Advertisement

You're blocking ads, which pay for BlenderNation. Read about other ways to support us.

Lagoa Multiphysics Demo

61

The following video was made using the Lagao physics engine - not in Blender. The first one to re-create it in Blender gets a cookie - and a spot on our homepage ;-)

About the Author

Avatar image for Bart Veldhuizen
Bart Veldhuizen

I have a LONG history with Blender - I wrote some of the earliest Blender tutorials, worked for Not a Number and helped run the crowdfunding campaign that open sourced Blender (the first one on the internet!). I founded BlenderNation in 2006 and have been editing it every single day since then ;-) I also run the Blender Artists forum and I'm Head of Community at Sketchfab.

61 Comments

  1. Oh, and the demo is very impressive. I think there's someone (cof, cof, farsthary) wanting to get even more famous by implementing something resembling it.

    (Couldn't edit previous comment...)

  2. You should add that this was rendered with a beta version of the Arnold renderer.
    It's coming back from the dead, no longer Sony ImageWorks in-house exclusive, he's putting it on the market :)

  3. the elastic materials that stick to and pull on other objects starting at 1:48 are my favorite parts in this vid. i would love to see something along these lines in blender. maybe it could be an adaptation to the fluid or cloth simulators.

  4. blendmaster1024 on

    @endi2: so true, so true. we can't even come close to faking it with sph ... it's just too much. now if sph had gpu acceleration, then you could do solid particle simulation like that ...

  5. This is really very impressive, but if this wasn't made from within Blender, why does the interface (where it is shown) look so very Blender-ish? Is that just Lagoa's "look"? Blender is the only 3D app I can think of that uses that particular interface design....

    That aside, this is an amazing engine. I like how you can control the granularity of the particles in real time; that would be very useful for previewing the simulation. I wonder if it lets you do that with the cloth system....

  6. Yeah, nice stuff!

    The gauntlet has been thrown down. Who can replicate this in Blender eh?

    I suspect this is one of the toughest challenges to come Blenders way, but I also suspect some genius with nothing else to do may crack it.

  7. More technical question...

    I wonder if they use any hardware acceleration for that (CUDA or OpenCL).
    It seems impossible to me to have this degree of interactivity through the CPUs only. Or they would use a lot of optimization...

    If it's sony imageworks, chances are that it was a Cell thing maybe...

    What is sure is that I would get a boner if something like that became usable on Blender, be it through bullet (any news btw? I heard Erwin was really really busy with a lot of stuffs lately...) or through Farsthary's project...

    It doesn't make it any less impressive :D (I had a boner just thinking about the stuffs one could do with such a thing...)

    And no link to buy/download/try.

    Dagnabbit.

  8. ^No, it doesn't look very Blenderish, it has overlapping windows. Also, Apple Motion uses those kind of sliders too. But the interface was inspired by Blender, so I suppose that's why. :P

    SPH in Blender is cool but it isn't stable enough to create such simulations yet (Or at least I haven't managed to get it to do that without crashing).

  9. I also think that a physical award/trophy should be set up for this?

    Say a Shapeways 3D print?

    A Suzanne trophy award?

    3 books/DVDs from the Blender Shop?

    A subscription to CD cookie?

    This is a worthy challenge that deserves some incentive!

  10. Additions to Blender like nodebased dynamics (ICE) along with Bmesh and a significantly improved render engine (preferrably realtime) could make a huge difference, especially in the VFX area.

    Either way, I believe you won't be able to get that speed of realtime performance in Blender until the core gets updated. Softimage was built with a high performance core. Afaik, Blender has not (yet?).

  11. Hello
    This has been done in XSI using ICE node based simulations.
    here is the website of Thiago who developed it. http://thiagocosta.net/
    Lagoa is like a framework of advanced physics to be used inside ICE trees.
    There is not much infos on how and when and which conditions it will be available, but he seems to have started a company : http://www.lagoatechnologies.com/
    Bart, maybe you could also correct “Lagao” to “Lagoa" in the text and add Thiago's link as well as saying it has been done in ICE.

  12. well... to make this all in blender (and for such smooth interaction like in video) we need unified physics engine, for all types of interactions (crushing, fluids, softbody, rigidbody, cloth, particles and so on), not effects separated into different simulators, beacuse they can't interact on such level.

    so by now we can only reproduce final effect by faking it, personally i think it's impossible to simulate it.

  13. @JaydenB & Occulus, it's Autodesk Softimage using its amazing nodebased programming system 'ICE'

    @Sandking: Amen! and after that focus on bmesh, render engine and nodebased dynamics pls

  14. Riccardo Covino on

    It will be included in next Softimage release.
    XSI has always been a step forward all the others, included Maya and MAX, despite its smalle market share.

  15. I'm afraid one cookie won't do it. It would take a years supply of chocolate chip cookies, or Pepperidge Farms Milano's.

  16. Let's focus on getting rigid bodies into a non-game-engine interface first, yeah? Maybe then unified physics for what we have. And then MAYBE we can THINK about something like this.

  17. This would be awesome if it were implemented in Blender. Doubt it'll be open-source, though :(

    But like Aden said, let's try to get Rigid Body physics working in non-GE first. I'm no programmer (at least for Python), but I don't even understand why it doesn't.

