## Wild Strata - Function

Started by Volker Harun, July 24, 2007, 10:00:12 am

#### BPauba

##### July 25, 2007, 07:08:36 pm #15
First off, Awesome work man! I especially love the detail! This thread inspires a topic that I would like to bring up though. How nice would it be if we had a "Functions" forum? "general" is nice but alot of the good threads (mostly functional) get swallowed up by alot of repetitive threads! I think it would be an AWESOME addition. What do you guys think?

#### rcallicotte

##### July 25, 2007, 09:59:35 pm #16
I agree.  If nothing else, it would give me something to read and think about.  I have lots to learn about functions.

Quote from: BPauba on July 25, 2007, 07:08:36 pm
First off, Awesome work man! I especially love the detail! This thread inspires a topic that I would like to bring up though. How nice would it be if we had a "Functions" forum? "general" is nice but alot of the good threads (mostly functional) get swallowed up by alot of repetitive threads! I think it would be an AWESOME addition. What do you guys think?
So this is Disney World.  Can we live here?

#### rcallicotte

##### July 26, 2007, 12:11:12 pm #17
Volker,

I have a couple of basic questions as I do a living autopsy on your shared file (Wild Strata) for this thread -

1.  What is the Height Variation doing?  Does it work in concert with the Depth Variation to control or affect the sin noise values?

2.  Is it planned that the Height Variation and Depth Variation both have Noise Octaves of 12?
So this is Disney World.  Can we live here?

#### Volker Harun

##### July 26, 2007, 03:58:15 pm #18
The octaves are not planned unless I am trying to keep them low. One of those fractals is the copy of the other one. There might be some similarity in both

#### Volker Harun

##### July 26, 2007, 04:21:53 pm #19
Q1:
Normally the Sin of get height would give you parallel data. Too boring. So:

The height variation should give values from 0 to 1 unless clamped by high and low colour. These values are multiplied by 360 (I thought that TG2 is calculating the sin of ° but it does in radiance, as I got tought today) So this value could be lowered to 3.2 for example - but it works with any number - call it the fine tuning.
At the output you get values that are added to the get height so the sin is not calculating the 'displacement' for the actual height but for a height given by the addition. This is where the variation is coming from.
Of course it is influencing the steep- and curviness of the strata and the distance between steps. Vertical variation.

The sin is giving us values of -1 to 1. Turn the high and low colour of depth variation to 1 - you'll see that it is getting too symetrical in depth. So the variation fractal is adding some noise to this depth going from 0 (= no influence) to 1 (=strong variation) at the moment. Feel free to set colours to negative or any other amount that is reasonable, i.e. from 0 to 2, -1 to 1. Variation along the normal.

I hope this helps.

Volker

#### rcallicotte

##### July 27, 2007, 12:02:07 pm #20
Volker, yes this helps.

I'm thinking about buying this book to catch up to some of the concepts here that seem to basic to grasping this program - http://www.amazon.com/exec/obidos/tg/detail/-/0240519353/ref=ord_cart_shr/103-3339306-1319019?%5Fencoding=UTF8&m=ATVPDKIKX0DER&v=glance

I'm ascertaining that you've come to your conclusions to make these node calculations work by your understanding of math and your understanding of the concepts behind procedural graphics.  Of course, you also seem to be intelligent and seem to be a creative talent, both of which are definitely a plus.  I'm almost afraid to ask what you do for a living, since it must be something amazing.
So this is Disney World.  Can we live here?

#### j meyer

##### July 27, 2007, 12:42:41 pm #21
Hi,
V2 to see if i could get something similar out of it quickly,but nope.
OK,no problem,just let´s have some fun with it i thought by myself
and started playing.The results are attached and i hope you don´t mind
that i post them here.Maybe someone can take it further.
Cheers,J.

#### Volker Harun

##### July 27, 2007, 02:39:38 pm #22
I like it. And I really enjoy your thoughts. I'd rather like to learn than just to present ,-)

The voroni seems to be faster - I guess, that I am going to use either of the functions. Mine is for rough and dirty rocks, yours is for subtle use.

Thanks for sharing

Volker

#### j meyer

##### July 28, 2007, 09:02:55 am #23
Glad you like it!  And thanks to you.

#### ProjectX

##### July 28, 2007, 10:57:22 am #24
Quote from: calico on July 27, 2007, 12:02:07 pm
Volker, yes this helps.

I'm thinking about buying this book to catch up to some of the concepts here that seem to basic to grasping this program - http://www.amazon.com/exec/obidos/tg/detail/-/0240519353/ref=ord_cart_shr/103-3339306-1319019?%5Fencoding=UTF8&m=ATVPDKIKX0DER&v=glance

I'm ascertaining that you've come to your conclusions to make these node calculations work by your understanding of math and your understanding of the concepts behind procedural graphics.  Of course, you also seem to be intelligent and seem to be a creative talent, both of which are definitely a plus.  I'm almost afraid to ask what you do for a living, since it must be something amazing.

To be honest, you're best off looking at a math-specific book, especially calculus and trigonometry. The graph sections are the bit you need. You need to imagine the rendered world as a massive graph, with values in the X,Y and Z axis. Shapes in graphs can all be represented with formulae, most (if not all) of which you can represent with nodes in the node editor. Draw a line in the graph along two axis, and you get that shape extended in the third axis. e.g. If I used a formula to draw two opposite curves (which I would add together with a node) in the X and Y axis, then plugged it into a displacement function, it would cut (or mold) an infinitely long cross-section of that shape, across the world in the Z axis. This was the basis of my original canyon shader. While this book will certainly help you with the varying terminologies and the varying methods of 3d renderers, to innovate in the node editor, and be truly in control of it, you need to know your math, and this is no math book. (I am by no means insulting your skill at math, it just happens that I recently had an exam in it, so I've revised it recently and it's still fresh in my mind).

I have been tempted to write a set of simple math tutorials that hold TG2 in mind (no useless things, just the stuff relating to TG2), the problem is I wouldn't know at what difficulty level I should start, and also I'm not a perfect mathematician, not even Uni grade, although some of my friends might be able to help me there. I'd certainly need the help of a good proof-reader who was proficient enough at TG2 to not start asking stupid questions like "what's a node" and "what do you mean by 'plugging a scalar multiplier with a const of 2 into a get x as scalar?'"

By all means buy the book, since it looks like it contains some good stuff in it, but i doubt that much will pertain to the node network (except, perhaps the section on transformations or geometric primitives)

As a side-note, I'd hazzard a guess that Volker is a mafia-boss, who occasionally dips into a bit of flower-arranging for a hobby .

@ProjectX