ROBERTO'S BULLHORN: Space Combat Challenge...

#21

Well the idea I had in mind is to use a pipeline similar to the one of a TV show.

In other words, most of the work would have to be done Upfront.
That means that before the challenge is presented in this forum
( and I post the crazy dateline) the CORE digital assets would be done,
and made available for the participants.

Of course I cannot promise a fully rigged alien character, BUT, a painted model of the creature
could be provided, along with concept art. These assets would be the product of a MODELING challenge (hard core modeling challenge) & a sketching challenge (Daily Sketch forum).

-R

#22

Having the assets built up first like that is a crucial step. It’s worked for the indie collbs on CGS…with the push of the larger number of people on official challenges it should have a good chance to work!

It’s always bugged me a bit to see unimaginable manhours and talent that the main challenges get put into them, but it never gets used for anything bigger down the line…seems a terrible waste to me. :shrug:
Seeing the efforts of the challenges be used for more than a personal display/portfolio would be a great bonus to your planned pipe. :thumbsup:

#23

More spitballing here-
Wasn’t gonna suggest this, but what the heck. Might it be worth contacting some of the people that have already created the amazing work on the challenges past and see if they would donate their work/talent for this endeavour? That would accomplish 2 things> ramp up the speed of aquisition, and also there would be a good idea of the quality of existing assets. :shrug:

BTW: I’ve had success recruiting work and talent for the shirowproject collab with this method . :slight_smile:

it’s a different genre but for the short cutscenes ide I mentioned above, this is the sort of thing I meant > http://www.worldinconflict.com/us/trailer.html
It could be built in 5-10 sec parts, section by section.

#24

I have been wanting to work on a procedural space battle for a while in PFlow with max. Hero ships would be hand animated but the background fighting, laser blasts and explosions would all be PFlow.

Sounds like I should get cracking on some research :slight_smile:

ETA: As above; What about a call for old Space Opera entrants to provide models with credit given? I am sure a lot of people wouldn’t be for it, but some people might be willing to offer some of the models for free/credit? Just a thought.

#25

I like that idea…

Ok how about creatures?
I am mostly concerned about rigs.

-R

#26

The thing about creatures is that there aren’t a lot of low-budget methods for doing crowd simulation. Most particle-based methods have a bunch of poses for each character baked out into static geometry, and switch between them on a frame-by-frame basis by a particle runtime expression. Then you have a couple expressions to keep the particles from hitting each other, and you’re done. You hand animate every guy who gets close to camera.

Massive and Behavior are significantly more complex than that.

I mean, with some really long particle coding, you could probably get something more complex than that, but I don’t think a whooole lot more complex. The method I’ve described wouldn’t have deformational motion blur (i.e. of arms and legs).

Sorry if I’m coming off as the naysayer here. I just like to play devil’s advocate, and I’ve had to do this stuff before, and know how hard it is.

#27

I agree totally. There isn’t some over the night trick to pull of crowd sims.

The only easy crowd system I’ve seen that works easily and quickly is birds. They can be done with particle systems, have simple animations, and it looks better when more birds are used.

#28

Good points…

The funny thing is that I was planning a crowd sim fxwars.

Any ideas how people could pull that off?

-R

#29

Not easily. It’d mostly take a lot of particle tweaking, and there’s definite ceiling to getting it looking good. To get believable crowdsim, you really do need something like Massive.

There’s a couple AI-driven particle systems out there, like BrainBugz, which allow for simplistic crowd behavior (flocking, pathfinding, etc.), but the problem, as far as I’m concerned, is always going to be getting actual deforming objects instanced, as opposed to swapping out static geometry at every frame. With PRman or MR standalone, you could write a shader or DSO which could do it (which Massive did), but it’s a little beyond the capabilities of standard Maya/XSI/Max. Though I dunno, is XSI’s Behavior cheap now? I don’t know much about that one.

#30

How many people have a $30,000 copy of Massive?

I’ve got a copy of Massive … in my dreams though!:wink: hehehe;)

Thought it was $18,000 though.

#31

Heh, you got me before I realized he’d ask “how” could people do it, not “how many”.

The full version, which allows you to make your own agents, is $30,000. Massive Jet, which allows you to use ready-made agents, is $12,000. Or thereabouts, I may have it confused.

#32

The full version, which allows you to make your own agents, is $30,000. Massive Jet, which allows you to use ready-made agents, is $12,000. Or thereabouts, I may have it confused.

hmmm … interesting … still very expensive. anyway thanks for informing me about the prices, cos I think the $30,000 one must have slipped under my eyes. :wink:

#33

This is from the CS documentation in Max 8.

“The crowd-animation system in character studio is designed to simulate the behavior of real-life crowds. A crowd simulation emulates real-life situations by animating delegates (helper objects that act as representatives). You give the delegates overall guidelines on how to behave, and the crowd simulation calculates their motion.”

So it can be done, but I’ve never tried.

As for the character itself you would most likely want to have a design where if they look the same it makes sense. Like lizard aliens without any cloths.

Once you try to do natural people. Then they need different faces, cloths, body types, and gender.

We’re already doing a project like this at work. For a stadium of about 70 digital people. All coming from one custom character rig that supports parametrics. You need to build a custom pipeline, scripts, and organize data structures to pull it off. It’s a lot of work, and in the end viewers make very quick judgements as to how real it looks. If you have to fix an error in the clothing or face. It can takes days to transfer those changes to all the many characters.

In the end it never looks as good as working on just 1 character. You waste so much time just trying to get your crowd to look right that time is lost on details of the character.

#34

Getting back on topic a little, i think that with a little propaganda you could get ppl into this.
I would love to see all the talent arround here put together, because i bet it can be done.

#35

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.