If TG2 is to move down the path of realistic terrain generation then we need to start looking at adding more geomorphology features, to start with Terragen really needs a procedural rock grain generator so that rock looks like rock as seen on this page http://www.angelstarcreations.com/rocks_minerals/rocks.htm as there are certain textures you can not achieve right now more on the subject can be found at this wikipedia article http://en.wikipedia.org/wiki/Particle_size
Real rock has pores (Voids between individual rock particles) and other macro surface detail not currently seen in Terragen or any other terrain generation software at this time ideally Terragen would support mineralogy and come with the ability to generate the thirty known rock making minerals quartz, feldspar etc and be able to generate come rock types such as sandstone easily.
Rock grain (Also known as rock particles) size, shape, distribution etc are an important component of rock and should be included when development time allows.
Regards to you.
Cyber-Angel
I think you can achieve certain amount of detail using powerfractals but I definitely wouldnt mind an easier way!
I think any of these can be reproduced, at least faked in TG2.
I'm with Volker here. Our aim is to provide flexible tools that can create the widest variety of realistic effects possible. Although basis functions are currently limited, there is a remarkable amount of capability in the existing shaders already. It may be a bit more work than a more purpose-built "procedural rock grain generator", but it's possible, and with the same tools you can do a lot more besides, which is really the point.
We could spend weeks of development time trying to perfect a system to reproduce each different type of rock and structure, or we could simply leverage the power of the community and see what people come up with, giving you guys a place to share (the file sharing area) and eventually building up a substantial library of useful effects and surface types. We've already seen this approach in action and it works quite well.
We do intend to provide some examples with TG2, and hopefully a "library" of pre-defined resources at some point (surfaces, atmospheres, etc.), but for now the community is doing a fantastic job of putting those things together on their own.
- Oshyan
The voronoi function can achieve grain or say a sandstone effect but ideally we'd want this in a fractal especially for those people not into tweaking around with the nodes but you can presently simulate everything in those rocks except for light scattering in the more translucent ones.
Oh well just a throat, and what I had in mind would maybe be needed for closeup work any way; but to my way of thinking surfaces in TG2 right now are to smooth (Yes you have displacement and other such methods) but lack that certain some thing that real life rocks have. At the moment (IMHO) you are going to have a hard time convincing a geologist (I happen to know one) that what they are looking at is a convincing representation of the natural world.
Yes, there are many fine examples of rock done with TG2 as it stands today and it is not the intent of this author to take any thing away from those individuals; my concern lies if you where to project a 4K TG2 scene on to an Anamorphic Projection Motion Picture Screen (2.35:1) and have the camera move past a rock face would you be able to convince you audience that what they where looking at is real (Suspension of Disbelief) and secondly how is a TD going match a TG2 Terrain to that of existing terrain captured in live action plates?
Any way TG2 is great (A little baffling at times) I look forward to the first commercial release latter this year, and yes reading the reviews as they come out and what they have say.
Regards to you.
Cyber-Angel
Like most things in the effects industry I think at a certain point you might be better off just resorting to "cheating". In other words use a high resolution image-based texture. Provided a clean texture with neutral lighting and an appropriately matching displacement map I think you'd be hard-pressed to tell the difference between real rock and TG2. Use of image-based textures is a perfectly legitimate technique and if done properly it doesn't reduce any of the advantages of using a package like TG2 (as opposed to real photography).
- Oshyan
Image based texturing seems very successful in TG from what I've seen but I never tried it.
We'll get better and better at getting more naturalistic results, especially as functions get added to TG. I'd suggest that in actual fact using procedurals will eventually lead to much more efficient results for the things you mention because of their infinite detail. You literally could see the grains in a rock and then move to mountain scale. It's still early days for this technology and it's not easy to use or learn and requires a lot of computer power but the results will be worth the effort.
Actually, the voronoi basis in a fractal is pretty crucial. I'd put that at the top of my list now for things needed since the colour gradient issue can be got around in Nodes. At present Blender can create better rocks than TG2. We can create the surfaces and grain of a rock but the larger forms really need voronoi fractal.
