Animating tree growth

Started by Mohawk20, October 11, 2008, 11:54:43 AM

Previous topic - Next topic

Mohawk20

I'm currently working on a new animation project called the Genesis Project (find all the info about it here: http://www.ashundar.com/index.php?topic=4533.0 ).

For this I want forrests to rapidly grow from one side of the screen to the other...
Now, I can do this a few dozen slim populations in side by side, all having a separate growth animation, but that would take quite some time, and I'm not sure my system could handle it.

I have thought of a possible solution, and I wanted to run it by you guys (specially Oshyan or Matt)..
My understanding of the populator system is that any seed will place the objects on a specific spot with specific coordinates. Would two moving populations have the same tree in the same spot?
Would it be possible to have 3 slim populations move position from left to right on the screen, the first with small trees, the second with medium small trees, and the third with medium large trees, and have a growing population of full grown trees behind that. If the seeds are the same in all 4 populators, would it then seem like the trees stay in the same spot and grow?
(I know I should need a lot more sequential populations for a smooth anim, but it's just a concept.)
Howgh!

bigben

No it is not currently possible, nor advisable, to move populations.  The placement of objects is currently linked to the population placement parameters so you also move the trees with the population. This will hopefully change in the future.
http://forums.planetside.co.uk/index.php?topic=3269.msg33511#msg33511

You would need a population for each "sprouting" of trees, each with it's own distribution mask. The creation of the masks will be the key to success.  There are a few different ways of doing this depending on what you are trying to achieve.

Oshyan

Maybe try a distance shader attached to the density input of 3 separate populators, all with the same seed and objects, but with varying object scalings. Maybe you'd need multiple distance shaders, come to think of it. But it seems like it might work, albeit in a rough fashion. It would be better if it were easier to control tree scaling in a non-contiguous world space manner.

- Oshyan

Mohawk20

Thanks for the reply guys...

I have done some testing after I posted the question, and I discovered what Ben said about the placement.
Moving populations, did work though. I thought myself that it could simply be solved by having a specific seed per spot, so with every incremental move, the seed also goes up by one. So for every 50 units of moving, the seed will be increased by 50 as well.

Unless seed increments are not that different from each other at one digit changes...
Howgh!

gregsandor

Why not just animate the population mask?

bigben

Quote from: gregsandor on October 12, 2008, 01:30:40 PM
Why not just animate the population mask?

Because the size of the trees is also increasing so each new tree would be the current be bigger when it first appears.

Mohawk20

#6
Well, I did this latest test with 4 populations moving down, keeping the seed on the location by animating the seed incrementally (every single frame animated in detail, luckily just 4 frames  ;)).

See below and compare...
Howgh!

Oshyan

Every seed will be *completely* different from the previous one. There is no real correlation or similarity between "adjacent" seed values (e.g. 4034 and 4035).

- Oshyan

Mohawk20

I'm going to revive this project and try to get this to work using painted shaders and colour adjustments.

But when looking again at the above examples, and Oshyan's comment on the seeds having no correlation, I noticed something.
Every population had a different seed, so the top one has a different seed than the one below, and they all have a seed incremented by one. Should be entirely different. But if you look closely, you can see a pattern like lines from top right to bottom left. The same 3 trees next to each other shows up in every new growing pop, though the seed is different.
This would imply that the seed value actually does have a correlation between values... or at least in the older build used for these tests.
Howgh!

rcallicotte

Mohawk, I do see what you're saying is true.  It would be cool to see about 50 of these one after the other to see not only populations, but clouds too, based upon increments of 1.
So this is Disney World.  Can we live here?

Tangled-Universe

Cool work mohawk, I'm really looking forward to see where this goes. This could give quite spectacular shots :)

Martin