because its so
fluid. I find this most interesting, though I'm
probably not explaining very well. Yes I've created
very smooth animation motion, but these boxes look like
they're alive and I'm not sure why… maybe is the
elasticity or "exponent" you put in, whatever its quite
amazing. >
You got it. The lifelike fluid motion results from
elastic "surfaces" (like that of a tennis ball). As
you guessed, the elasticity is controlled by the
exponent. I find it remarkable that computing the
keys is so quick (10 frames/minute, for 200
particles interacting with 5 masses, on a 486-66).
<I point all this up just because I think future
examples might benefit … what seems obvious isn't
always…>
I'll use the information you've given me on the
next one. Thanks!
Ed K
>> Hmm, I wonder if the script itself can assign one view to the camera. <<
I think just a comment to go to the camera view and render is all that's
needed. Terminology is what led me astray. I see the word Script and I'm off
looking for one. <g>
>> I find it remarkable that computing keys is so quick… <<
Yeah well I decided to push it and put 2000 to generate 2000 boxes, and its
pretty dang fast, course I'm using a P90 but still. Too many particles however,
looks neat though.
>> I'll use the information you've given me on the next one… <<
Please do, its easy to feel intimidated even when knows quite a bit, and as
this whole thing is really not terribly complicated, it would help people get
alot more excited about it.
Really neat stuff Ed, I can see hundreds of permutations on this. Like tiers
and tiers of cascading spheres.
Thanks much, I can see alot of promise in this area. Keep up the good works.
🙂
John
<Yeah well I decided to push it and put 2000 to generate 2000 boxes, and its
pretty dang fast, course I'm using a P90 but still. Too many particles however,
looks neat though.>
How fast was it?
Ed K
>> How fast was it… << Ed,
I'll have to do it again to give you a precise time, maybe later this morning.
I've got some other stuff rendering. One thought is to use Tetra's instead of
Boxes. I don't see any real advantage to boxes and this would make it possible
to double the number easily. I was looking into how you can assign a particular
object and duplicating it rather than building them from scratch. The other
advantage this would have is that mapping coordinates and materials could be
already applied. If you follow up on this I'd like to see the code. I know what
I'd like to do but the "how to" is just developing.
John
Ed,
>> How fast was it… Part II <g>
Well it was so fast… just kidding.
It took about an hour to process the boxes for 70 frames and another 38 mins or
so to render them, so figure about 2 mins a frame all told. Very fast IMO.
That's below or par for a typical animation.
Something I should have mentioned earlier is that I've been working with
KeyFactory which is a collision detection system. All is relative in terms of
time. Several times with a couple of thousand objects its taken 4 to 6 hours
to build all the keys. That's my base reference for whats going on here.
I mention it additionally because, I was pretty frustrated by the situation,
which had as much to do with my inability to figure out what was going on, as
my expectations regarding what to expect.
I would hope that you'll take my comments in that light. When I said it was
"nuts" it was on two levels, one the frustration of what I went through, and
two the end result… which "nuts" doesn't do justice… "pretty wild" is
better, "incredible" might begin to describe it… and I think I'll just leave
it at that. Besides I think Gary said it all… 🙂
John
Ed,
I am interested in scripting particles to make some special shaped
particle flows. I have tried the stock spurt, vapor, turbo, dust, etc. and
they are all too chaotic to fit into a flowing shape. Can the KUB scripting
extensions help create something like a laminar liquid flow of particles along
a path?
Thanks,
Sanford
Can the KUB scripting extensions help create a laminar liquid flow of particles
along a path… <<
I would think. I know Ed's mostly been working with gravity. But my thought was
to combine MetaReyes and a particle system to create what you're after. It has
some incredible possibilities, if its doable. Just haven't gotten around to
exploring it.
John
I don't have Metareys yet, but I have played with it at work.
I have only seen the first version. I need to combine particles and
metaballs. what would be really great is to have impact detection,
which the Scripting language is supposed to be able to do…but I
havn't found an example other than "bounce." Then the liquid could
impact and flow around an object…
Sanford,
>> Then the liquid could impact and flow around the object. <<
Actually that would work just like Ed's waterfall example, which shows
how that effect can be obtained. He has an exponent which causes the
particles to flow around the object as opposed to hitting it and
bouncing off.
