Planetside Software Forums

General => Image Sharing => Topic started by: Tangled-Universe on July 27, 2009, 12:12:33 PM

Title: "new" principle for canyon mesas
Post by: Tangled-Universe on July 27, 2009, 12:12:33 PM
Hi Guys,

Unfortunately I don't have much time right now to elaborate on how it's done.

Basically I replaced the base-terrain setup of my canyon-technique with just 2 powerfractals in order to generate mesa-like structures.
So now the entire planet is made up of these mesa-structures.
I will try to further develop this technique, things like a better talus for example.

I think I have already isolated the fan-like displacements.
The lighting is a bit strong and harsh, but not really the focus anyway.

Martin
Title: Re: "new" principle for canyon mesas
Post by: RArcher on July 27, 2009, 02:01:50 PM
Martin,

The wall variations are really fantastic!  If you have indeed figured out how to get rid of the sharp fan displacements and cutoffs then this is damn near perfect.
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on July 27, 2009, 02:05:38 PM
Quote from: RArcher on July 27, 2009, 02:01:50 PM
Martin,

The wall variations are really fantastic!  If you have indeed figured out how to get rid of the sharp fan displacements and cutoffs then this is damn near perfect.

Thanks a bunch Ryan ;D
And yes, I've isolated the problem and already tested the solution, it's fixed now as far as I know :)
Title: Re: "new" principle for canyon mesas
Post by: Gannaingh on July 27, 2009, 02:55:07 PM
I agree with Ryan, the wall variation is superb!
Title: Re: "new" principle for canyon mesas
Post by: Hetzen on July 27, 2009, 04:16:56 PM
I think I understand what you've done. I also think a 64bit version of TG2 will help. More levels of grey?
Title: Re: "new" principle for canyon mesas
Post by: FrankB on July 27, 2009, 04:17:11 PM
cool, definitely different with these "other" strata :-)



Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on July 27, 2009, 04:59:36 PM
Quote from: Hetzen on July 27, 2009, 04:16:56 PM
I think I understand what you've done. I also think a 64bit version of TG2 will help. More levels of grey?

I'm not exactly sure what you mean, but you mean that with more levels of gray the fractals could be more accurate and wouldn't cause fan-shaped displacements?
In this case I had enabled "continue spike limit" in a couple of powerfractals and they seem to do the trick for solving it in this case.
There was also a height-restriction which was off and caused minor problems.
Title: Re: "new" principle for canyon mesas
Post by: Hetzen on July 27, 2009, 05:05:14 PM
Not tried the spike limit, but I do know that TG doesn't like creating vertical triangles, somehow you need to temper that. More 'grey' that can be compressed with a greater bit resolution would help from what I've found and understood.
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on July 27, 2009, 05:09:48 PM
Ah like that, yes TG2 doesn't really like extremely steep surfaces, but they're still quite manage-able.
I agree with you that a greater bit-depth could aid, but I'm not sure this is really how TG works, since a lot is being "simplified" to make it work faster.
A possible solution for the problem you describe is to play with the patch-size of your compute terrain and normal.
Title: Re: "new" principle for canyon mesas
Post by: aymenk2003 on July 27, 2009, 07:07:18 PM
if we ignore the fan-shaped displacements... then it will be a PERFECT image...nice very nice ...let us see more ...
Title: Re: "new" principle for canyon mesas
Post by: choronr on July 28, 2009, 07:57:16 PM
You vertical surface is shaped beautifully. This is what you see here in various areas ...keep on please.
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on July 29, 2009, 04:24:16 AM
Thanks guys :)

I think I have solved the fan-shaped displacements.
It was due to restrictions of my experimental sand-shader. Better said: lack of restrictions ;D
I also improved the strata further and as far as I can tell from the crops it will be the best I ever made :)
So I'm curious as well how it turns out. Will post it in the next 12 hours.

Martin
Title: Re: "new" principle for canyon mesas
Post by: inkydigit on July 29, 2009, 05:26:58 AM
I await with baited breath!
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on July 29, 2009, 03:45:05 PM
So here it is....no more fan-shaped displacements and more properly constricted and scaled sand. Though the ripples could still be a bit smaller imo.
I also tried an other atmosphere but I don't like it.
Further I added another strata-shader to the upper part of the mesa-structure.
If you compare this image with the previous one you can see this one gives a nicer layered style with some little breakups by a renewed voronoi-function.
The renewed voronoi-function is a blend of Frank's advanced cracks clip-file and the one I've been using for many months already.
You can find Frank's advanced cracks clip-file here: http://forums.planetside.co.uk/index.php?topic=6432.msg67919#msg67919 (http://forums.planetside.co.uk/index.php?topic=6432.msg67919#msg67919)

I'm going to try to do some work on the talus and then I consider it almost finished. Suggestions are always welcome of course :)
Thanks for stopping by again!