I would like to add the point, that once TG2 gets a "macro" feature, we'll be able to package all those complicated function networks in a simple-to-use macro node, with just a few sliders to adjust the essential parameters. That's what I'd really like to see as soon as the new renderer is finalized with its multi-threading support. To me, macros are a great productivity component, too.
This goes way beyond the current experiments on terrain manipulation, but ranges to complete atmospheres, cloud types, celestial bodies, and so on and so on...
I imagine that if the macro capabilities have been available right from the beginning, the overall range of available effects within TG2 must have grown enormously by now, considering all the functions that reside in the file sharing area already.
It's unfortunate that the function clip files are close to being unusable for most people, unless they are really simple, and the "user" takes the time to study the thought process of the creator that led to the actual function network. So in fact just a handful of people are truly skilled enough to really leverage the already shared "effects".
So in my mind, macros are high on the list. ;-)
Regards,
Frank
Hi Frank,
I double this ... My heart starts pounding thinkink of Macros ,-)
Volker
Quote from: Volker Harun on October 23, 2007, 06:48:23 AM
Hi Frank,
I double this ... My heart starts pounding thinkink of Macros ,-)
Volker
I kinda knew you were on my side ;-)
This sounds really wonderful as long is it comes with clear documentation (written by Oshyan :)) and yet I know I've learned so much by tearing into people's clip files and asking questions. I'm clearly not upset by this absence since it meant knowing this app.
Quote from: FrankB on October 23, 2007, 06:25:51 AM
It's unfortunate that the function clip files are close to being unusable for most people, unless they are really simple, and the "user" takes the time to study the thought process of the creator that led to the actual function network. So in fact just a handful of people are truly skilled enough to really leverage the already shared "effects".
So in my mind, macros are high on the list. ;-)
Regards,
Frank
Quote from: calico on October 23, 2007, 09:24:51 AM
This sounds really wonderful as long is it comes with clear documentation (written by Oshyan :)) and yet I know I've learned so much by tearing into people's clip files and asking questions. I'm clearly not upset by this absence since it meant knowing this app.
Yeah... similair like the feeling of endless gratefulness that overcomes you from time to time, when you think about all the useful things you learn about seeding and harvesting corn, just because there's no place to simply buy popcorn from a shelf ;-)
I've found I can do more with an application if I understand it as opposed to knowing some of its working parts. Fortunately, this community has provided this opportunity. :-*
Looking forward to the macros. 8)
Quote from: FrankB on October 23, 2007, 09:31:21 AM
Quote from: calico on October 23, 2007, 09:24:51 AM
This sounds really wonderful as long is it comes with clear documentation (written by Oshyan :)) and yet I know I've learned so much by tearing into people's clip files and asking questions. I'm clearly not upset by this absence since it meant knowing this app.
Yeah... similair like the feeling of endless gratefulness that overcomes you from time to time, when you think about all the useful things you learn about seeding and harvesting corn, just because there's no place to simply buy popcorn from a shelf ;-)
Quote from: calico on October 23, 2007, 10:36:26 AM
I've found I can do more with an application if I understand it as opposed to knowing some of its working parts. Fortunately, this community has provided this opportunity. :-*
all true. I've been kidding.
Cheers,
Frank
Ahhh. Okay. ;D
I haven't been. LOL I don't envy someone coming into this application a year from now when all of this interactivity is not as necessary. I think we're lucky at this stage (especially you and the other Alpha testers, Frank), since we're all searching for the ins and outs. Someone coming in cold next year will need to wade through old conversations, since most of us will be on to other things. I guess. This might be another subject altogether, though...I'm looking at the Subject to this discussion and I don't think I'm on mark. ;D
A lot of the functions I've used involve trying to manipulate the voronoi noise so that's why I put a voronoi fractal at the top of my wants list. I think once Matt gets past the UI, multi core etc then a voronoi fractal should be priority. It's really important for any procedural generator to have this. At present TG2 is essentially a one fractal app. Voronoi fractal is the next most useful after Perlin. You have to consider that TG2 is fundamentally a procedural generator. Blender has more fractals. This is TG2's weakness. I'm finding that this is the limit now with it's capabilities. Look at this render from Mojo and you'll understand why we have to have Voronoi fractal to get all those lovely voronoi octaves:
http://img77.imageshack.us/img77/8081/volcanicsz4.jpg
We can get the surfaces of those rocks in the links from Cyber-Angel but we can not get the variety of rock forms that voronoi can provide.
