PDA

View Full Version : SG_Fertilizer2



gerardstrada
06-04-2009, 05:01 PM
Hello,

Just to let you know Sebastian Goetsch has released a new version of the SG_Fertilizer:

SG_Fertilizer2:
http://www2.informatik.hu-berlin.de/~goetsch/Fertilizer2/

In addition to the previous features, we now have the option to define a secondary growth process (using a second weightmap). Besides there's a 64-bits version available!

Many thanks to him for this new version!

Please, feel free to comment and spread the word :thumbsup:



Gerardo

adhesiveX
06-04-2009, 05:15 PM
Thank you Sebastian.

papou
06-04-2009, 05:22 PM
Thank you very much Sebastian, and thx Gerardo for the news.:thumbsup:

h2oStudios
06-04-2009, 05:27 PM
Damn, another toy to play with;). I'll be sure to check this out. And as adhesiveX says, thanks, Sebastian Goetsch, and all the other cool people that keep us Lwavers in good supply as being an artist with basically no know-how creating tools for myself it's always a treat when someone drops something cool on my lap.

EDIT: Whoops, thanks Gerardo.:thumbsup: For the look out, for I don't search plugs enough as I should.

biliousfrog
06-04-2009, 06:07 PM
WOOOOOOHOOOOOO!!!!

Fertilizer is awsome and I've been waiting for the 64bit version for ages.

Thanks Sebastian, if you're reading this I promised to send you some money to help with development ages ago...it's a long story but I didn't have the funds and then forgot all about it but I promise to give you something in return for this great plugin...LightWave should have this tool natively.

Dodgy
06-04-2009, 06:31 PM
Hmmm, as soon as I select a weightmap, it crashes LW. This is the 32 bit version.

gerardstrada
06-04-2009, 07:12 PM
Tested here with LW 9.6 (32 and 64-bits) without crashes. Can you share more info, please? what LW version are you using? Win or Vista? subpatch object? subd order? first or second weightmap? Better if you can post a simple scene reproducing the error.



Gerardo

Auger
06-04-2009, 08:05 PM
Excellent! Thanks to Sebastian.

JMCarrigan
06-04-2009, 08:22 PM
I don't know why it does what it does. I tried one thing (attached) and then changed the 2nd weight map to just having a value rather than an envelope. The result is also attached.

Any idea what's up?

Dodgy
06-05-2009, 07:03 PM
Here's the file I've been trying it with. Load the object, apply fertilizer, choose the Falloff weight map, crash.

This is on 32 bit win.

gerardstrada
06-05-2009, 09:38 PM
It works fine on 64-bits, but it crashes in 32-bits as you say. Problem is fixed when we delete the skelegons in Modeler. Until a response from Sebastian (and if for some reason you are not able to use the 64bits version), better make an object copy, convert skelegons into bones in Layout, save the scene and delete the skelegons before applying SG_Fertilizer2.



Gerardo

Dodgy
06-05-2009, 10:32 PM
There's a new version with this fixed :)

gerardstrada
06-05-2009, 10:40 PM
Yes, Sebastian just fix it while I was writting the previous message. That's speed! :D



Gerardo

flakester
06-06-2009, 11:00 AM
Awesome stuff, a very big !w00t! to you both, gerardstrada and Sebastian Goetsch. :thumbsup:

Many thanks

--
flakester

mav3rick
06-06-2009, 02:17 PM
amazing great exciting MMMMMonsterkilll

3DGFXStudios
06-06-2009, 04:09 PM
Cool > ........... Thank you very much!!! :D

gerardstrada
06-06-2009, 07:26 PM
Awesome stuff, a very big !w00t! to you both, gerardstrada and Sebastian Goetsch. :thumbsup:

Many thanks

--
flakester

Thanks Flakester, but I have nothing to do with this. I'm just posting the news here. All our gratitude for so useful tools goes to Sebastian Goetsch :beerchug:



Gerardo

gerardstrada
06-15-2009, 06:33 PM
Just to let you know Sebastian has fixed a bug that led to distort a geometry after loading objects with several plugin instances applied:

http://www2.informatik.hu-berlin.de/~goetsch/Fertilizer2/



Gerardo

JeffrySG
06-20-2009, 06:13 AM
Does anyone happen to know if there is a MacUB version in the works?

zapper1998
06-20-2009, 08:50 AM
wow

thank you

gerardstrada
06-20-2009, 04:29 PM
Does anyone happen to know if there is a MacUB version in the works?