Martin
Title: Re: "new" principle for canyon mesas
Post by: inkydigit on July 29, 2009, 05:45:25 PM
looks much better, Martin!...a great scene for a river? the talus will be the icing on the cake, so to speak!
Title: Re: "new" principle for canyon mesas
Post by: choronr on July 29, 2009, 06:16:15 PM
Quote from: Tangled-Universe on July 29, 2009, 03:45:05 PM
So here it is....no more fan-shaped displacements and more properly constricted and scaled sand. Though the ripples could still be a bit smaller imo.
I also tried an other atmosphere but I don't like it.
Further I added another strata-shader to the upper part of the mesa-structure.
If you compare this image with the previous one you can see this one gives a nicer layered style with some little breakups by a renewed voronoi-function.
The renewed voronoi-function is a blend of Frank's advanced cracks clip-file and the one I've been using for many months already.
You can find Frank's advanced cracks clip-file here: http://forums.planetside.co.uk/index.php?topic=6432.msg67919#msg67919 (http://forums.planetside.co.uk/index.php?topic=6432.msg67919#msg67919)

I'm going to try to do some work on the talus and then I consider it almost finished. Suggestions are always welcome of course :)
Thanks for stopping by again!

Martin
This is looking very good Martin. I would suggest some slight color adjustment to the yellow color on the rock faces; probably darker and a bit less reflective. Try adding a tint of blue in your red/orange colors as well. The sand is also looking good as well - in the canyon and on the foreground.
Title: Re: "new" principle for canyon mesas
Post by: CCC on July 29, 2009, 08:49:46 PM
Could you embed some stones inside the canyon walls by any chance? Some more color would work well but in harmony with the real thing.
Title: Re: "new" principle for canyon mesas
Post by: Thelby on July 30, 2009, 02:41:10 AM
Looking very nice in this. The strata is really good.
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on July 30, 2009, 03:38:56 AM
Quote from: choronr on July 29, 2009, 06:16:15 PM
This is looking very good Martin. I would suggest some slight color adjustment to the yellow color on the rock faces; probably darker and a bit less reflective. Try adding a tint of blue in your red/orange colors as well. The sand is also looking good as well - in the canyon and on the foreground.

Quote from: CCC on July 29, 2009, 08:49:46 PM
Could you embed some stones inside the canyon walls by any chance? Some more color would work well but in harmony with the real thing.

Thanks you for the suggestions!

@Bob: I'll see what I can do about the yellows. It's probably the brightness which makes you think there's reflectivity, because there isn't. I certainly going to try to add some more blue in the red/orange. I'm a little bit color-blind so it's always useful to hear this kind of stuff.

@CCC: that doesn't sound too hard to do. What do you have in mind? Patches of rocks or complete (low) coverage? And which colors, darker or lighter?

Martin
Title: Re: "new" principle for canyon mesas
Post by: CCC on July 30, 2009, 05:15:39 AM
Well, when ever there are canyon strata layers there are all sorts of rocks everywhere in many cases so what about both clusters and low coverage that tend to follow the strata lines if that is possible and i'd say both light and medium/light warm greys and sandy brown greys. Some faded red brown would look interesting. This looks like sandstone so i think those colors would look harmonious to the terrain itself.
Title: Re: "new" principle for canyon mesas
Post by: Gannaingh on July 30, 2009, 12:40:02 PM
Looks great!
Title: Re: "new" principle for canyon mesas
Post by: littlecannon on July 30, 2009, 12:52:49 PM
Martin, Anything you do with canyons is cool.... this is looking very nice. I think I can see an exploded stone on the foreground in the middle, right near the bottom of the image (splitting hairs here). How did you get the overhangs at the top? Was it a restricted Redirect with a Billow PF? The sand is looking superb now.
Cheers, Simon.
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on July 30, 2009, 03:21:24 PM
Thanks Darth and Simon :)

Quote from: littlecannon on July 30, 2009, 12:52:49 PM
Martin, Anything you do with canyons is cool.... this is looking very nice. I think I can see an exploded stone on the foreground in the middle, right near the bottom of the image (splitting hairs here). How did you get the overhangs at the top? Was it a restricted Redirect with a Billow PF? The sand is looking superb now.
Cheers, Simon.

Yes I saw the stone too...don't know what caused this. I think it's exactly the same thing Dune posted about in the general forum. Just bad luck I guess.
I guess you mean the overhangs at the top in the background? They are unintentional :) I indeed do use redirected powerfractals for creating overhangs.
Title: Re: "new" principle for canyon mesas
Post by: Cyber-Angel on July 31, 2009, 12:23:35 AM
Quote from: CCC on July 30, 2009, 05:15:39 AM
Well, when ever there are canyon strata layers there are all sorts of rocks everywhere in many cases so what about both clusters and low coverage that tend to follow the strata lines if that is possible and i'd say both light and medium/light warm greys and sandy brown greys. Some faded red brown would look interesting. This looks like sandstone so i think those colors would look harmonious to the terrain itself.

