Planetside Software Forums

General => Terragen Discussion => Topic started by: CCC on September 05, 2009, 04:10:29 AM

Title: Future Implimentations
Post by: CCC on September 05, 2009, 04:10:29 AM
Branch Noise Flavor.

Examples of this noise can be large scale canyon networks digging into the terrains. Smaller canyons then can be added onto other large scale terraced layers along with rivers at the base of the canyons which again can use the same noise but the only difference being is that it is clamped to be made narrow/thin. Other examples say the noise were displaced on inversion can be weird dunes and hilly structures. This does not have to be a complex l-system, just rather basic. Turbulence can be added to control uni-directional and or more omni-directional patterns as well as over all smoothness or rigidness of the winding channels. Perhaps a taper can be added where the channels become more narrow as they end.

An example below of what such a noise can do. Imagine layering this noise several times and altering the look of each noise and including rivers. If the seeds are the same then the multiple channels can follow one another per layer.

Title: Re: Future Implimentations
Post by: cyphyr on September 05, 2009, 07:48:04 AM
Great :) !!
Is this something your working on or a feature you'd like Planetside to implement in a future update. If the latter I suspect we'll have to wait until the SDK is released. If the former ... "Bring it on ...!!"
:D
Richard
Title: Re: Future Implimentations
Post by: Henry Blewer on September 05, 2009, 08:19:04 AM
This is good news. I tried using 2 color L systems from ChaosPro, but it did not work out well. It ended up being easier to let Terragen 2 cut the river/s into the height field.
Title: Re: Future Implimentations
Post by: CCC on September 05, 2009, 03:44:02 PM
Quote from: cyphyr on September 05, 2009, 07:48:04 AM
Great :) !!
Is this something your working on or a feature you'd like Planetside to implement in a future update. If the latter I suspect we'll have to wait until the SDK is released. If the former ... "Bring it on ...!!"
:D
Richard

Heck no. Terragen 2 can not produce such a noise but i'd rather have them or someone create this noise with the SDK. I'd do it myself but i am no programmer and have a tough enough time understanding terragen.
Title: Re: Future Implimentations
Post by: CCC on September 05, 2009, 04:27:00 PM
Speaking of an SDK. A multi-pass rendering option would be nice. It's good for those matte painters out there.

http://www.asilefx.net/catalog/product_info.php?products_id=112
Title: Re: Future Implimentations
Post by: Volker Harun on September 09, 2009, 04:56:59 AM
I have an idea about this ... not the time now, but.
The base idea is to modify the 'Get position in texture'. When it reads for example X=100, Y=0, Z=20 you can exponential increase these values with functions. Doing a power of 2 will give you values of 10.000 | 0 | 400.
In other words, the farer you get from 0|0|0 the smaller steps are. On the other hand: if you have an output for a displacement you can divide by the length of that vector, so farer distance, smaller displacement.

Now, you only need to control the what and where ;)

Volker
Title: Re: Future Implimentations
Post by: CCC on September 09, 2009, 05:10:20 AM
Then there may be hope yet.     :)
Title: Re: Future Implimentations
Post by: Henry Blewer on September 09, 2009, 08:11:01 AM
It is an interesting concept. Please keep posting! I seem to learn about functions better when you do stuff. I've a couple days off to play, so maybe I'll have something cool to show. :o
Title: Re: Future Implimentations
Post by: Volker Harun on September 09, 2009, 09:25:27 AM
Attached is a simple Mandelbrot that feeds a Voronoi Diff Scalar ... and the shader preview.

You might want to go into a direction like this ;)

I have no idea how to blend areas properly for now.

Volker
Title: Re: Future Implimentations
Post by: cyphyr on September 09, 2009, 09:35:25 AM
Great work VH, thanks for sharing :)
Richard
Title: Re: Future Implimentations
Post by: Henry Blewer on September 09, 2009, 09:43:15 AM
Thanks Volker. I saved the image so I can see what's up. The preview image looks intriguing.
Title: Re: Future Implimentations
Post by: Volker Harun on September 09, 2009, 10:33:21 AM
Here is a quick, dirt render.
The nice point of using those formulars is that the falloff is not symetrical and gives nicer patterns close to 0|0|0.
Title: Re: Future Implimentations
Post by: Henry Blewer on September 09, 2009, 03:09:30 PM
Can the large patch be centered anywhere? The rest of it would make great pavers, or dried mud.
Title: Re: Future Implimentations
Post by: Volker Harun on September 09, 2009, 03:15:11 PM
In the upper post with the node network: Send the get position to a transform shader and give those results to the X- and Z-scalar.
You can change the origin inside this transform shader.

Beware, this Mandelbrot-distortion is only 2-dimensional - on far away distances the scale of the Y-axis is relativly large.

Volker
Title: Re: Future Implimentations
Post by: Henry Blewer on September 09, 2009, 03:18:24 PM
A distance shader with constraints multiplying by scale may fix that. Thanks!
Title: Re: Future Implimentations
Post by: CCC on September 09, 2009, 04:19:11 PM
I see variants for the voronoi but how is this good for large scale canyons which the thread was started on? Perhaps i am missing something here.
Title: Re: Future Implimentations
Post by: Volker Harun on September 09, 2009, 04:32:37 PM
Hi CCC,

As I stated, I am a bit busy with other things ... thatswhy I just showed, how to get the falloff, which is in my humble opinion the essentiell part to get those structures. You mentioned in your first post the L-System, which drew me into this direction.
:)
Volker
Title: Re: Future Implimentations
Post by: cyphyr on September 09, 2009, 04:35:51 PM
The variant voronoi is just a starting point, VH likes us to work for our prizes :) That way we get to learn more :)
I was trying to see how a line could be split again and again and the voronoi above has a property like this. Maybe we could adapt the voronoi to work on a line.
:)
Richard
Title: Re: Future Implimentations
Post by: CCC on September 09, 2009, 04:39:11 PM
Quote from: Volker Harun on September 09, 2009, 04:32:37 PM
Hi CCC,

As I stated, I am a bit busy with other things ... thatswhy I just showed, how to get the falloff, which is in my humble opinion the essentiell part to get those structures. You mentioned in your first post the L-System, which drew me into this direction.
:)
Volker

Ok, no worries then. Believe me what you are doing is no doubt rather nifty.    ;)
Title: Re: Future Implimentations
Post by: CCC on September 09, 2009, 04:42:29 PM
Quote from: cyphyr on September 09, 2009, 04:35:51 PM
The variant voronoi is just a starting point, VH likes us to work for our prizes :) That way we get to learn more :)
I was trying to see how a line could be split again and again and the voronoi above has a property like this. Maybe we could adapt the voronoi to work on a line.
:)
Richard

Well, i was never good with math. I think "mountains/hills" and not "power fractal" and that is what i see as far as terrains go. The technical stuff stumps me all of the time.    ;D
Title: Re: Future Implimentations
Post by: Shere Khan on September 09, 2009, 04:44:10 PM
Quote from: CCC on September 05, 2009, 04:27:00 PM
Speaking of an SDK. A multi-pass rendering option would be nice. It's good for those matte painters out there.

http://www.asilefx.net/catalog/product_info.php?products_id=112

amen to that!
Title: Re: Future Implimentations
Post by: Volker Harun on September 13, 2009, 03:35:53 PM
I am just thinking loud, as my brain is working on a similar project.

As we see above, we can use the distance from origin to scale down features - simple.
Now we need a function that does not produce voronoi cells, it has to produce lines. A function like this: http://forums.planetside.co.uk/index.php?topic=7275.msg77527#msg77527

Still the masking is a problem ...