TSE provides the means to do the kind of CD you're talking about as
well and there is an example with pool balls on a pool table
included, which was also created by Ed.
Get the feeling Ed's got a monopoly on this stuff. <g> No but he's
certainly far more knowledgeable than most and as far as writing the
program, he could do it alot faster and more efficiently. He might be
willing to do it on a contractual basis, provided he has the time. I
think you might want to consider this seriously, as it will
undoubtebly be to your advantage. And no I'm not plugging his
efforts, I'm just letting you know that its not as simple as it might
appear, so if you're on any kind of a deadline it would definitely be
to your advantage. On the other hand if you've got the time, it would
be a great way to explore TSE, as you can literally write a whole
program to create the entire animation from objects to keyframing
etc.
Hope this helps, and if you've got further questions Martin D. is the
one to talk with. I just thought the idea of using MetaReyes in this
manner would be a good approach, which Martin D. has confirmed is
do-able, much to my delight. 🙂
John,
Thanks for the idea. I am reading and downloading things this week to
become more familiar with Ed's projects.
Sanford
Hi Sanford,
<< what would be really great is to have impact detection, which the Scripting
language is supposed to be able to do…but I havn't found an example other
than "bounce."…>>
Have you used Larry's recent contribution to the Script section of the lib yet?
File Info for POOL1.ZIP
Section: Scripts Contributor: 74010,3076 Size: 15571
Submitted: 3/26/95 Type: Binary Accesses: 135
Title: Ball/ball and ball/cube collision detection
Keys: COLLISION DYNAMICS BALLS CUBES GRAVITY IMPACT
A replacement for BALLS.K3D. Handles object motion after impact correctly,
includes sphere/cube collision detection and processing, balls can both bounce
and roll. Includes small 3ds file as an example.
jonas[adesk]
Jonas,
I will look at the pool ball script. It would be a good place to
start. Thanks.
Sanford
If you want to mutter incoherently at 3am (when you are suppose to get
up and go to work at 8am) about collision detection, why you can have
a collision at time t and then at t-1, and why objects created in
different viewports give different results when the moon is full,
drop me a line. I may not have the answer, but I will sympathize
with you <g>.
Larry,
I already mutter incoherently at 3am, and I fear that scripting is
going to add a new dimension to that dubious activity, but how else do you get
from here to "there?"
I would appreciate any help in understanding how to set up scripts for
purposes not covered by examples in the manual, such as flowing sand. I have
just recently tackled scripting and am not quite ready to generate scripts of
my own from scratch. Would you recommend the sections on scripting in the new
edition of Inside 3D Studio? Or is there a better resource of in-depth
understanding? I confess that the Version 4 manual leaves me with a great
number of unanswered questions.
I have been studying pool1….it looks cool, but I havn't studied it
long enough to understand what's going on. In about a week I hope to be able
to ask an intelligent question or two. Thanks
Sanford
>> not covered by examples in the manual, such as flowing sand.
>> …is there a better resource of in-depth understanding?
May I suggest a good book on graphics principles such as "Computer Graphics
Principles and Practice" by Foley and VanDam. It will give you helpful
background on the fundamentals. Follow that up with a good vector mathematics
book and you will be all set.
– Dave
David,
Vector mathematics. Well, that reminds me of my electrical engineering
days. I will check out the books. I am waiding through the exercises in the
3DS book. I already know basic a little, so I can write scripts to move
objects. But generating particles? I assume that one of the IPAS programs is
better than others. Do you have a favorite? I find Vapor hard to control,
even though it has lots of settings. Any ideas about how a script would call
and execute AXP's? (like explode on impact)
Sanford
Hi Sanford,
<< Any ideas about how a script would call and execute AXP's? (like explode
on impact) >>
Just like the keyscript (*.k3d) KXP that ships with R4, The Script Extension
has a AXP equivalent. It is assigned the same way as current AXPs.
Well it's kinda like a mutant BASIC. The commands in general are pretty
simple, but the occasional weirdness isn't documented anywhere.