It is a nice image - but we do have voronis .... it is just a matter of combining several scales and using maybe a warp-shader ,-)
That's a lot of hassle though Volker and certainly not something the average user is going to do. For example a perfect use of a voronoi fractal would be in our tower technique or clamp and mask as it really should be because that's fundamentally what's going on. It's not just for crazy extreme towers. If we had voronoi fractal we could clamp and mask it to get beautiful outcrops of rock.
We do however have the alpine fractal. Anyone tried doing anything interesting with that ;)
The alpine fractal actually has a voronoi basis and if you guys tweak round with that you'll find it pretty versatile but it displaces the wrong way for rock clumps and I see no way of changing this so in essense it should be easy for Matt to make a voronoi fractal that displaces convex voronoi clumps instead of concave. It could be more or less similar to the alpine fractal. Why Matt didn't do this is beyond me.
???
I got around it 8)
I've forced the Alpine fractal to create amazing rock textures. Test renders coming soon.
Okay, efflux. More, more, more. This sounds cool and I want to experiment. What are you doing?
The alpine fractal is slow rendering so this will take some time. It's not ideal but with various tweakings you can get some cool rock effects. I'm not sure how much of the alpine fractal is actually voronoi but it has some other really cool settings anyway.
The alpine fractal appears to use voronoi as it's basic structure then some other function that Matt has come up with to create the details. The voronoi with no warping is horrible and there is no way I can find to truly get rid of the voronoi ridges. However the alpine fractal has some cool effects that when tweaked about just by themselves create superb effects. The secret is with inverting the displacement. You can also merge the alpine with power fractals. We can't make the voronoi into "cracks" rather than ridges but it's still cool. I'll post some images soon.
I don't like those scalloped effects in the alpine fractal but of course Matt did this to get an alpine effect. It could have been a little more versatile in my opinion though.
I edited this message because what I said about the warping is not that simple but fairly easily possible. I didn't do it the way I initially had thought.
Here's my first two tests and it appears with some tweakery you can get cracked rock without that scalloped alpine look and you can simply use this as a basic voronoi to play with and no blue function nodes. I didn't use any. Render times are longer but if you initially tweak the detail out and then bring that back up at render times it's usable.
(http://img520.imageshack.us/img520/1355/alpine10ai5.jpg)
The first one involves giving the fractal a negative displacement value and warping it. The rest I leave to you. The second one is positive displacement, also warped slightly and sent through a transform shader to stretch one dimension. Playing with the deposition values are important to try to avoid that scalloped look.
The alpine fractal is unique but not so cool by itself as a terrain. However when tweaked and combined with perlin you get some great results:
(http://img155.imageshack.us/img155/9448/alpine30qo5.jpg)
This is a more interesting terrain than just perlin alone. There are a lot of flowing lines and cells from the alpine fractal but I like that. Surfaces over the top would improve it. Blending terrains in TG2 causes step like problems. What causes this? It happens every time I blend terrains so much so that I rarely do it. This is no good because the only way to get truly interesting terrains is to blend them. However, this feels like something bigger in scale. It also has an altitude clamp for the cliffs but when merged with an inverted alpine fractal, the cliffs are broken up in a very nice way. Some of the most interesting terrains are ones where there are lots of steps of some sort but blended so the steps get distorted and look natural. Perfectly horizontal stratas and big perfectly stepped ledges don't usually look too good to me.
Unless a macro facility can be edited online, I don't regard it as a help!
Quote from: mogn on October 24, 2007, 01:28:37 AM
Unless a macro facility can be edited online, I don't regard it as a help!
What do you mean, edited online?
I like the first and the third one (of efflux' latest) looking very
promising,gotta give that a try i guess.
efflux, these last two look very promising. Nice work. Is creating this effect too much information to pass on through a tutorial?