That's being discussed in this thread:

http://www.newtek.com/forums/showthread.php?p=895049#post895049



Gerardo

JeffrySG
06-20-2009, 08:31 PM
^thx for the link! :D

Exception
06-21-2009, 04:36 AM
Nice!
Looks like a great plugin.

I had a crash copying and pasting the plugin from one object to another though... and when containing some illegal geometry (1 point polygons, etc).

gerardstrada
06-21-2009, 08:07 PM
I don't get a crash here copying and pasting. Perhaps it could be due to another reason? What LW version are you using (9.x)? 32b or 64b? It's better if we can post a simple scene reproducing an error.



Gerardo

Exception
06-22-2009, 08:16 AM
's been fixed by Sebastian. This is an awesome plugin!

Here's a quick tree growth mockup. Might look convincing with some tweaking (and perhaps a better model :) )

For now it's difficult to just use *any* model because geomtery needs to be physically attached to each other. A little script plugin which would connect 'floating' geomtrey to the closest point of a certain defined center object would help greatly with this. It's the reason why I can't use more complex tree models. They're never single piece branch systems. Using boolean 'add' or 'union' only gets you so far, as on complex models it becomes almost unfeasible.

biliousfrog
06-22-2009, 09:21 AM
I used the original fertilizer plugin with Trees Designer to create some veins, lots of fun. Also works great with the free ivy generator.

gerardstrada
06-22-2009, 04:06 PM
Yes, bug fixed for some serious degenerate polygons:

http://www2.informatik.hu-berlin.de/~goetsch/Fertilizer2/

I agree, it also works great with Trees Designer and Ivy generator. For more realistic results, it's better if it's a continuous geometry - though it can be several of them in cases of veins and roots. For discontinuous geometry (branches separated in groups), we need a Wmap for each group.



Gerardo

Exception
06-23-2009, 02:09 AM
I don't think a wmap for each group is a good solution. It's still a lot of work, and has many limitations.
Treesdesigner is nice, but not as good as OnyxTrees for example (which creates discontinuous geometry). A simple script that makes 2 point polygons between separated branches and a main stem would be a lot easier to manage.

gerardstrada
06-23-2009, 02:26 AM
What many limitations? Why a lot of work?
Though for discontinuous geometry we might need a totally different approach (let's read: new plug-in), I think it's a very good solution, at least for continuous geometry :) mainly because continuous geometry is more realistic and it works for close-ups. It's not a lot of work also because if you have noticed, we don't need to set up a ramp for the Wmap through the entire geometry, we simply apply a 100% value for the whole (continuous) geometry and 0% just for the beginning of the growth. That's all :)



Gerardo

Exception
06-23-2009, 03:59 AM
What many limitations? Why a lot of work?

Err... have you tried WMapping 500+ discontinuous branches? And that would be for a low-res tree; try something with the stems included.

I never said the entire plugin wasn't a good solution. Obviously I meant the individual mapping of discontinuous parts as a suggestion to not need a script. I think the plugin is great.

Also, while continuous geometry might be more realistic in the theoretical sense, in reality, the tree generators that out there that produce continuous geometry (mainly TreesDesigner) are, in general, less well developed than those that produce discontinuous geometry (Xfrog, Onyx). The close-up argument holds very little ground, if you look close up at a high-res output of Onyx tree.

A very simple light-weight solution would be to have a little script as I suggest. It would also help for many other applications that are not trees.

gerardstrada
06-23-2009, 04:59 AM
Err... have you tried WMapping 500+ discontinuous branches? And that would be for a low-res tree; try something with the stems included.

Of course not! But I was talking about continuous geometry.


I never said the entire plugin wasn't a good solution.
Nope. You was talking specifically about weightmapping each group. But it seems the group you were talking has a discontinuous geometry while I was talking about continuous geometry (which is the work scenario for this plugin right now).