So the language itself is simple to learn, the question is what are you going
to do with it? That's where it gets tricky. Every application naturally has
vastly different computational requirements, and very few overlaps between
applications. In other words, it going to take you longer to figure out the
math/logic of how to do something as opposed to doing the actual programming.
I haunted various libraries for about 3 weeks when working on Pool1 to get a
handle on the best way to handle the motion after impact. I found some helpful
hints on how to make estimates using sliderule <g>, but little that was
directly applicable. The I finally found a fairly recent mechanics textbook in
a cowrkers office that pointed me in the right direction. The code I wrote
still has at least one major problem if you want true reality based collisions.
I'll let you figure out what it is, and why I couldn't use a single dummy
object moving around the "bounding" box <g>.
I think part of the secret of flowing sand is not to worry about particles
moving through one another. Collision detection is a massive CPU hog. Just
think of the sand as water, and parametrically (sp?) model the terrain in the
script.
I've looked at the new Inside 3D Studio at the section on scripting. While
interesting, I think that just sitting down and starting to do something will
bring you up to speed the quickest.
LAM PS: Get all 4 volumes of Graphic Gems. They may not answer all questions,
but they at least can give you ideas on how you might be able to handle various
problems.
Larry,
I ran the "pool" that comes with 3DS. All the colition reactions seem
to lag quite far behind the impact events…I will run pool1 tomorrow. Thanks
for the suggestion of the books "Graphic Gems" but I confess I have not heard
of them. Where are they available? Someplace I can call? I am in Los
Angeles.
I've seen them at Microcenter and Border's book stores. Should be readily
available.
LAM
Hi Sanford,
<< Thanks for the suggestion of the books "Graphic Gems" but I confess I have
not heard of them. Where are they available? >>
The books are very good at hand holding you through many fundamentals. Each
section(chapter) addresses another 2D/3D graphic problem, explains the solution
and then provides source code in c and later in c++. The source code can also
be found in….
File Info for GGEMSD.ZIP
Section: Software Dev. Kits Contributor: 70673,1605 Size: 370422
Submitted: 10/7/94 Type: Binary Accesses: 44
Title: Graphics Gems I/II/III Code (DOS)
Keys: GEMS GRAPHICS SOURCE GLASSNER ARVO KIRK MONOTRICHOUS
"The code for the articles in the Graphic Gems series (I, II, and III). This is
a botched down version for DOS (filenames, CR+LF). I haven't had the time to
deal with IV. Check GGEMSU.ZIP for it in its original form. Get the books!
.Graphics Gems – Anrew Glassner – Academic Press – ISBN 0-12-286166-3 .Graphics
Gems II – James Arvo – Academic Press – ISBN 0-12-064480-0 .Graphics Gems III
– David Kirk – Academic Press – ISBN 0-12-409670-0
Gus"
PMJI, Sanford;
If you find something like this please let me know. I'm trying to work out
something special that requires I hold together a great many particles of very
definate color values. This is something on the order of a pointilist type of
image that has a great number of particles making up the total picture. Only
animated of course. If you've ever seen anything by Beardsley I think you know
what I'm going for.
Murph'
James,
You probably know that Ed, uploaded a flic of his likeness which explodes. You
could talk to him about that. What I've experimented with with good success is
Disintergrate on a tightly spaced grid. Then its just a matter of sizing and
placement of the nodes. Then you can texture map just like a regular plane. I
also used X-mas and applied Explode to it as well.
Both work, just depends on what you're looking for. As I recall the only
problem was a 64k face limit which I got around useing two planes but mapping
them simultaneously.
Hope this helps,
John
Thanks John!
Great, I'll try this.
Murph'
John –
Your answer to Murphy at 4:02 this AM came through blank. <G> Was that a
statement or just up too late!
Anne
Anne,
>> Your answer to Murphy… came through blank… <<
He thanked me for some assistance, and I had replied "Your Welcome". But
apparently I lost it when saving it out…
So… YOUR WELCOME MURPHY! 🙂
John
Any time John! <G>
I figured that's what happened.
Hi Anne!