I went this angle due to being a bit stuck with limited basis functions and the issues of getting more varied rock forms - partly sparked of by this thread. It seems you can still get quite a bit within TG2's present limits. What you see in the previous images is mostly just using the alpine shader which I never explored much before. If you follow the things I mentioned and just play with that shader you should find similar results. I have messed further with surfaces but as usual I always stray away from real world forms. Maybe in an H R Giger world you would see this.
http://img138.imageshack.us/img138/9633/warp10wb8.jpg
Once again. No blue function nodes.
This is three perlin billows warped and noise stretched in the power fractal's settings, then simply merged. Each has different feature sizes and stretched in different dimensions of X,Y,Z. I'm getting more into just getting good surface details etc now rather than taking TG2 to the extremes of it's technology capability.
You need to play with all these fractal settings. They are very interdependent. You will only find interesting stuff by exploring. No knowledge of how any graphs work is needed for this.
Hehe when I saw he image I had not read your commentary and my thought was, "Well thats a good example of somewhat wild natural forms, now how are you going to re-produce them in Terragen"
You certainly get lava flows and fossilized mud flows just like that.
Richard
This is the cool thing with TG2 and as far as surfacing is concerned, I haven't explored it enough. TG2 gives beautifully natural rich surfaces very easily. In Mojo you could get those forms and actually more spectacularly but it would be very difficult to give them that natural rock form look.
Negative displacement is important for the alpine fractal because it can give you totally different results. Try changing the displacement amplitude to -1000 for the terrain. You'll see what happens and to avoid the terrain ending up 1000m too low you can use displacement offset to bring it back up but I'm experimenting with smaller scale surfaces not so much terrain. With perlin, this negative displacement is already provided for in the fractal. If you negatively displace a perlin billows you will get ridges. The perlin mix basis are combinations of these two forms. Perlin is fast rendering and the most useful basis. If you can do something with perlin then it's best but some things you simply can't do, then voronoi is the next most useful and it's utilized in the alpine fractal. It's also what the fake stones use.
Quote from: efflux on October 23, 2007, 09:40:33 PM
...Blending terrains in TG2 causes step like problems. What causes this? It happens every time I blend terrains so much so that I rarely do it. This is no good because the only way to get truly interesting terrains is to blend them....
By stepping I'm assuming you're referring to the thins sharp drop running across the third image... This main source of this from what I can see would be the nodes used for the mix controller in the merge node. If these jump sharply in value, particularly from negative to positive values you would easily get stepping like this. I got this with some of my terrain masks, and eventually fixed it by clamping the masks to 0-1 once I realised what was going on.
There might be some value in clamping fractals used for mix controllers to positive values?
OK, thanks. Yes, it's the sharp line and it's been giving me problems before. I will look into your ideas but need to check out my actual blending methods used. TG2 is rendering right now. Funnily enough, I have just solve step problems in another area of experimenting on that voronoi thread.
I notice yet again one of my Imageshack images has disappeared here. I have all my images at Imageshack. It makes it easy to see everything I have uploaded at various sites. I've used Imageshack for ages without a single problem but now my images are simply disappearing from Imageshack and then I can't delete images that do not exist. My gallery at Imageshack is basically broken.
Imageshack is now useless. Do not use it.
Has anyone tried File Nanny?
http://www.filenanny.com/index.php
Quote from: Sethren on November 15, 2007, 01:36:12 PM
Has anyone tried File Nanny?
http://www.filenanny.com/index.php
not working for me...
It doesn't work for me either so a bad sign from the start.
Odd, they must be having issues today because normally i use this service quite often.
These services need to be 100% reliable. It seems they are not which means that you go on a thread with a crucial screenshot or something and it isn't there. There is an image missing on this thread so people will not know what hell I'm talking about. It's quite convenient though because for example I was just able to quickly show someone a bunch of TG2 tests etc all gathered at Imageshack and the Planetside system seems a bit slow anyway. I can fix that missing image but I'm waiting to see if Imageshack get back to me because my gallery of images there is simply broken. I can't even delete missing images because it can't find the images to delete.
That is strange. The missing image has suddenly reappeared and my Imageshack gallery is all sorted now. Maybe somebody there actually read my mail.
I agree but in this world nothing is 100% reliable. :-[