are you serious about physics? check out XSI results

pretty neat and dense as well.

Just saw the clip, and read some of the comments. Impressive to say the least. It will take a month of Sundays to render 5 frames from any of those animations on my laptop. Fortunately I have taken a small respite from animations for now, else I would be out begging for a demo of that.

Very impressive stuff, like an very enhanced version of some of the PhysX demos.

The first simulation they showed was impressive as if every little grain of sand was geometry. Though I wonder how it would handle fully procedural demolition of solid objects (the only thing I didn’t really see in the video)

both the animations and the real-time previews blew me away…

I read he did it using something called ICE in Softimage, which is kind of a nodes-based user-level editor for effects. A nodes-based, user-assembled “script” outperforming many C++ code out there, that is mindblowing…

Oh my god!!!

it’s true he did it in ICE which is Softimages Node editor for simulations etc, everything in Softimage is node based. To apply texture you hook up nodes. etc.

although he mentions, (and I suspect uses to a high degree) there’s custom C++ written nodes. Softimage plugins are written in C++ (and python I think…)

he’s using in ALL cases SPH simulations and I don’t get the big fuss about it? it’s been around for ages.

Raul Fernandez Hernandez, is working on some SPH solvent for blender.

although what impress me the most w/ Softimage is their code base must be really well written. the simulation and added C++ custom nodes, runs very fast.

xsi can make serious things, blender cant
this is sad
blender is unusable
we need more features


In my language there is a sentence for that: “A poor pupil always has bad tools” …

And if you are very good in what you do, with the tool(s) you use, you’ll have too much to create to spend your time in negativity. IMO

I like your comments. They’re all like a programming language sharing same library, structures and functions.

include <conio.h>
using namespace std;

int main()
cout<< blender is bad, xsi is good. We need more features.>>


I think it never hurt asking Farsthary or Stephen if they can do some work towards getting something like the video shows…

Seems good.

And about the sand render… looks good… probably yafaray can give some realism like this nowadays… but firsts are firsts…

Yeah. Very nice versatility, results and performance :yes: Being nodes based makes it sound even cooler.

I’ve got to agree with Endi here. XSI can make serious stuff, blender cant. Facts. We’re doomed :no: Seriously this is sad.

:smiley: I’ve been watching the awesome work that’s being done: SPH by Raul and particle nodes by Lukas. Looks like a nice combo :cool:


dude! I’m not saying we NEED more features right now. I’m saying Raul is already on it! and has done a lot of previews of his SPH solver code. even started to optimize.

Ok, maybe time to buy Softimage, and start hanging on and start working on Cars 4 then ;D

the charcoal demo was hardly realytime, the realtime stuf was always in the amount fo 2000 particles. IMO the elbeem lib in blender is a way bigger feature.


dude! I’m not saying we NEED more features right now. I’m saying Raul is already on it! and has done a lot of previews of his SPH solver code. even started to optimize.

Ok, maybe time to buy Softimage, and start hanging on and start working on Cars 4 then ;D

the charcoal demo was hardly realtime, the realtime stuf was always in the amount fo 2000 particles. IMO the elbeem lib in blender is a way bigger feature.

all the neat stuf we still need to bake ;D but maybe in the future we can take use of the GPU to simulate fluids / smoke / particles.

Thats just plain daft, i should hope Softimage can do some serious stuff its £1700 per seat. XSI was a good product when it was Softimage now Autodesk have got there hands on it i wouldn’t touch it with a barge pole.

Ive been a XSI user since 4.2 and the ICE system was brought in V7 the last real release before Autodesk bought it out.

Now if you want crap customer service, crap licencing and get the privilege of buying a £800 a year maintenance contract for nothing more than point releases be my guest.

Yes ICE is cool and can do some pretty funky stuff but the problem is with it is that the vast majority of the stuff you see are custom nodes wich people have produced and don’t release them. There has always been a complaint that there is not enough training for ICE so getting to know anything other than the basics is really difficult.

Most of the tutorial makers such as i3d and digital tutors have switched to max and maya tutorials so that says a lot about the user base of Softimage.

RIP xsi

Wow. Digital Tutors is still releasing training for XSI though obviously not at the same rate they release it for Maya, et al. i3d is much smaller than Digital Tutors so they have to go where the money is and my understanding is that making training for 3d software isn’t a lot of money to begin with considering the limited market and rampant piracy. And Sotimage has had a smaller share of users compared to Maya and 3ds for many years now. None of that equals “RIP XSI”. That’s just ignorant. Softimage blew by all sales expectations that Autodesk had set last year (per and while I sympathize with no love being lost for Autodesk Softimage isn’t like Maya. Maya generally ships with so many issues it would almost be foolish to buy it without maintenance. Softimage, on the other hand, still has most of the developers it had when it was under Avid and most releases are fairly solid so a freelancer would probably be fine without maintenance.

I am sure a lot of people are doing things with ICE that are very proprietary but it is also, in many ways, still in its infancy. It is also sort of heavy programatically speaking and if you are not a programmer then unfortunately you probably have to wait until people start selling solutions. I think with the power of ICE and the user base compared to something like Houdini Softimage’s future is actually fairly bright. If I were Houdini I would be worried.

I don’t think any of this is anything that Blender needs right now but it sure would be nice to have sometime in the future. I’d be surprised if some developer didn’t start to build the framework for it in the not too distant future just because it would be a lot of fun to play with.

If you want a taste of node-based effects in Blender then there is the particles branch being worked on, there was work on math nodes commited today

Now if his branch will be able to play with the physics simulations of Aligorith’s GSoC branch, Farsthary’s SPH, along with the current simulation options then Blender could become a powerful VFx package like XSI is now.:wink:

Blender can do great things. When watching the demo, I thought, this is all very impressive, but how likely am I to use any of this? Blender has much more important things to do like improving the internal renderer (which is happening), than adding feature that only 1% of users would use seriously.

The demo was a bunch of VFX stuff but ICE is way more than just vfx tools. I believe it is now possible, for example, to build character rigs in ICE using no bones, thus cutting down on the number of objects in a scene and speeding things up. ICE is a whole framework for affecting geometry above and beyond just being some sort of particle engine. With a few exceptions it doesn’t necessarily do anything you can’t do in other ways but it potentially allows you to do those things faster, with fewer resources and to iterate on it. If the equivalent of ICE were fully implemented in Blender more than 1% of people would use it seriously, I guarantee.

I LOL whenever anyone bites at endi. It is an ongoing joke! the longer it goes the funnier it gets!

Amen to that. It took me a while to figure it out… :mad::smiley:

But he does have a point there, whether intentional or not, this time.

On another note, how many times does this need to come up in threads, I’ve seen a few here of late…