New Rocks

Started by efflux, December 06, 2012, 07:10:45 PM

Previous topic - Next topic

efflux

This is called New Rocks because I started a thread about four years back about this general angle. I've removed the tgd from that one though.

Here's a couple of graphs which create rocks. They are quite mushroomed. That is just a feature of these particular graphs. In Rocks 7.0, it's Rock profile that will adjust that and the blend start and end also have effect on how the rocks surface blends in. The overhang form is controlled in the colour tab of a fractal. In Rocks 6.0 it's Low High profile that mushrooms them. The rock "stems" surface fractal has no displacement. This would break at the join to the ground.

I may work out a way of linking the Grow and Low High profile in Rocks 6.0 because those things would be better linked. It just needs to be done a certain way. "Growing" the rocks isn't as good as I'd like.

There are both completely different. I could make lots of variations. The way I build these all depends on the outcome I want. If I don't need it to have a certain feature I don't built that in. Various power functions hooked in at different places can alter the rock shapes. that's just a matter of the shape you want.

The names on the adjusters may not be entirely logical. Rocks 7.0 has "Strength". that's a power which tends to puff the rock out before clamping.

Some adjustments will break the graphs. Watch for that. Breaks in displacement blend will occur but these graphs are not fool proof shaders that can be used any old way. To explain where breaks may occur would take some time. Just watch for it. Don't apply displacement to the "stems" on Rocks 6.0 for example. Some settings will wildy change the effect of others. I don't know that the settings I have made are the safest to start with in this respect.

Rocks 6.0 can be clumped in their distribution. Rocks 7.0 doesn't have that. Rocks 7.0 is just not as integrated into a scene the way the Rocks 6.0 is but Rocks 6.0 is perlin based so needs a bit more control of it's distribution.

In general, Rocks 6.0 should be kept small (probably smaller than I have made them here). Rocks 7.0 tends to work better for increasing the rocks to much larger scale or taller. Rocks 7.0 can be more like rocks sitting on small hills of the surrounding surface, maybe eroded at their base. This all depends on how you tweak them. In general, l use Rocks 7.0 (or variants of it - some less mushroom effect) if I want larger rocks that don't mask off perfectly with the surrounding ground like the fake stones do. I may add other variants.

Rocks 7.0 general type of form (voronoi based) can also have another feature built in which I haven't done yet that would allow you to have rocks of variation in surface i.e. one rock has a colour and displacement of one surface fractal and another a different surface and lots inbetween. Like the way you can have colour variation on the fake stones. That requires the graph to get bigger though. I will also be building in a way to distribute the rocks more unevenly in a similar fashion to fake stones. Using a seperate fractal to laterally displace Rocks 7.0 seems to be quite cool. This is what makes them good for bigger rocks.

efflux

#1
Here's another variant of Rocks 7.0. It's a bit more suitable for having more standard rock forms.

EDIT: tgd removed. A better one is on a later post. It's a slightly different graph but also has colour differences built in.

efflux

Here I changed settings in Rocks 6.0. No change to the graph though. Because these are perlin they can eventually all almalgmate into one rock covering rather than separate rocks.

efflux

Two more earlier variations on Rocks 6.0. The first is closer to fake stones territory but the perlin doesn't have the voronoi edges.

efflux

Another variation from Rocks 7.0. Just seeing what this looks like with water.

efflux

One feature of Rocks 7.0 is that they will pick up displacement from the underlying surface. This can cause them to look more interesting or can make it look worse depending on the surfaces used. The graphs can be swapped and mashed around in any which way. Try Perlin on Rocks 7.0 for example. There is also a way to altitude distribute the rocks so that they don't fade out with the surface layer distribution. I'll eventually branch off various ideas to more specifics.

Chinaski

My hero!

I was just playing (since one week) with your "old" rock clip file. It's a very interesting approach because there is no cahotic collisions between rocks, we can easily mix the surface texture, and it's not too heavy to compute (my PIII thanks you).

