Particle Flow Discussion

#801

I have the techsmith codec installed and the first vid is working…I don’t know why the second part isn’t…but thanks for all your help Allan

#802

I used to use it in the last placed I worked for but it had a harder workflow which I can`t seem to remember now nor do I have a license here.
it did have some nasty fragmentation and other cool oparators but it seems like cebas stoped supporting us maxers and develop it solo for cinema 4d.

I am trying a new aproch where I load my fragments and instead of having the skin animate the ir mesh I use shape mark to move them with the original skined mesh…

I hope it will work…

:slight_smile:

#803

Cinema4D guys threw Cebas a lot of money and said “hey can you make TP for us?!”

I’ll just say Cebas is very much still sticking with max and working on a lot of very impressive goodies which’ll be anounced at siggraph.
I’m not one for software wars, although have kept pretty clear of cebas for my own personal reasons, although starting to be whored over a little bit as they are doing some pretty mind blowing things. Scanline are a great example of this, as they’ve built their own fluids engine within TP for use on a few of their projects.

I’ll say what you want to do particle studio and TP can do. although yeah $1000 for a TP license is kinda step. maybe you could cheat it? have an animated gradient ramp mask creeping up the hand which also emits particles or plugs into a vol select to emit from the selection (right now it’s faster) and have the pieces kind of emit instanced geometry aliggned to the surface from that. that could get you some interesting results?

#804

I guess you should call it privet R&D :slight_smile:

I got some progress tho… I got my particles moving with the deforming emmiter and now I am trying to read each particles rotation in frame 0 and substruct the difference made by the shape mark at frame 1 then use this value on top of the shapemark rotation for the rest of the frames.

I am doing this as a proof of concept nothing more nothing less and I realy need to get the pflow scripting expirience.

I will post more when I`ll manage to get the rotation fixed.

looking forward for cebas new stuff and I hope to see a next version of TP since version 1 seemed a bit unstable at the time (year and a half)

if you have an idea about the rotation issue please help…
thanks :slight_smile:

#805

rotation issue if you were to make them prefractured objects linked to the particles that have to orientate based on a deformable mesh will be a pain.
If you’re lucky you can set up some faces that you just want to orientate to the normal. otherwise you’re going to have to set up a bunch of vertices in which it needs to normalize the orientation to which not to knock on pflow but to do with scripting all of this stuff especially for more than a thousand peoples would just get too slow I think.

ideally what you need is just a propper fracture system which is coming, although it’s not here right now so that’s why I’m recommending other stuff. ideally trying to hack a fracture system for deformable surfaces isn’t really too practical. my example is a huge workaround but not the ideal method :slight_smile:

#806

I am not trying to use skin on the fragments anymore.

instead I use shape mark to get my new particle fragments to stick to the surface.
the only problem is, at frame 0 we get the original fractures transform and since I centered all pivots their orientation is the same (z up).
what I did was put a shape mark using my deformin mesh (not fractures) as the object and my particles turned to rectangles which stick to the objects surface.
then I put a script op that give it back its shape.
now I am trying to calculate the base rotation then the new rotation derived by the shape mark op and apply this delta on each particle for each frame as a substraction of the shape mark state.

another way will be to shoot rays out of my objects and align the normals of the fractures to those rays but it seems complicated.

any ideas?

#807

no… what you said sounds right. I’d rather try and read the normal of a select face ID if at all possible opposed to a ‘z-ray’ as they’re goddamn slow.
we did something similiar a few years back for a pitch for a census TVC in Sydney at Ambience where we were doing a pin cushion type effect, and the idea was to animate a bunch of primitives, and then run a simulation which would generate a pinboard of X density and it would fire Z rays down to work out where to place the pins etc. on every frame.

Dan Kripac wrote the script which kicked butt although we litterally spent 3 days to do that sim :slight_smile: and it wasn’t that many pins at all. Ideally using a zdepth ‘map’ would have been more ideal I think, and read in normalized or clamped luminence values or render the zdepth as a float.

erm so yes that’s one way to do it, although everything discussed so far sounds like it’s gonna require a lotta simming time for orientation etc :wink:

but sounds like a good idea
personally I’d go the simpler route, although you’ll never learn anything going the easy way either so it might be worth it :wink:

#808

thanks for the insight :slight_smile:

I took the privilige to send you the file I am working on by email since it was 30k and we can upload 20k here and I can`t seem to remember the uploader… :frowning:

anyway my only task now is to read the particles orientation before shape mark then have shape mark do its magic
then read the difference at frame 0
and have this delta as a substraction of the shape mark rotation since my pivots are all oriented to world cords’

thanks mate and if I make it to siggraph you got a beer from me too!!
:beer:

#809

say could I clone my fracture objects and give them a new name like fracture_pivots*
then I could some how use pflow to change those new objects pivot like
pcont.particleTM = fracture_pivots[i].pivot.transform?
or orientation?
would something like this work?

if it is possible then I can then use the new objects with the corrected pivots as my fractures and they will be aligned perfectly since their pivot been edited.

tell me what you think

#810

ack kinda under the gun here so brains hurting too much to rerun that through my head but it sounds about right. I had originally created dummy’s representing the original positions of my objects so that way if they were external objects they know their original positions at frame 0 etc.

if you were to assign the pivot again you’d do something like

for i in 1 to fracture_pivots.count do
(
pcont.particleindex = i
fracture_pivots[i].pivot = pcont.particleposition
)

if you wanted the pivot to follow the particle? or do you want it vice versa? eitherway yes that can work
if it’s the other way around, keep in mind with scripting I think you need both particlespin and particleorientation both turned on in the channels opposed to just one or the other…

anyway hope that helps, will look at your scene tonight if there’s time, otherwise sometime this week.