I think I know what been refereed too here, do a Google search for "Rockfall" and have a look a what comes up on Google Images. Just a theory.

Regards to you.

Cyber-Angel  ;D   
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on July 31, 2009, 06:32:20 AM
Thanks cyber, appreciate it.
I'll work more on this in the weekend, first have to finish other stuff.

In the meantime I perhaps could use some suggestions on how to make a river through these mesas.
I'd like to have a river which doesn't reach the talus but still has a bit of "beaches" (so, say a kind of offset from the talus).
Anybody a straight idea?

Martin
Title: Re: "new" principle for canyon mesas
Post by: CCC on July 31, 2009, 08:35:41 AM
Rockfall, in other words talus build-up caused by thermal and fluvial erosion. What i was referring to is just the rocks embedded in the rock strata layers themselves. I think the talus itself would be more easy to achieve then the strata rocks but i could be wrong.

Is there a clamp node you can use (sorry, don't have TG open at the moment). Clamp the lower terrain for a lake shader for the river. Perhaps the clamp could be softened by another node, maybe smooth erode? That way the edges blend well into the dry terrain just before hitting the talus.
Title: Re: "new" principle for canyon mesas
Post by: buzzzzz1 on July 31, 2009, 09:08:08 AM
Quote from: Tangled-Universe on July 31, 2009, 06:32:20 AM
Thanks cyber, appreciate it.
I'll work more on this in the weekend, first have to finish other stuff.

In the meantime I perhaps could use some suggestions on how to make a river through these mesas.
I'd like to have a river which doesn't reach the talus but still has a bit of "beaches" (so, say a kind of offset from the talus).
Anybody a straight idea?

Martin

I'm nowhere near as advanced as most of you, but just a thought here. Have you considered checking smooth shoreline in all your displacement settings including the fractal breakup shaders?  You can also adjust the smoothing altitude and amount there.
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on July 31, 2009, 10:40:46 AM
Thanks for your suggestion Jay.
I'm aware of it, but I never tried it seriously and will definitely look into this.

However, I already have flat surfaces between the mesas.
What I do not want is water up to the beginning of the talus.
So I'd like to make a sort of thin/small "ditch" exactly in the middle of the mesas. See what I mean?

Martin
Title: Re: "new" principle for canyon mesas
Post by: Gannaingh on July 31, 2009, 01:27:50 PM
I suppose if it came down to it you could just paint your own ditch with a painted shader, (that's what I'm having to do in my most recent project) though at that point it sort of mucks up the spiffy procedural workflow.
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on July 31, 2009, 02:09:19 PM
Quote from: darthvader1 on July 31, 2009, 01:27:50 PM
I suppose if it came down to it you could just paint your own ditch with a painted shader, (that's what I'm having to do in my most recent project) though at that point it sort of mucks up the spiffy procedural workflow.

Indeed :) But ultimately, if I can't find a decent procedural way I must use the painted shader. Such is Terragen 2 life, sometimes :P
Title: Re: "new" principle for canyon mesas
Post by: RArcher on July 31, 2009, 02:37:54 PM
Martin,

I think that if you take your Mesa structures, invert them and play with the settings a bit then displace that output downwards, it should carve out some channels.  I'll send you something that seems to work, just needs a little more adjusting.
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on July 31, 2009, 03:26:03 PM
Quote from: RArcher on July 31, 2009, 02:37:54 PM
Martin,

I think that if you take your Mesa structures, invert them and play with the settings a bit then displace that output downwards, it should carve out some channels.  I'll send you something that seems to work, just needs a little more adjusting.

I was working on the exact same approach, though I haven't found a way to create an offset of a fixed distance. Probably needs some blue nodes work which I'm not good at at all.
I've received your file, will check it right away, thanks :)

Martin
Title: Re: "new" principle for canyon mesas
Post by: domdib on August 10, 2009, 06:35:14 PM
Coming rather late to this discussion, but just wanted to say that I love the canyon walls in particular. Is this going to be a Canyon pack Mark 2?
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on August 10, 2009, 06:55:37 PM
Perhaps, but I'm not sure yet, since there are still some minor problems here and there.
My intention is to send it out to Canyon Pack purchasers, so no seperate product.