Also, while continuous geometry might be more realistic in the theoretical sense, in reality, the tree generators that out there that produce continuous geometry (mainly TreesDesigner) are, in general, less well developed than those that produce discontinuous geometry (Xfrog, Onyx).
Maybe, but have you tried Ivy generator? it's pretty good and available for Win, MACOSX and Linux.. (Tree Designer and Leaves Generator are also very good, though don't know if they are still available).


The close-up argument holds very little ground, if you look close up at a high-res output of Onyx tree

I've seen it and it could work depending on the tree maybe, but we'd need a lot of res and I don't think a discontinuous geometry could be used by any VFX house for a close-up in a movie or TVspot if we really need a realistic result. i.e: in the last HDRI3D magazine, Javier 'Netvudu' Meroņo wrote an article about a VFX shot (a close-up) for the Estigmas movie, they consider to use Fertilizer (but pitifully there wasn't available a 64-bit version in that time), butr, guess what? they used continuous geometry for their roots models, why? more realism, I can guess.


A very simple light-weight solution would be to have a little script as I suggest. It would also help for many other applications that are not trees.


Yes, it's a good idea for those cases :) Let's ask to Sebastian. In the meantime, other solution might be to have a 2point-poly scheme of the three (continuous geometry) so that apply the SG_Fertilizer, bake the deformation and apply it later with something like FX_Metalink perhaps.



Gerardo

Exception
06-23-2009, 09:19 AM
Nope. You was talking specifically about weightmapping each group. But it seems the group you were talking has a discontinuous geometry while I was talking about continuous geometry

My whole point concerned discontinuous geometry from the start.


Maybe, but have you tried Ivy generator? it's pretty good and available for Win, MACOSX and Linux.. (Tree Designer and Leaves Generator are also very good, though don't know if they are still available).

Ivy Generator is nice... for Ivy. I already mentioned Treesdesigner: it's nice but nearly not good enough. Onyx and Xfrog are the unbeaten industry standard for plants and trees. Vue plants certainly aren't either. Speedtree ones are continuous it seems, but hardly realistic (game models), and only for 3ds Max. Lenne3D are repackaged Xfrog plants, and Bionatics is not affordable for anyone. Take a look at them. Before any other solution has as many library items as Xfrog/Onyx have, we're several years and a lot of money down the line. If you're talking professional, realistic trees, you're talking Xfrog/Onyx.


I've seen it and it could work depending on the tree maybe, but we'd need a lot of res and I don't think a discontinuous geometry could be used by any VFX house for a close-up in a movie or TVspot if we really need a realistic result.

That's just nonsense. Xfrog is not 'realistic enough'? And Treesdesigner is? I don't think you've used these tools?


butr, guess what? they used continuous geometry for their roots models, why? more realism, I can guess.

Logic fails here. There could be 1000 reasons why their vegetation had continuous models (SpeedTree being the number 1 reason for it, I bet). "Realism" has nothing to do with it. And I'm sure their leaves were not continuous, and I doubt the stems were.

To cut this silly thing short, and to repeat my first post:
Discontinuous geometry is important, it would be great if we had a quick solution to make this work in Fertilizer2, as currenty it requires a lot of work, and a simple script could do the job.

biliousfrog
06-23-2009, 10:03 AM
VMapping an X-Frog/Vue/Onyx tree is quite straight forward, even with thousands of branches/leaves. Load up the UV window, select the UV's for the branches and apply a VMap. The downside is that you'd need to offset the growth from top to bottom unless you want the trunk to grow first. It would take a little trial and error but wouldn't take very long to actually create.

Exception
06-23-2009, 11:52 AM
VMapping an X-Frog/Vue/Onyx tree is quite straight forward, even with thousands of branches/leaves. Load up the UV window, select the UV's for the branches and apply a VMap. The downside is that you'd need to offset the growth from top to bottom unless you want the trunk to grow first. It would take a little trial and error but wouldn't take very long to actually create.

I'm not sure if you tried, but anything that is not connected to the starting object (the one with a weight point set to 0%) will never appear.

