The illusive 'Grass Clumps'...

[dhavalmistry]

here are some of the good grass models from turbosquid (ofcourse they are free )

http://www.turbosquid.com/FullPreview/Index.cfm/ID/182767  (.max)

http://www.turbosquid.com/FullPreview/Index.cfm/ID/270766  (.3ds)

[sonshine777]

That clump of yours looks pretty good, guy. Thanks.

There are two color shaders in it to give a more realistic appearance.

[Costaud]

here are some of the good grass models from turbosquid (ofcourse they are free )

http://www.turbosquid.com/FullPreview/Index.cfm/ID/182767  (.max)

http://www.turbosquid.com/FullPreview/Index.cfm/ID/270766  (.3ds)

These are the exact same plants from 3Dplants.com

[Buzzzzz]

Regardless of settings variations, I am unable to render the 'any' grass clumps in the positions represented in the preview window. Any thoughts you would like to share on this subject would be appreciated.

Bob,

I just encountered a population of grass below the terrain because of displacements. I followed what I understood Matt to say and attached the population to the last shader in my surface shader group and the grass was back on top where it should be. I don't know if I needed it but in addition I placed a "Compute Terrain" at the bottom of the shader group before attaching it to the planet. Anyway it worked and I'm happy. If you want some help with your scene or don't follow what I'm saying, just give me a call later. Isn't Verizon Great! We talk Terragen from Arizona to Maryland for free. 

[choronr]

Thank you Jay and all those who responded to this thread; and, to Planetside for making this forum available to us all. Sharing our problems and remedies will help us better understand the intricacies of this wonderful program.

[choronr]

Problem solved; thank you all; Matt and Jay for your input - I now have visible grass clumps by connecting the clumps to the last shader in the group.

[Kevin F]

Hi, I'm still having problems placing a population of bushes on the terrain.I can see the bushes being placed initially and then then terrrain covers them. The terrain is slightly displaced and is stratified and has three sets of fake stones. I've read most of the posts on this and seen Matts advice but I don't seem to be getting anywhere!
Buzzz said:
"I encountered a population of grass below the terrain because of displacements. I followed what I understood Matt to say and attached the population to the last shader in my surface shader group and the grass was back on top where it should be. I don't know if I needed it but in addition I placed a "Compute Terrain" at the bottom of the shader group before attaching it to the planet. Anyway it worked and I'm happy"

Question - how do you attach a population to a shader? What does the output of a population conect to? I can't make it conect to anything - always a red line - no green.  And since my surface shader group is no more than the base layer and fake stones - what can I connect it to?

Help please.
regards
Kevin.

[Buzzzzz]

Hi, I'm still having problems placing a population of bushes on the terrain.I can see the bushes being placed initially and then then terrrain covers them. The terrain is slightly displaced and is stratified and has three sets of fake stones. I've read most of the posts on this and seen Matts advice but I don't seem to be getting anywhere!
Buzzz said:
"I encountered a population of grass below the terrain because of displacements. I followed what I understood Matt to say and attached the population to the last shader in my surface shader group and the grass was back on top where it should be. I don't know if I needed it but in addition I placed a "Compute Terrain" at the bottom of the shader group before attaching it to the planet. Anyway it worked and I'm happy"

Question - how do you attach a population to a shader? What does the output of a population conect to? I can't make it conect to anything - always a red line - no green.  And since my surface shader group is no more than the base layer and fake stones - what can I connect it to?

Help please.
regards
Kevin.

Hi Kevin,

Try this: In your objects settings choose the Terrain Tab, click on the small box next to compute terrain and chose assign shader, then select your last fake stone shader in the shaders group. Try a crop render to see if it helps. Also make sure sit on terrain is selected, which it should be because it's set by Default.
Please let us know if it works.

Good Luck,'Jay

[Kevin F]

Hi Kevin,

Try this: In your objects settings choose the Terrain Tab, click on the small box next to compute terrain and chose assign shader, then select your last fake stone shader in the shaders group. Try a crop render to see if it helps. Also make sure sit on terrain is selected, which it should be because it's set by Default.
Please let us know if it works.

Good Luck,'Jay

[/quote]



Hi 'Jay,
No this doesn't work. I had already tried this and a few other connections! I've attached the .tgd and the .tgo for the bushes.
I've stripped the file down to just the terrain and the bushes so there are no fake stones but the problem is still there.
It seems to be directly related to the displacement of the terrain - take it off and the stata shader off and there's no problem. I suppose the strata shader is actually a very specific displacement. But the thing is I need both these displacements to acheive the particular fake stone shapes I get with them on. (if and when you see them you'll see why )
Why does the shape of my fake stones vary with displacement switched on or strata switched on? I thought the stones were placed on the displaced terrain These are obviously placed on the terrain before displacement and then altered accordingly by the displacement!
In just what order is the file rendered/processed?
Any help appreciated.
Regards
Kevin.

[Buzzzzz]

Hmmm? I would be happy to look at your file but I can't seem to find it?

