How long are you going to want to stay on a massively wide establishing shot full of tiny people? A few seconds?
I used to play a little game on my powerbook 140 where hundreds of little people were given a random map to move about-- hills, valleys, ladders, etc, and you could aim various calamities at them: floods, earthquakes, lightning bolts, and so on.
You misunderstand my point. Itâs not that a realistic reacting crowd canât be done. But it takes MUCH more time and power to render. Since these are obviously joke-renders, why would the âartistâ go through the trouble of doing that. And then heâd need to either use a render-farm or wait 2 days for the render, while these videos can be rendered live in real-time on a modern home PC. If you look at the single people, youâll see that they only have the one behavior âmove in a straight line from point a to point bâ. They donât even âknowâ how to get up after falling (which could have made the videos that much funnier). This shows you, that there is no virtual intelligence plugin running at all - which makes sense if the thing you want to try out is the crowd control algorithms.
LOL Bid seriously?? Why would it take longer to render more ârealisticâ crowd? Im guessing you have never done a crowd sim.
Massive and Softimage can render 100,000 in about the same time it takes to render 100.
There is a big difference between render time and sim time. Both things you describe wouldnt take much longer to sim or render.
Btw, anyone doing a serious crowd sim would cache the point cloud and render that instead of rendering while simming frame by frame.
Honestly what you wrote doesnt make much sense in the real production world. Not many apps can distribute a simulation so most likely its on a single machine doing the sim, which is most likely just as powerful as the average gaming machine. Of course Maya is ancient (both software and civilization) and single threaded so⌠well whatever. what you wrote isnt very true.
oh lord. i had forgotten all about that game. thanks for the memories.
Iâd bet that it was Populous. I used to play it on my Amiga 1000.
I know itâs fake, that those arenât real people. And I laughed so hard.
Iâm not a bad person, am I?
Also, hereâs the follow-up on how the arrested men were tortured:
Â
What the âŚ? Are you guys serious? Either you are putting A LOT of work into understanding me wrong or my grasp of the english language is worse than I previously thought. So on the off-chance you arenât trolling, but really donât understand me, Iâll explain yet again:
-
YES, REALISTIC CROWD RENDERING CAN BE DONE. (Some examples: https://www.youtube.com/watch?v=zXiVNF9uyTU )
-
These videos look like they may be rendered in real-time on a home PC - for fun. If you think your boss is OK with you using the render-farm for your joke-vids, go for it. Why bother is a question youâll have to answer yourself.
-
If you follow the start of the discussion you joined, youâll find that we were talking about REALISTIC BEHAVIOR of the single actors. Maybe itâs because Iâve been working with artificial intelligence for years, but I donât consider a small randomized decision tree per actor ârealistic behaviorâ - why do you think most automatic crowd sims only âworkâ if the camera doesnât stay on one actor for more than a few seconds.
If you can show me a demo that renders that many actors with full-blown virtual intelligence (meaning they at least react to the environment and the other actors in a realistic way, try to get out of harms way, etc,) in real-time on todays PCs, feel free to post a link. I love to be amazed.
Good bosses know that allowing a bit of on-the-job fun can do wonders for the crewâs morale.
Exactly how much resource use then counts as âtoo muchâ is an exercise in cost-benefit resource allocation.
So again, crowd sims for the most part are not something you distribute. They need to simmed on one machine and one machine only. you dont need your bosses renderfarm, you need one computer and a little bit of time. Massive did Lord of the Rings like a decade ago. There is a very big difference between rendering and the simulation. Once the simulation is baked it should take only seconds to render each frame.
Them run good, but not smart
Itâs not turtles all the way down.
.
.
.
itâs
.
.
.
.
.
dicks!!!
.
.
.
.
.
.
.
.
.
.
.
.
no homo
That depends on so many factors. Yes the âsimulationâ itself is largely independent but we had renders take more than a minute per frame on a distributed render farm in my last job and not for anything really fancy or complex. Itâs not really my area of expertise because I concentrate on the real time rendering field, but making any claim for rendering time is same as making a claim for how long is a piece of string. But by the same token any form of baking whether complex simulation logic or other is also totally open to whatever the factors are determining the calculations needed. A simple scene with simple point light bake might still take me many hours on a 6 core cpu and thatâs just to produce in effect a single texture map, not multiple frames.
thanks yeah im sure the crowd simulation algorithm being intentionally used here in a playful and unrealistic way will welcome your useful criticism on boingboing.net
What do we do but take light-hearted posts seriously, and dark-hearted posts lightly? You hath disgraced him, and hindered him half a million; laughed at his losses, mocked at his gains, scorned his nation, thwarted his bargains, cooled his friends, heated his enemies; and whatâs your reason?
He is a BoingBoing commenter. Hath not a BoingBoing commenter eyes? hath not a BoingBoing commenter hands, organs, dimensions, senses, affections, passions? fed with the same posts, hurt with the same sticks and stones, subject to the same imaginary diseases flesh is prone to, healed by the same apple-cider vinegar, warmed and cooled by the same winter and summer which have remained roughly the same anthropocene-epoch claims to the contrary, as a lurker is? If you prick us, do we not bleed snarky comments? if you tickle us, do we not ROFLMAO? if you poison us, do we not die? and if you wrong us, DO WE NOT POST IN ALL-CAPS SPEWING SEMI-RELEVANT GIF AND YOUTUBE LINKS?
Yeah, I must admit, that I used the term ârenderâ quite loosely above, since most people arenât familiar with the steps involved in the whole process. Of course I meant the time including pre-baking stuff and post-rendering processes. On that topic: 2 or 3 years ago, I read an interesting article about distributed on-demand rendering of physics and virtual intelligence. With the systems they used they actually got better times out of this than by prebaking everthing. And of course they didnât need all the memory the cache the whole thing in advance. But since this is not my line of business I have no idea what the big players use today.
wow
Another 40,000 coming every dayâŚ
The villany you teach me, I will sudo execute
, and it shall go hard but I will better the instruction.
Â
Â
On a completely irrelevant tack, I always initially mentally-pronounce your name as âcemen-tennialâ for some reason.