  18. For those interested in the underlying technology:
    http://graphics.cs.kuleuven.be/publications/LenaertsPhD

    Personnaly I don't think there is much interest in trying to add 2 ways interactions between the actual solvers in blender (Fluid, SPH, Smoke, SofBody, Hair and the coming rigid body) because it will result in huge amount of code all specific to some particular situation, resulting in almost unmaintainable code. Also it would lead to much more difficulty to have a stable simulation.

  19. It's still pretty hard to make "simple" things in blender like water, smoke, fire, not to speak make them realistic, make them without too heavy workload and days of time.
    Better work that problems out before going for the long run...

  20. Ubisoft = Money = Higher Quality.
    Blender = Open Source = Good quality.

    Let improve what we have now instead of wanting to aim to high
    and come up with nothing.

  21. If we could hire these guys (or one guy?) to make this in Blender (if it's possible without some foreign licensing) then it will be really cool! It's a bit pitty to see this demo and think about that it's not Blender... Farsthary is really cool but he's busy for now as I got from his blog.

  22. recreate it... in Blender? You seem to demand too much. Blender simulators is faulty, you'll get problems even for small simulations. It would take tremendous troubles and headaches to do, at most, something close to this.

    The prize is not much an incentive imo.

    Still cool stuff. Something Blender users would just have to gawk about; and please don't raise your expectations too high. Blender is awesome but it has limitations.

  23. few years ago a lot of things that now blender manages seemed a dream
    (water, unbiased renderings, softbodies, a good game rendering, hair, modificators...)
    now i sit down and wait
    they did it, so blender can do that, and somebody will find the way :-)

  24. Amazing! Yeah, I'd definitely love to see something like that for Blender. Maybe I'll even build it myself one day.

    @ jedihe (the very top post):

    You misspelled "misspelled." ;3

  25. Just wait till the great Blender developer Lukas Toenne finish his particles node system, you will see what amazings things will come...

  26. Looks rather not usefull.

    The "mud" seems to be a lot of blocks. So it looks like ... well... a lot of little block behaiving like mud.
    Maybe if all blocks where connected ( by hand? ) to blobs it would look like mud, but more likely look like blocky blobs. Or am I to pesimistic here? Did noboby noticed that its just blocks?

    "throwing rabbits" is even more, weird. Its loads and loads of cubes that are sticky... How is that usefull? As mesh deform shape? Then how to connect that to a real form? I'm puzzled.

    I'd go for Drews comment "it needs more lens-flare and robots"

    Maybe blender can add some smart shader that renders particles lit as blobs so stuff like milk would be easy to render.
    But oceans with with caps and caustics in transparent water still seems far away to me, even with Lagoa. ( Note how most examples have non transparent material ).

  27. @Joeri: It really depends on the resolution. The high friction granular particle simulation (the first one the looks like soil) for example is of very high resolution simulation. The other is of much lower resolution so you kind of notice the individual particles. Its well explained in the last bit.

    I think this is a developer challenge as I could see nothing similar to this is Blender. Wonder who will try to join.

  28. @joeri67

    I would like to play with interactive things there. And I see that those games with a "spider net" cube and a torus are very useful! Just imagine how quiclky it will help to create different environment objects or just some good looking trash for filling scenes. "Thoothpaste" thing is really cool! It can be used to create very funny animations. Even if it will be connected with metaballs for visualisation - it will be very good. Not fast maybe... but it's better than nothing ;)
    Just remember what was Blender about two years ago...

    Yeah, Lagoa makes physical interactions - it's not a render, obviously.

  29. ICE is the node based simulation engine of Softimage(it still has the older ones not node based).
    This was made those nodes and i think some created new nodes.
    In Siggraph 2010 Autodesk said the Lagoa(Portuguese for small lake) will come in next Softimage.

  30. @ralmon
    Yes, the coders need more motivating well proven statements like this:
    "Blender simulators is faulty, you’ll get problems even for small simulations. It would take tremendous troubles and headaches to do, at most, something close to this."

    As far as i can see the physical model behind is pretty much the same Farstharry used in his new particle code.

    Well and the video reads "Meshed using emPolygonizer" and this is what is really missing with the new particle code. The El'Beem module (aka fluids) does a grid based meshing. May be a generalized version would apply to particles too. Well but that is the work part in coding compared to the fun part coding particles.

    BTW have you noticed the "stray" particles in 0:44 - 0:46 .. looks like the similar simulation techniques have similar problems.

    Finally i want to repeat the obvious:
    Marketing videos are made to make you want to buy the product, not really to show what you actually can do with it. Oh yes i spent my childhood dreaming over that colorful LEGO prospects.

  31. @blendmaster1024: SPH can be done with GPU. It is actually very suitable for this type of processors. See e.g. http://www.sphysics.org where a few colleagues of mine will soon publish a GPU SPH solver. Unfortunately the code will be closed source as far as I know, although the binaries will be made publicly available.

Leave A Reply

To add a profile picture to your message, register your email address with Gravatar.com. To protect your email address, create an account on BlenderNation and log in when posting a message.

Advertisement

×