Martin
Title: Re: "new" principle for canyon mesas
Post by: RArcher on August 10, 2009, 10:09:11 PM
Martin was kind enough to let me play around with his mesa techniques.  Here is a very rough render of a combination of his displacements as well as some other additions.  I'm not particularly happy with the colours and plants, but the canyons walls are great.
Title: Re: "new" principle for canyon mesas
Post by: Henry Blewer on August 10, 2009, 10:17:36 PM
The walls really are good. Could a power fractal overlay correct your color issue? It would be easier than recoloring every layer, besides I think the color is good.
Title: Re: "new" principle for canyon mesas
Post by: Tangled-Universe on August 11, 2009, 02:35:18 AM
That looks nice Ryan! Which setting did you use for the GI? The shadows are a bit dark and so are also the plants' inner shadows.
You found yourself a nice POV!

Quote from: njeneb on August 10, 2009, 10:17:36 PM
The walls really are good. Could a power fractal overlay correct your color issue? It would be easier than recoloring every layer, besides I think the color is good.

Yes, good suggestion!
Title: Re: "new" principle for canyon mesas
Post by: Phylloxera on August 11, 2009, 07:35:41 AM
Quote from: RArcher on August 10, 2009, 10:09:11 PM
Martin was kind enough to let me play around with his mesa techniques.  Here is a very rough render of a combination of his displacements as well as some other additions.  I'm not particularly happy with the colours and plants, but the canyons walls are great.

The rock is realistic, the vegetation is less!
Title: Re: "new" principle for canyon mesas
Post by: littlecannon on August 11, 2009, 06:56:53 PM
The rock is... stand back, jaw hit floor amazing :o. The POV is great RArcher. I can't wait to find out how to do this. I guess this is using the "Merge Strata MagicĀ®" you like to weave Martin?
Cheers, Simon.
Title: Re: "new" principle for canyon mesas
Post by: CCC on August 11, 2009, 07:11:03 PM
Looks good other then need some stones embedded inside the strata layers.
Title: Re: "new" principle for canyon mesas
Post by: Gannaingh on August 11, 2009, 07:25:49 PM
I agree, the rock wall looks fantastic! Good work TU/Ryan!
Title: Re: "new" principle for canyon mesas
Post by: MacGyver on August 11, 2009, 08:02:37 PM
Is it just me or is it really that amazing how mathematical functions can imitate nature? It has always been a stunning realization to me how interwoven these two so seemingly seperate entities are, one a construct of the human mind, based on axioms, the other one in front of your nose ;D
By the way - really outstanding work! :)
Title: Re: "new" principle for canyon mesas
Post by: Seth on August 12, 2009, 01:13:48 PM
great walls and very nice pov :)
Title: Re: "new" principle for canyon mesas
Post by: goldfarb on August 12, 2009, 01:33:11 PM
Quote from: MacGyver on August 11, 2009, 08:02:37 PM
Is it just me or is it really that amazing how mathematical functions can imitate nature? It has always been a stunning realization to me how interwoven these two so seemingly seperate entities are, one a construct of the human mind, based on axioms, the other one in front of your nose ;D
By the way - really outstanding work! :)

check out Vol Libre... one of the most important films ever made...

http://vimeo.com/5810737

and a bit from Droidmaker, via kottle.org:
http://kottke.org/09/07/vol-libre-an-amazing-cg-film-from-1980
Title: Re: "new" principle for canyon mesas
Post by: MacGyver on August 12, 2009, 03:30:35 PM
Quote from: goldfarb on August 12, 2009, 01:33:11 PM
Quote from: MacGyver on August 11, 2009, 08:02:37 PM
Is it just me or is it really that amazing how mathematical functions can imitate nature? It has always been a stunning realization to me how interwoven these two so seemingly seperate entities are, one a construct of the human mind, based on axioms, the other one in front of your nose ;D
By the way - really outstanding work! :)

check out Vol Libre... one of the most important films ever made...

http://vimeo.com/5810737

and a bit from Droidmaker, via kottle.org:
http://kottke.org/09/07/vol-libre-an-amazing-cg-film-from-1980


Thanks for sharing, I'll definitely check it out :D
Title: Re: "new" principle for canyon mesas
Post by: Stridemovies on November 01, 2009, 09:16:27 PM
Hi Martin, Love the Strata and the scale. Are you still working on the image any chance of making the file available.
Also on the canyon pack do you need to get the TU Snow pack to produce a sand covered base as in some of the promo shots.
Title: Re: "new" principle for canyon mesas
Post by: TheBlackHole on November 01, 2009, 09:46:33 PM
It would be cool to see some rocky mountains and overturned blocks of rock created by an ancient planetary cataclysm. That would be PERFECT for the planet Jhasnak, a planet in the same system as Dgnyre, the gas giant in my image "Unknown Moons: Ru'Ep".
Title: Re: "new" principle for canyon mesas
Post by: MGebhart on November 01, 2009, 10:06:52 PM
Perfect!

Marc
Title: Re: "new" principle for canyon mesas
Post by: littlecannon on November 04, 2009, 12:04:59 PM
Hey TU.... any further developments on this... As I have bought the canyon pack I'm eager to see how you did this.
Cheers,
Simon.