If you do a reverse growth, you'd have to manually pick a point on each disconnected piece and set it to 0%. This translates in a massive amount of work for a complex tree.

It's not the application of the Vmap that is the problem. It's the fact that each piece needs to be treated separately with either a separate weight map setting that is in harmony with the growth pattern of the rest of the tree, or manually connect each piece with a 2 point polygon.

biliousfrog
06-23-2009, 12:59 PM
It's been a while since I used it but I thought that you could assign more than one instance of fertilizer using different maps?...so you could have one for the trunk, one for the branches and one for the leaves?

As I mentioned, it wouldn't be as organic as growing from a single point but it is possible.

gerardstrada
06-23-2009, 08:08 PM
That's just nonsense. Xfrog is not 'realistic enough'? And Treesdesigner is?

At least Polas gallery looks better.


I don't think you've used these tools?
Some time ago I have to make a close-up for a full CG TV-spot where some ants walked over some branches. We tried some Xfrog's and Vue's trees, and we ended up modeling the whole thing with MagicBevel by hand because it was the only way to get a realistic result. Continuous geometry of course :)



Logic fails here. There could be 1000 reasons why their vegetation had continuous models (SpeedTree being the number 1 reason for it, I bet). "Realism" has nothing to do with it. And I'm sure their leaves were not continuous, and I doubt the stems were.
It wasn't a whole tree, only roots. They used the L-Systems from Houdini. He mention the L-System of Xfrog too, but Houdini provides them the proper tools to design the animation of the evolution of - continuous - geometry. But a similar thing could have been done with a 64-bits of SG_Fertilizer, I guess.


To cut this silly thing short, and to repeat my first post:
Discontinuous geometry is important, it would be great if we had a quick solution to make this work in Fertilizer2, as currenty it requires a lot of work, and a simple script could do the job.

Maybe. You could send an example to Sebastian about what you pretend.



Gerardo

Exception
06-24-2009, 02:51 AM
As I mentioned, it wouldn't be as organic as growing from a single point but it is possible.

Possible... and a lot of work :) Especially if the branches started at different points, and had many forks into other (diconnected) branches.


At least Polas gallery looks better.

You're joking! you must!

randomd Xfrog gallery image:
http://www.xfrogdownloads.com/greenwebNew/gallery/images/Landscapes/landscape3.jpg

Random Treesdesigner gallery image:

http://www.polas.net/trees/td15/tree03_spheres.jpg


Some time ago I have to make a close-up for a full CG TV-spot where some ants walked over some branches. We tried some Xfrog's and Vue's trees, and we ended up modeling the whole thing with MagicBevel by hand because it was the only way to get a realistic result. Continuous geometry of course

and... modeled by hand for your specific purpose. So not quite using any of the aforementioned tools, so not actually an application which we've been talking about at all. The L-Systems from houdini is also not an actual full blown tree generator.

In any case, this discussion has become useless. Good luck using continuous geometry when making a realistic forest... especially when you'd like to use some tree species that actually exist, like, eh... a european hackberry, a silver birch or a kentucky coffee tree.

gerardstrada
06-24-2009, 04:03 AM
You're joking! you must!
Not kidding :D Don't be fooled by the lighting&shading. At that distance I only see lines in the Xfrog model, but very natural connections in TreesDesigner geometry, and we are talking about close-ups, right?



and... modeled by hand for your specific purpose. So not quite using any of the aforementioned tools, so not actually an application which we've been talking about at all.
We need to use them to finally know if they were useful or not. And we don't use them because, at least in that time, discontinuous geometry looked something like this:

http://imagic.ddgenvivo.tv/forums/b1.png

While continuous geometry looked something like this:

http://imagic.ddgenvivo.tv/forums/b2.png

Besides, leaves weren't connected in a natural way and their geometry was intersected too.


In any case, this discussion has become useless. Good luck using continuous geometry when making a realistic forest... especially when you'd like to use some tree species that actually exist, like, eh... a european hackberry, a silver birch or a kentucky coffee tree.
For a realistic forest, I don't care if they are continuous or discontinuous since they gonna be shown in PGs where nobody notice it :) We don't have to solve each take in the same way, you know?



