PDA

View Full Version : allan mckay particle tutorial help


mistral
03-13-2007, 04:32 PM
Hi....I watch the allan mckay tutorial about fracture ( http://www.allanmckay.com/index.htm ) but i've a problem...I tried to reproduce this excercises with a simple horizontal plane without problems but....after this i've tried to destroy a statue with this method:

the problem is that when the particles (displayed as GEOMETRY) falls down to deflector, they bounce on this deflector not for the geometry but for the particle "volume" extremly smaller than pieces of geometry obviously creating some interpretations between visualized geometry and the deflector....

how I can do?

please help me.......

thanks a lot

superhypersam
03-13-2007, 06:06 PM
Its a current limit of Pflow, only the center point respects collisions.

There is a solution using box 3 and offsets, but its not ideal.

U can also bake out your parts uisng a script, then shoot it threw reactor, but its a bit of a process.

Thinking Particles is great for this sort of thing, but is complex.

heh lots of work eithier way.

cheers

mistral
03-14-2007, 06:06 PM
oh no....ok.....so you think that thinking particle could help me?

is it so hard to understand and use?

thanks a lot

:sad::sad::sad:

BrandonD
03-14-2007, 06:58 PM
This is problem long common to all particle systems in packages. The problem stems from the simple fact that particles aren't really objects, just infinitely small points in 3d space. It's relatively easy to apply a collision to a point in space with velocity. The thing is that when we work with particles, we're often instancing geometry to their locations - so in other words, on a frame by frame basis the 3d app is snapping objects to the location of particles, hence the transform matrix of the particles drive the geometry. In this way the particle takes the place of the geometry's pivot point. Incidentally, this is why in Particle Flow you can't spawn from instanced geometry - just particle locations.

For the instanced geometry applied to particles to affect the particle motion you need some kind of integrated rigid body dynamics solver. This is something Thinking Particles added, if I remember correctly. It's a potentially difficult problem to solve and, at least in the past, we've done all kinds of tricks to avoid it.

superhypersam
03-14-2007, 07:57 PM
what Brandon sez!


Tp has a great built in solution, its the shape collision solver.

does a great job on rigid body stuff, and has the added benefit of of it bin connected directly with the particle system.

IMHO the best soulution for rigid body stuff in max


The downfall is Tp pretty complex, but the task is a complex one.

they do provide a tutorial that covers the subject, but its not great on explaining settings.

cheers

PiXeL_MoNKeY
03-14-2007, 08:18 PM
This is one of the features that should be a part of PFToolbox #2, however there is no stated release date for it. I am sure you could hack something up in Toolbox #3 or I am sure Oleg could throw something together to get a basic working version. You have to remember that PFlow doesn't even have inter particle collision as part of the shipping package.

-Eric

superhypersam
03-15-2007, 12:58 AM
yes there is a box 3 solution that invloves calculating the parts offset distance, but not very elegant, and a lil hard to control.

Box 2 should be great, but like u said, not there yet.

CapitanRed
03-15-2007, 10:49 AM
Mistral, if you want to spend the money for Thinking Particles, then do it!! it's great, and not very difficlt to understand. it's just different to Pflow.
there is also a TP-Forum:
http://www.3dplanet.de/cgfluids/

mistral
03-16-2007, 11:14 AM
ok:

@brandonD: Yes I know that the particles are extremly small points in 3d space but snapping 3d fragments on particle in this case it seems too long and difficult cos in my scene there are about 1600 fragments....so....you know....it's so complicated....

@superhypersam: I tried to use rigid bodies and fracture (included in reactor features) but the fragments are too near and so during the simulation it crashes also modifing the collision tollerance...


@PiXeL_MoNKeY: If I well remember I've seen your discussion on oleg forum about a script that create line from particles movement...is it right?
do you think that with Toolbox #3 could help me?I've just Toolbox #2....


@CapitanRed: I tried Thinking Particles just a day ago but the results are quite similar to pArray results...also pArray could fracture object but the fragments are not so realistic...similar to thinking particle results (I've downloaded a fragment tutorial from cebas website)

CapitanRed
03-16-2007, 01:00 PM
well...nobody said one day will be enough to understand and find out the possibilities of TP.
keep on rying ;)
but it gives you enough controle over fragmenting and shape based collisions.

watch the bullettimevideo on "my site"

it is done with fragment operator, which is controled with a alpha map and an omni.
the bullet is pushing the fragments away(collision shape operator)

this is just an example. there is much more, which i havent explored yet ;)

mistral
03-16-2007, 10:38 PM
ok....thanks a lot....i'll see the video immediatly...

stay tuned:)

mistral
03-16-2007, 10:44 PM
i've seen the bullet video and i noted that the pieces are too regular and quite artificial...while with a script (found in internet) i could fragment a mesh in a clearly realistic way...the fragments are irregular and seems real...but otherwise I've the same problem with particle volume...:(

thanks a lot at all

supremepizza
03-17-2007, 08:05 PM
.......oops! :)

mistral
03-18-2007, 07:18 PM
ooops?????????

Digital Masta
03-25-2007, 09:35 PM
Instead of making yet another thread on it, I figured I post it in here...

I'm having some issues with another tutorial, its the one where the ball dissolves to become the teapot.

My issue is when I turn the Find Target's point to "by script vector" it doesn't do anything. Clearly it has to be the script but since I don't really have much script experience at all and the script isn't error'ing out on me when I close the script I'm not sure what to look for. It's also hard to see the script from the video tutorial and there is no max file to look at either.

Would anyone happen to know whats wrong with my script OR know what the script is supposed to look like so I could compare it with mine

Here is an image of a piece of the script

http://i35.photobucket.com/albums/d168/digitalmasta/script.jpg

PsychoSilence
03-26-2007, 12:36 AM
i think i remember that the first tutorial on his second advanced visual effetcs dvd deals with the same kinda effect. i will not post the final file from the dvd since itīs copyright protected content but maybe iīll strip it down when i find the time :)

kind regards

anselm

CapitanRed
03-26-2007, 07:31 AM
(
pCont.particleIndex = i
pCont.particleVector = pCont.particlePosition
) -- <------ your code must be inside the brackets!!

Digital Masta
03-26-2007, 03:09 PM
Wow...I feel stupid. :banghead:


Thanks! :thumbsup:

CGTalk Moderation
03-26-2007, 03:09 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.