About the physics development sprint

I was reading the blender news from yesterday, and I maybe too tired at the moment, so if I ask some stupid questions, please forgive me, but in the report about the development sprint it says

Target: speed up and synchronize development of physics and simulation related code in Blender. We currently have several modules doing different aspects of physics without much integration or a well described relation.

Physics systems are for example:

  • Particle systems (free flowing particles for fire, gas, etc)
  • Hair or fur (grass, plants, and duplicators as well)
  • Soft Body
  • Cloth simulation
  • Fluid simulation
  • Rigid body (currently only in game engine)
  • Hinges and Constraints
    (Q1) Will the all the various physics systems be integrated into and set physics module, similar to XSI Simulation set where all physics, can be accessed from a singular module?

This is how I understand it to be… anyway

(Q2) By the ‘free flowing particle fire , gas’ statement… will voxols be implimented?:confused:

I do understand this development sprint will be done to speed things up to get peach underway with a good start, can someone explain the implications this will have on the current system?

Like I said, I am tired (Baby = no sleep) and if I read it more carefully maybe I can come to the right conclusion myself, just interested.

P.s. I hope this is the right thread to put this in… sorry if it isn’t:(

Yeah it kind off brought up a lot of questions. Answers to which I don’t know either, hopefully someone will post here soon to enlighten us, and then someone else will post and correct the person who has enlightened us.

Thanks gat, I am glad I am not the only one. Well patience is a virtue, one I need to develop, well good luck Daniel Genrich, Jens Ole Wund, Nils Tuery, Janne Karhu, Campbell Barton, Brecht van Lommel and Ton Roosendaal! :yes:

as i understood that ‘sprint’ will = to the dev’s getting together and just speaking & planning for the future, as for what will be decided/more details about what actually is discussed in there, we will probably get to know more properly after the event…?

Thanks! Felix

Couldn’t they just chat online? We have the technology! lol :stuck_out_tongue:

because things are more clear that way, and theres no cross posting, like there is on the irc, perhaps…? :stuck_out_tongue: and im not saying they wont do any coding at all, probably some tests and stuff will be part of it too, but i doubt they plan to impliment some new features in such a short time… :confused:

In person productivity for programmers can increase dramatically especially for design decissions.

As to what it means - my understanding is that the way forward for unifying the systems will be discussed. In particular unified baking and baking editing tools, and allowing the various systems to pass data to and from each other during simulation so that they can be coupled.

As to will voxel rendering be coming - that isn’t on anyone agenda as far as I know.


thank letterRip for the info. I didn’t think voxols were officially being considered since Ton already mentioned on another thread that that would be a feature for another open project. Just seeing fire and gas made me think otherwise.


Hit the nail.
My intention was to communicate and find out synergy ( +1 bullshit bingo points ) .
Well the whole thing was about what we have right now … how can we improve that … what do we want for the future and what needs to be done to get there.

Discussions went quite deep and the results are satisfying IMHO

well i don’t expect anyone to understand that kind of stuff … but i think this meeting resulted in a reasonable plan to coordinate and continue coding tools based on physics simulations.

Sorry… no new stunning features … just an effort to keep blenders code clean and ready for new stunning features.

That was…


P.S. we used ancient voodoo techniques like talking … listening … watching facial expressions … gestures … using a paper based flip chart, pens and pencils to draw a sketch and other magic that does not work very well online :slight_smile:

bjornmose, That was a very exciting thing to read :slight_smile: . While I don’t understand 100% why each single decision is made, it does look like you guys came up with a good way forward for blender.
I have two questions:

1- did Ton at any point talk about how simulation/baking will work (or not-eek!) for linked objects/groups and things animated with proxies?
2- does “cloth and softbody cant be integrated” mean that the simulations from each cannot affect the other; or does it (I hope) just mean that they cannot be “collapsed” into a single modifier with settings to tune from softbody case to cloth case?

thanks for synergizing our assets in an upwardly horizontal paradigm.

@Slikdigit, i think the answer to your second question is that softbody and cloth will be 2 different modifiers.
not being able to let a softbody affect cloth would just be stupid :stuck_out_tongue:

Great to see what to expect. Thanks for the link bjornmose!

P.S. we used ancient voodoo techniques like talking … listening … watching facial expressions … gestures … using a paper based flip chart, pens and pencils to draw a sketch and other magic that does not work very well online :slight_smile:

I thought that was how you programmers get things done

I saw it was clear right away what the intention was. First, they synchronize thoughts. Sketch a plan for a future unified frame work. In itself that is so exciting as hell. I mean, if a general collision API won’t get your heart pumping, then I don’t what else will. Simulation can be programmed kernel like and general stuff like collision, dynamics and interaction between simulation system can be handled using those API. Very flexible, extensible and robust if you ask me. That OpenPM thingy should score high in the hit charts too. If I understood it correctly, it is a framework to manage multithreading? That means (Like Gensher said) that cloth sim will speed up noticeable and the goes for the fluid sim. This whole thing sounds like Blender 2.60-ish to me.

any news about fluid-enhancements?

eg: rigid-body / soft-body with fluid interaction??


i think its name was OpenMP witch stands for(iirc) Open Multi-threaded Physics(could be wrong about it though) but thats what ive understood… :confused:

edit: nope was wrong, check the link(the name) though basically, indeed seems to be all about multi threading… blargh, ok, time to get some sleep…