Gerardo

Exception
06-24-2009, 11:38 AM
Not kidding :D Don't be fooled by the lighting&shading. At that distance I only see lines in the Xfrog model, but very natural connections in TreesDesigner geometry, and we are talking about close-ups, right?

Lighting & shading? Fooled? Have you ever looked at a real tree? They look nothing like Treesdesigner trees. Treesdesigner trees look like they're from the land of Nod, or out of some fantasy rendition of Tolkien's stories. I think it would take a massive amount of effort to attempt anything like a realistic tree species in Treesdesigner. So what, you have continuous geometry, but totally awkward tree shapes. Completely does not make sense to me. Sorry, Gerardo, I think you're missing the entire point.

And we're not talking about close ups. You are talking about them. I'm talking about how to grow realistic trees with Fertilizer, and what would be the best way to deal with (the apparent) problem of discontinuous geometry. Onyx and Xfrog are the only ones that make realistic, species correct, indexed and library-documented trees. They both output discontinuous geometry, so we're going to have to live with that.


And we don't use them because, at least in that time, discontinuous geometry looked something like this:

Why are you explaining the difference between continuous and discontinuous geometry? you think I don't know the difference?


For a realistic forest, I don't care if they are continuous or discontinuous since they gonna be shown in PGs where nobody notice it :) We don't have to solve each take in the same way, you know?

We were talking about growing trees using fertilizer, not any other method, last time I checked. This includes growing a realistic forest, with realisitic trees if one wants.

gerardstrada
06-24-2009, 02:19 PM
Lighting & shading? Fooled? Have you ever looked at a real tree? They look nothing like Treesdesigner trees. Treesdesigner trees look like they're from the land of Nod, or out of some fantasy rendition of Tolkien's stories. I think it would take a massive amount of effort to attempt anything like a realistic tree species in Treesdesigner. So what, you have continuous geometry, but totally awkward tree shapes. Completely does not make sense to me. Sorry, Gerardo, I think you're missing the entire point.
Have you tried TreesDesigner? We have the tools to get any shape we need with a realistic finishing. That's why it's called TreesD e s i g n e r. And that's the entire point, realistic trees that can work with SG_Fertilizer. The style is up to you.


And we're not talking about close ups. You are talking about them. I'm talking about how to grow realistic trees with Fertilizer, and what would be the best way to deal with (the apparent) problem of discontinuous geometry. Onyx and Xfrog are the only ones that make realistic, species correct, indexed and library-documented trees. They both output discontinuous geometry, so we're going to have to live with that.
Yes, I'm talking about close-ups because is the extreme case for a realistic geometry. But those discontinuous geometry can be noticed even in EPs depending on the species.


Why are you explaining the difference between continuous and discontinuous geometry? you think I don't know the difference?
Because those sharp connections aren't realistic AT ALL. But it's good to know that you already know that; so I won't bother you with this again :D



We were talking about growing trees using fertilizer, not any other method, last time I checked. This includes growing a realistic forest, with realisitic trees if one wants.
In that case, you could try two things at this moment: Or you switch to an app that outputs continuous geometry. Or, if you don't think you could get realistic results with SpeedTree:
http://www.speedtree.com/trees/images_4.2/5_Shot_Detail%202.jpg

or TreesDesigner:

http://www.polas.net/trees/td15/tree08_planes.jpg

or even Ivy generator

http://www.ewers-fx.de/P002_Paint_afterbeforeafter.jpg
http://graphics.uni-konstanz.de/~luft/ivy_generator/images/Mark_Dunakin_01.jpg

Or any other app that generates continuous geometry for this kind of models...

then, you could try to check if some of the apps you like, can export a polychain scheme (or curves) of the trees you want to make growing. Talking more seriously, you can use this as a proxy object for the deformation with SG_Fertilizer.

I've tried a quick test here with a crappy discontinuous model (those that you like):

http://imagic.ddgenvivo.tv/forums/DCGtest1.gif

I just exported a low-res version of the tree and merge its points gradually until I end up with a very simplified version (very degenerated geometry). I suggested you the rest of the procedure in one of my previous messages. The ends of the branches are messed up because it's not an accurate polychain (it's not even 2pointpolys in the entire geometry because I made the merging process in a couple of minutes), but I guess a proper polychain should work.