Murph'
Sanford (and Ed):
>> Can the KUB scripting extensions help create something like a laminar
liquid flow of particles along a path? <<
Absolutely, but I wouldn't go looking for any generalized liquid dynamics
simulations.
On the other hand, if your project has some "useful limitations", you may be
able get by with some artful cheating. A particle system following a path is
essentially a flocking system. Ed's approach to particle systems (based on
attractive and repellant forces, rather than collision detection) should be
perfectly suited for flocking: the forces could be used to control the
particles and move them along the path.
If you solve the basic motion problem, the biggest challenge will be getting
whatever liquid appearance you seek. I don't have any special suggestions, but
I'd also point out that both The Script Extension and, more specifically, Ed's
particle techniques (based on simple volumes) are perfectly suited for use with
Metareyes Metaballs.
Martin
Martin,
I will look into what has been previously discussed in this
forum about Ed's particle system setup. I have only recently gotten
into the scripting area. I'm trying to recreate naturalistic liquid
and particle flows. Also sand…I'm working on a scene from my
version of Rommel' North Aftica campaign…command cars racing across
the desert, sand storms, tanks crashing through trucks and dunes,
Howitzer's firing and bombs blasting. Particles, particles….Any
help with impact detection and particle behavior would be
appreciated. Is there anything written specifically about applying
impact detection to particles?
Sanford
Sanford:
>> I'm working on a scene from my version of Rommel' North Aftica
campaign…command cars racing across the desert, sand storms, tanks crashing
through trucks and dunes, Howitzer's firing and bombs blasting. <<
Yowza! All computer generated?
>> I'm trying to recreate naturalistic liquid and particle flows. Also
sand… <<
I don't want to sound discouraging, but you'll find no EASY solution.
On one extreme, you can get your hands on a supercomputer and run a
scientifically "accurate" liquid dynamics simulation. On the other
extreme, you can paint up some clever animated texture maps and cheat
a bunch. It's possible that LenzFX Lumina (the new iteration) would
help you out with the clouds and airborne particles–I don't have the
software yet, so I don't know how far Aaron's gotten into the
Alias-like density rendering stuff.
As far as impact detection, you have applications like Keyscript,
which contain some rudimentary collision-detection functions, and
"canned" applications like Imagine Key Factory, which get more
sophisticated. Note that neither of these options are really oriented
toward simulating volumes of particles; they're more oriented toward
giving objects particle-based behavior. You'll find it a trivial
matter to bring your computer to its knees doing collision detection
of any sort.
I would look first for solutions based on image processing and texture
mapped effects, possibly combined with some of the options above for
larger debris.
Martin,
Thanks for the thoughtful answer. Yes, I am attempting to do my war
scene all on the computer. I have already discovered how much time it takes to
render a dense cloud of Vapor. Even tornados take a long time when you max out
the particle count. My plan is to use video post to assemble different
animations…with Alpha, and some 2D photoshop effects. I will have an old
French fortress in the scene as well. Because everything is stylized, I will
be able to take many liberties with the effects.
I havn't tried LensFX but I've seen what it can do. Very pretty.
Great for cannon fire and bombs, etc. Smoke will most likely be 2D…except in
closeups. Lots of experimentation ahead.
Sanford
Is Watrfall.zip part of TSE, Ed? I couldn't find it anywhere on the board, so
I was curious.
At the moment, I'm leaning heavily towards getting TSE, solely based on the
work you've done. Having lurked I did note that you were promising a particle
system script that (I assume) was the basis for many of the things you've
created. I'm particularly interested in the water flic you posted that you
said also used AniPro. I'm interested in creating similar water — just how
important was AniPro to the creation of that effect?
I hope KUB is at least paying you royalties or something — you've certainly
deserve them!
Mike,
PMJI but yes Watrfall.zip is part of TSE and requires it to run.
>> I'm leaning heavily towards getting TSE… <<
Its a good move if you got any programming experience, and the
possibilities with MetaReyes and Particle systems are really
extrodinary. This has the potential to be a whole new area for
creation. I'm pounding away on some other projects but as soon as I
get a chance I plan on trying Waterfall modified with MetaReyes.
Should be cool…