[Kevin F]

Hmmm? I would be happy to look at your file but I can't seem to find it?

DOH! forgot to attach.
The .tgo is bush 1 from:

http://www.balaskas.net/files/CB_Desert_Bush_Pack.zip

Regards
Kevin

[Oshyan]

Stones are applied as a displacement effect over existing terrain. Any displacement you apply afterward should affect them.

I took a look at your .tgd and there are a couple things going on. First off you have your populator set to a position of -100 on the Y axis placing it below the terrain. You may have done this intentionally to compensate for other issues but since it is an absolute offset it will seldom get you the results you want if sitting on terrain is the problem you're trying to resolve. So first thing is to correct that by setting Y to 0.

Second, populations tend to start having problems with negative altitudes in terrain. For some reason much of your terrain is at a negative Y, which is odd because there's no negative offset or anything. I think it's due in part to your relatively high noise variation in your base fractal terrain. It's not necessary to change this but it helps to be aware of this problem at least.

Finally, populations also sometimes have issue with properly sitting on strata displaced terrain. In this case in particular you have some pretty extreme and perhaps inappopriate settings given the actual altitudes of the terrain you're working with. I understand it may give the results you are trying to achieve, but in terms of considering the absolute settings they don't make much sense in the context of your terrain.

In any case there are generally a couple things to try when solving problems like this.

Already mentioned was changing the terrain which the populator points at to be the last node that provides displacement. Generally you will only want to do significant displacement in the Terrain group so that the Compute Terrain node can do its job of providing a computed normal. In this case your major displacement is done appropriately.

Another approach is to add a Compute Terrain/Normal into the shader chain just after your last major displacement and then set your population to point to this node for its terrain key. This approach doesn't always work either and it will also increase render time.

A final possibility, and one which I think works in this case, is to look at the granularity of the normal being used to compute object placement for your scene and consider whether some finer-scale displacement really needs to be accounted for. Sometimes the problem is that the populator is basing its object height on isolated sample height values that place it inappropriately below or above an otherwise fairly uniform terrain. This can happen when small displacement is sampled at a small level of granularity. You can control the scale that the normal is calculated at with the Gradient Patch Size setting in the Compute Terrain/Normal node. In this case setting it to 100 (meters) seems to fix the problem, suggesting that less fine evaluation of the surface is in order. But it's something to play around with regardless.

I've attached a .tgd that appears to fix the problem on my end. Please verify.

- Oshyan

[Buzzzzz]

Great! I come back and Oshyan has a fix! You Da Man! And thanks for the tip on the Gradient Patch Size. 

[Kevin F]

Thanks for the reply Oshyan and all the detailed information BUT.....
applying the changes in your fixed file to my original file still doesn't work even with the population Y value set to 0.
Your reply raises a number of issues:
First fake stones - if they are applied as a displacement effect over existing terrain, why do I get totally different results when the displacement in the terrain is turned off/on and when the strata layer is turned off/on? i.e. in what order are effects carried out in TG2?
because I didn't apply any displacement "afterward" - did I? I though the stones were place on the displaced terrain?

Second, the populator was set to -100 on the Y axis because the terrain was randomly generated and slightly displaced to a mainly negative level completely by chance. I generated the terrain, found the pov I wanted and set the populator to a Y value "above" the average for the ground in view, which was -145 meters.
If this is because   "..... it's due in part to your relatively high noise variation in your base fractal terrain. It's not necessary to change this but it helps to be aware of this problem at least."  Is a noise variation of 2 relatively high? relative to what? and why is it a problem?

Third, I totally agree with your comments on my strata settings, but to be honest I was simply experimenting with different values. It just so happens that with these extreme values I produced fake stone that were amazing! (see sample pic).

Finally, as for "to look at the granularity of the normal being used to compute object placement for your scene and consider whether some finer-scale displacement really needs to be accounted for".
I don't understand any of this!

It's obvious from most of this that proper documentation and instruction is needed for TG2 urgently. Your extra and valuable information is appreciated but not available anywhere else!

TG2 is still in danger of being regarded as a hit & miss piece of software, with a lot of nice somtimes superb yet random effects being produced by people who don't really understand how they are acheiving them. Look at the stones from my original render for example - I've no idea really how I created them and consequently don't know how to modify or control them. This all gives fuel to the "terragen is not art" lobby. Sorry for going on for so long.
Regards and thanks for your help.
Kevin.

[sonshine777]

@Kevin F - TG2 is still in danger of being regarded as a hit & miss piece of software, with a lot of nice somtimes superb yet random effects being produced by people who don't really understand how they are acheiving them. Look at the stones from my original render for example - I've no idea really how I created them and consequently don't know how to modify or control them. This all gives fuel to the "terragen is not art" lobby. Sorry for going on for so long.
Regards and thanks for your help.
Kevin.

It's still a "TECH PREVIEW" that has only had one patch since it was first opened up to us. Give them a break.

Also I have seen plenty of cases were results are completely repeatable and controllable. (leaving RANT mode)