I was wondering if SpeedBoolean could help here. If you have a low-res version of the tree, you could use SBUnion and MergePoints, every time that polygons get connected. You may end up with a damaged model, but if all polygons are connected in some way, you may run SG_Fertilizer over it and then mdd_scan it and use something like DP_MDD_MetaPointer to translate the growth animation. The other way could be a script you talked about, which connects the crossed polygons with 2pointpolys or whatever. Find the edges of the crossed polygons can be made with something like Edge Detect, but the thing is what is the criteria so that these vertices could be connected with the middle of a polygon if we don't intersect them. Maybe there's is already a script like that out there. The other solution would be a totally different approach from Sebastian.



Gerardo

Exception
06-25-2009, 07:14 AM
Have you tried TreesDesigner? We have the tools to get any shape we need with a realistic finishing. That's why it's called TreesD e s i g n e r. And that's the entire point, realistic trees that can work with SG_Fertilizer. The style is up to you.

Yes I have. You miss the point I've been explaining so very often... go design a Silver maple or a western australian eucalypt... go do it, see how long it takes you. It'll take you days, of not weeks. Why? Because it's not an actual botany simulation, it's just a geometrical design tool that has nothing to do with botany in the first place. And then another reason is that if you have a huge library of species-correct paramters lying around, why would you even go though that hell of designing each and every species by hand?



Yes, I'm talking about close-ups because is the extreme case for a realistic geometry. But those discontinuous geometry can be noticed even in EPs depending on the species....

Because those sharp connections aren't realistic AT ALL. But it's good to know that you already know that; so I won't bother you with this again :D

Complete nonsense. Just because one part of a branch isn't physically connected to another doesn't mean it doesn't look good, or that it's a 'sharp connection'. You clearly have never worked with Onyx or Xfrog.



In that case, you could try two things at this moment: Or you switch to an app that outputs continuous geometry.

Sigh... I just gave you a whole list of tree programs and all their advantages and disadvantages, yet you persist in stating the stupid solution: Switch to another program.
Ok, you tell me a program that makes realistic trees, makes continuous geometry, has a species-correct database of parameters available, and is affordable, and I'll buy it.


Or, if you don't think you could get realistic results with SpeedTree:


Er... do you think that tree could actually exist in the wild? Look carefully at those branches... it's totally fantasy.
And...
Does it work outside of 3ds MAX? Nope.
Does it have a species library? Nope.
What species was that tree? *fantasy*
How useful is that fantasy tree program if you need to have a small forest of Elms, Oaks, some proper existing types of dissiduous underleafs and a set of evergreens? That's right: not very useful at all... is it?


or TreesDesigner:
http://www.polas.net/trees/td15/tree08_planes.jpg


My goodness. That's not a realistic tree! It's some kind of Elven forest elder tree... that can speak and sing songs. And if you think it's realisitc, please give me the geonome of that tree. I'm sure it lives somewhere in the northern regions of an obscure patch of tropical rainforest. On the moon.


or even Ivy generator

Yeah... nice. If you like sick trees that have vines all over them... or more odd miscellaneous non-existent species all made by hand by someone with too much time on their hands.



I've tried a quick test here with a crappy discontinuous model (those that you like)

Ok, you missed the point entirely, and refuse to concede in any way. Now I 'like' crappy models?
Discussion seems to be useless at this point.

Nice suggestion for a workflow, but it looks seriously funny and it's still quite a lot of work. Imagine doing that for 100 different trees. No thanks.

gerardstrada
06-25-2009, 12:52 PM
Yes I have. You miss the point I've been explaining so very often... go design a Silver maple or a western australian eucalypt... go do it, see how long it takes you. It'll take you days, of not weeks. Why? Because it's not an actual botany simulation, it's just a geometrical design tool that has nothing to do with botany in the first place. And then another reason is that if you have a huge library of species-correct paramters lying around, why would you even go though that hell of designing each and every species by hand?
HAHA! Did you mean weeks for a process like this???

http://www.polas.net/trees/exampl/tree4.jpg

C'mon! It seems you don't really have never used it! Did you know you have presets, right? TreeDesigner creates far better (continuously meshed) branches than XFrog, and it also does a better job in placing the leaves and keeps polycount far below 100K for really big and realistic meshes.