My goal was to deform the soil near the rock, to have more interaction between them... Basically like that:

https://secure.flickr.com/photos/maxxwellsmart/4241026157/
https://secure.flickr.com/photos/corica/5240870424/
https://secure.flickr.com/photos/danskiedijamcojr/8163653337/
https://secure.flickr.com/photos/24562498@N03/8116785401/
https://secure.flickr.com/photos/chrisk1982/6919572624/
https://secure.flickr.com/photos/borealnz/3313714500/

And I was stuck here:



I'll see what I can do with your upgrated files... And be busy for a while. Thanks! ;)
You don't understand me ? That's normal, I don't speak english.

efflux

#7
Hi. I'll take a look at that graph. I think you get into difficulties if you mix up using scalars with vectors to do stuff here.

There are actually lots and lots of ways to do these things. On another file I have somewhere I have a slider which kind of flows another displacement around rocks but it doesn't touch them. I need to revisit that. It's not the same angle as here in what I was doing so needs patching around to have something definitely stone forms. For example, I think snow often melts around stones. That's the effect.

My feeling is that fake stones are great if you want the stones to look very separate from surrounding ground. Often this is the case in reality but often ground either rides up the stones or is different arround the stones. Blending the stones in other ways becomes problematic because you can't get what's inside. Also, these big boulders in Rocks 7.0 don't have to be big boulders. They can just be a slightly raised surface. Make them very large with less height and overhang.

I'm going to delete a previous tgd and add the one below because this rebuilds the idea of the graph in my first Rocks thread i.e. two surfaces can cover the rocks by various degrees. This time the whole arrangement works as a surface layer - rocks 1 and 2 surfaces and ground. This way you can add other surfaces into those powerfractals whatever way you want. I removed the smooth step and used gain for all the surface blending. Smooth step may be better for the eroded looking ones. I'll revist these different approaches.

Also attached is a screenshot of voronoi Cell and Diff. Notice the shape is the same as long as seed and size are the same. That means we can use cell to control masking of diff. You can use it to distribute the rocks as well as control the surface to get different blend levels of surfaces. I'm not going to do distributions with the cell here though for now because I'm working on simpler voronoi ground cover forms, not big rocks and I'll bring the cell blending in there.

efflux

Chinaski,

I'm not on my Windows system at the moment so can't check this out but in theory, fake stones should use the same shape as voronoi diff if you tweak them the same way hence it should be possible to simply get stuff to flow around the stones. That's kind of the opposite to what I'm doing here. This should be tested with fake stones if it hasn't been already.

efflux

#9
One thing I'm not clear on about the fake stones is how they treat colour and displacement differently i.e. you plug a fractal in with colour and displacement but the fake stone has to treat those differently to blend with the ground. Does anyone have any clue on how that happens? There is no point in exactly recreating fake stone but that aspect of them feeds into some ideas with using the graph to also treat surrounding ground.

Zairyn Arsyn

thanks for sharing these, looking forward to playing with and rendering these nodes, later.
:)

WARNING! WIZARDS! DO NOT PREDICT THE BEHAVIOR OF OTTERS UNLESS YOU OBEY BIG HAPPY TOES.

i7 2600k 3.4GHZ|G.skill 16GB 1600MHZ|Asus P8P67 EVO|Evga 770GTX 4GB|SB X-FI|Antec 750W
http://zlain81.deviantart.com/

efflux

Here's a test where I loaded in a ground surface to show why I'm doing this with the rocks. The ground surface rides up the edges of the rocks. Maybe sand would be good if anyone has something along those lines.

I may test ways of still having a tight join to the rocks from the ground but increase of displacement around rocks. Avoiding rocks and swirling around is no problem but I have to build another graph. The ground surface is some added voronoi. I'm working on some low level surfaces that also clamp and distribute but that's a different topic.

N810

Very kool rock experiments efflux.  8)
Hmmm... wonder what this button does....

efflux

I'm just doing some tests to see what these look like in a better environment. Distributing them is now the key thing. I'm going back to a more basic format to test that. This render is basically the previous two files set up but with different surfaces. One is a voronoi surface I've worked on which is on the rocks. There is obviously terrain and the rocks surface layer is simply altitude distributed. That means the rocks fade out. I can make it so they distribute without doing that but that will come a lot later. The graph will get too big.

efflux

I replaced two of the earlier Rock 7 files. Colour is controlling lateral displacement. Colour was set to 5 then displacement multiplier to 1. I've reversed that. It's the same thing but probably more logical to have the displacement multiplier being where you displace.

A new file is coming which is another variant on the basic idea.