|
|
Basic flocking behavior question I am trying to get a handle on the Particle Tools and am having quite a rough time of it. Does anyone have a very simple QTZ or tutorial that simply shows how to get some number of objects onto the screen and flocking about? There seems to be some hat trick I am missing. I can't keep the boids from flying off the screen never to be seen again. The demos and sample projects I have seen are quite complex and while I would love to delve into them at some point, am currently interested in a much simpler base to work from. Thanks in advance!
|
The kineme boid fish demo would be useful to you. it won't even open for me -- crashes QC every time. the n-body demo is also good. in the n-body comp, if you replace the emitter gravitate object with a particle object it simplifies the concept of bodies gravitating around a point. (also perhaps add a couple sample & holds after the mouse patch to sample viewer coord with the mouse button, otherwise the gravitational point will keep moving off screen).
(i wish kineme would have a look at upgrading the particle tools suite. it's always held so much potential but also felt more like proof of concept. i'd pay for the plugin if it insured stability and support. always been crashy for me, usually trying to simply get up and running with included demos.)
Thanks for the help. Got a very basic model running. At least it's doing somthing. Now I just have to get it to do something I like to look at.
A few questions ...if you don't mind! :)
-Was Particle Tools crashy for you in 10.5? -What would you want to be different? -Is setting stuff up hard?
The boidfish demo had some problems because of K3D being updated to fix some bugs. Particle Tools hasn't, and probably shouldn't, be updated to fix that, because there's a whole host of other issues remaining that make it very sketchy for widespread release.
Vee has a lot of code (the codebase PT uses) that is totally bound to 32 bit. So, it becomes a matter of either re-writing Vee and updating the parts that are foul, or taking the functions and slowly substituting them with stuff written from scratch, one by one.
I've actually done a crap-load of work with the PT code, know it pretty well, and have tried both approaches. I think that in order to get the really high particle counts we would likely want, Vee should probably be let go of, and another system put together. Particle Tools makes sense to me, as far as how emitters, forces, etc., work with constructing a scene, but I notice it sort of doesn't to most people. So, it seems like maybe that would have to be addressed as well.
I think particle systems with good physics are one of the coolest things when it comes to "whiz bang/that's awesome" feelings from watching cutting edge graphics.
The hardest thing for me, coming at it from scratch is that there is no centralized how-to/manual. Searching through the forums is great but sometimes things slip through. My initial problem was a function of not knowing where the 32/64 bit setting were, so after looking through everything in QC assumed it was a function of a different release.
I come from design background (theatrical lighting) and not a programming/code background. So there are a lot of assumptions made about knowledge base that makes much of this opaque to me. But I learn fast, so once I get it, I've got it and can work with it.
Even basic stuff like "A plugs into B plugs into C." Is not obvious. Some of the tools I can't yet figure out what they should plug into. SO I try the hunt and peck method. I'm happy to do that because I am interested in the software, but would be a lot easier if the toooltip rollover had that included. Some does and that is awesome. But not everything.
As for my initial flocking question, it just took a while to get the settings right so I could even look at sample files and then find the projects that were not so complex and code heavy that I could read what was actually happening.
My solution (attached) is no where near the most elegant, by any means. But I got the basic system working to a point I can begin the see if it will do what I want. :-)
Where is the Vee sourcecode BTW?
What about replacing Vee with another Particle Engine entirely?
Vee, okay.
I don't remember so much about Particle Tools crashing in 10.5 -- took me a while to getting around to exploring the tools. i think maybe less crashy in 10.5, as i did dig around some and do not remember crashing being a huge issue. and K3D is the hitch in the boid fish comp, okay. suddenly i'm embarrassed cuz now i remember reading this.... still, last crash was with prefs set to not open viewer, still crashing. shouldn't i be able to go in and disconnect it?
I'm okay with the basic PT setup at this point, used to it. it's versatile and open enough with all the patches and ports that can be added for objects, children, etc. The "proof of concept" comment comes after coming back to the demo comps time and again and more often than not getting crashes. (means I could never use this for performance, and i would like to.)
Particle Tools was pretty mysterious for a while. i get the PT setup okay at this point. i wish the math / physics notation was documented for practical use (epsilon, theta, etc.) what do some of these params actually do? are they all active? where can i look this stuff up? maybe now i can at least look up Vee.
what has helped me with PT has been the slow and ongoing process of understanding how data flows in QC -- pulled from the right, etc. somehow when that started to click, coming back to PT made a lot more sense. after that, documentation in the different patches and the demo comps have been decent for reference. is this helpful?
um, what are the two methods again? rewriting Vee vs starting from scratch? Sounds like a rewrite gets your vote?