Complete nonsense. Just because one part of a branch isn't physically connected to another doesn't mean it doesn't look good, or that it's a 'sharp connection'. You clearly have never worked with Onyx or Xfrog.
Do you want to see my example again? :D Again, the biggest drawback in the apps you mention is that both just "stick" the childbranches in the mainbranch: so you never get a realistic mesh. it could look Ok in PGs but in nearer takes, its mesh is not realistic AT ALL.


Sigh... I just gave you a whole list of tree programs and all their advantages and disadvantages, yet you persist in stating the stupid solution: Switch to another program.
What it's stupid is to try to get stuck to incompatible solutions.


Ok, you tell me a program that makes realistic trees, makes continuous geometry, has a species-correct database of parameters available, and is affordable, and I'll buy it.
For continuous and realistic geometry you have TressDesigner (you have presets, you know?). The free option is Ivy. But they don't like you. The other option that I know that has species-correct database of parameters available and including the possibility to make grow trees (continuous geometry) is for Maya, sorry, NatFX. You can also buy the models already made with it in their website or customized models as a third-party service. Don't know what you can afford or not but since you don't care about discontinuous geometry and since you want to make grow an entire forest, why not Vue?


Er... do you think that tree could actually exist in the wild? Look carefully at those branches... it's totally fantasy.
And...
Does it work outside of 3ds MAX? Nope.
Does it have a species library? Nope.
What species was that tree? *fantasy*
How useful is that fantasy tree program if you need to have a small forest of Elms, Oaks, some proper existing types of dissiduous underleafs and a set of evergreens? That's right: not very useful at all... is it?
My goodness. That's not a realistic tree! It's some kind of Elven forest elder tree... that can speak and sing songs. And if you think it's realisitc, please give me the geonome of that tree. I'm sure it lives somewhere in the northern regions of an obscure patch of tropical rainforest. On the moon.
Yeah... nice. If you like sick trees that have vines all over them... or more odd miscellaneous non-existent species all made by hand by someone with too much time on their hands.
:sleeping: Again, style is up to you. But it's clear that if it's not from Onyx or Greenworks, you don't want it. In that case, as a user of those packages, what you should demand to those companies is continuous meshes for their models. Not for working with SG_Filter or any other nodal solution, but just for a matter of realism and workflow. I've seen those discontinuous meshes in Xfrog for years, and they have never changed that.


Ok, you missed the point entirely, and refuse to concede in any way. Now I 'like' crappy models?
Discussion seems to be useless at this point.
Don't be so susceptible, Tom. I was referring to those discontinuous models you like. With 'crappy' I was referring to my model.


Nice suggestion for a workflow, but it looks seriously funny and it's still quite a lot of work. Imagine doing that for 100 different trees. No thanks.
Of course not. That's because I used the old DTRee from DR, which doens't export a guide curve with the structure of the tree (however it shows it in the preview).

But in order to do it quickly and get the proper results, idea is to use a guide curve exported by some of your wooonderful apps (because, since they don't generate a continuous mesh, at least they can do that now, right?). With that guide (converted to 2point-polys) you could be able to use SG_Fertilizer with nice results.



Anyway, I think you are doing wrong with sticking to a single solution for such a complex effect. Because even if you solve the discontinuous problem of your models, you need to make grow at least 500 models. You could try to solve that with TexturedFilter in GE (a procedural texture for a random growth), but you'll need many trees in a single layer. In that case, I'd try a mix between SG_Fertilizer, MDD_Scan an Displacement Node Editor. There, is more easier to get a random growth, even for individual objects. For PGs I'd try also with DP_PolyMove or even PartMove, and maybe you could even use your beloved discontinuous meshes :)



Gerardo

Exception
06-25-2009, 04:00 PM
HAHA! Did you mean weeks for a process like this???