#811

last Q when you have the time…

I only want to change the pivots orientation so what should be the syntax for such a task?

anyway thanks again and if there is anything I can do to get that gun away just tell me!!

:slight_smile:

#812

sure, blow up newline :wink:
if you mean an objects pivot it’s $.pivot
if you mean a particles pivot… I don’t have max in front of me right now … but you need to set up rotation by whatever place in world space.

#813

I got some realy nice things going on but still I wasn`t able to redo the rotation of the shapemark and have it only transform my particles instead of completly re align them.

about the pivots, I wasn`t able to simply use .pivot = *.rotation so I ended up moving the particles data to point helpers and then using pflow took a clone of the fracture objects array and gave them the transformation of the particles.
I then took all my new fractures and choosed affect pivots only and align to world.
so finaly I had the fractures moving but it was the objects them selfves and when I tried to feed them into a second flow they went crazy…
ohh…
maybe some sleep will do.

thanks anyway mate.:cry:

#814

you want to change the pivot of rotation right? so a particle or object could be 100 units away from 0 space and rotate around the center of world space oposed to it’s center? or do you mean using the pivot for alignment? as pivots are positional information.

yeah that kind of stuff is a pain in the butt, and personally any kind of scripting for rotation stuff I try not to go anywhere near, some of the guys here understand quats but it’s something I usually avoid wherever I can.

Like I said before though I think it might be going a lil overboard with the way to do it so you’re certain to run into some problems :slight_smile: but it’s all good

#815

I find it quite hard to explain so I`ll restart…

1.I got an object took it apart to fragments using parray.
2. centered all pivots (their alignment is the same)
3. I added bend and twist for the object and animated it for 100 frames.
4. I used pflow to generate the particles out of the fragments objects aka bobo`s simple chunks.
5. I wanted to get the original object deformation so I added a shape mark op. (which takes position and rotation into concern so I got them all aligned like spikes deforming above my object.
6. what I am trying to figure is, I got my fragments at the orientation I want and they get their rotation changed by the shape mark so I want to find a way to substruct this change while keeping the shapemarks transformation only retouching it so that they all align to their origianl fragments mesh and transform acording to the object but from the original state and not the first state that the shape mark give us.

its hard to explain I`ll upload something to the uploader tommorow.

10x

#816

Ok, i’m postin this out of desperation … and … first of all, i am sorry but just couldn’t read all the 55 pages here … i will when i have time … so … if this already was brought up, sorry again.

My question is: is there a way to make particles build on top of eachother?

Imagine a slightly downwards hill.
Now imagine a mud flood coming down and hitting a building.

Apart from some minor problems, already made it slide down the hill, and stop when hitting the building wall, while the other mud that doesn’t hit the building continues to flow nicely done.

So how to make the particles mound up behing the building? Any good tip?

Thanx so much in advance for any pointers :slight_smile:

p.s. : I’m makin the mud with a blobmesh, as u may already have guessed :slight_smile:

EDIT: Bahh! As soon as i hit post, i instantly thought of a way of solvin this… tell me what u think please.
Couldn’t i make a test for the particles speed and, if its zero or very near, make:

a) particles rise (with some warp like gravity or something), so that the next particles continue to come down the hill and hit the wall and so on and so on

b) scale, so they become bigger givvin the illusion of mud building up

#817

getting particles to interact and collide with eachother and other objects isn’t really something dumb particles do by default :wink: You can use a speed by surface or speed by icon to get the particles to move how you want, although they won’t layer on top of eachother etc.
You can hack it to get some pretty good results, although I’d recommend checking out glu3d as it’s a fluids simulator and is build specifically for what you’re requesting…

hope that helps!
It’s free at this stage if I remember correctly.

#818

U N B E L I V A B L E!

Damn! This Glu3d thing is awesome!

amckay, i already saw refernce to it on the starting posts of this thread, but i someehow just didn’t download it …

Thanx for the tip! It seems it will do exactly what i need, but now i have to go home … tomorrow i’ll set it up!

#819

Hi Allan what script did you use for the Birthscript in your Fracture Tutorial???
I want to set up a bullet that hits a window and shatters it and this could be very handy…or is there a better way to do this…with parray??
but I want to do this with only pflow…is this possible???

#820

glu3d is a great product - wish it was a little better known as the guys have worked hard to push it out there and are doing some very interesting things. Hoping to help out with it a little bit more down the line as it definitely doesn’t get the acknowledgement and fan base that it deserves.

hamu73, the birthscript was a small part of a presentation on pflow I’ll be doing in august. Pflow mind you can’t fracture by itself right now, parray’s used just as one way to break the pieces apart and be fed into pflow.

I wrote a system for breaking glass, it’s 100% code. it’s not finished yet, but on it’s way, just haven’t hme :\ jason crosby here at DD did a kickass system with thinking particles for that alias series, where erm whatsface jumps off the building and shoots in the glass, very very cool system.