Err... nop, but that's not an actual existing tree species, now is it?


C'mon! It seems you don't really have never used it! Did you know you have presets, right? TreeDesigner creates far better (continuously meshed) branches than XFrog, and it also does a better job in placing the leaves and keeps polycount far below 100K for really big and realistic meshes.

I have used it a-plenty. And it's great. Until you need a REAL tree. Until someone says to you": give me a garden with Canadian Chestnuts. then what? Presets are not going to help is there's not 'Canadian chestnut' preset, now is it?



Again, the biggest drawback in the apps you mention is that both just "stick" the childbranches in the mainbranch: so you never get a realistic mesh. it could look Ok in PGs but in nearer takes, its mesh is not realistic AT ALL.

They work fine for everything but having a camera stuck straight inopt a single branch, and that is a super-rare ocurrence worth modeling the branch by hand.


What it's stupid is to try to get stuck to incompatible solutions.

Incompatible with WHAT? Your thinking methods?



For continuous and realistic geometry you have TressDesigner (you have presets, you know?). The free option is Ivy. But they don't like you. The other option that I know that has species-correct database of parameters available and including the possibility to make grow trees (continuous geometry) is for Maya, sorry, NatFX. You can also buy the models already made with it in their website or customized models as a third-party service. Don't know what you can afford or not but since you don't care about discontinuous geometry and since you want to make grow an entire forest, why not Vue?

Why Vue? It's also discontinuous! And bad qualiy. And limited library. And limited export... and I already own it. NatFX is not always continuous either, neither are any of the others... And presets, like I said, are useless unless you actually have some. And why buy Maya and NatFX if you can do it right there in LW with tools you already have? I don't get your argument. It makes no sense.


Again, style is up to you. But it's clear that if it's not from Onyx or Greenworks, you don't want it.

Gerardo, there's no arguing with you. You give suggestions that don't solve the problem. I clearly state the requirements and you come with things that clearly break them. What kind of dicussion does that lead to but some sheer nonsense. Now I'm a Xfrog / Greenworks fan that doesn't want anything else? I just suggested you to name a single app that could make continuous geometry with LW and had a botanic database. You couldn't name any. So, what is there left to say?


I've seen those discontinuous meshes in Xfrog for years, and they have never changed that.

Because they work fine for many purposes. I guess most Xfrog & Onyx users don't habitully stick their camera into a branch connection. And even if they did, if you export with high quality settings, it's hard to see the difference.

Now that I've been destined to like 'crappy models', be a fanboy of Xfrog and Greenworks and rigid in my thinking for wanting a botanical database without having to buy maya or another 3D application, I sign off on this hilarious tantrum fest, and go do more useful stuff. Have fun there, Gerardo, and go try to make a realistic asian cherry tree with Treesdesigner.

gerardstrada
06-25-2009, 05:29 PM
Err... nop, but that's not an actual existing tree species, now is it?
I have used it a-plenty. And it's great. Until you need a REAL tree. Until someone says to you": give me a garden with Canadian Chestnuts. then what? Presets are not going to help is there's not 'Canadian chestnut' preset, now is it?
I'd begin to tweak that preset for a Canadian Chestnuts. It's easy to get some variations with TreesDesigner tools.


They work fine for everything but having a camera stuck straight inopt a single branch, and that is a super-rare ocurrence worth modeling the branch by hand.
It works for GPs for sure, but depending on the specie, you can see those discontinuous meshes even in EPs.



Incompatible with WHAT? Your thinking methods?
Not, your own tools!


Why Vue? It's also discontinuous! And bad qualiy. And limited library. And limited export... and I already own it. NatFX is not always continuous either, neither are any of the others... And presets, like I said, are useless unless you actually have some. And why buy Maya and NatFX if you can do it right there in LW with tools you already have? I don't get your argument. It makes no sense.

Yep, they are also discontinuous, but that's precisely what you like, so what's the problem if it already has the growing capabilities you are looking for. If you get a bad quality with it, it's not because the software C'mon! it has been used for matte paintings by ILM! (GPs).

http://mirror-eu.e-onsoftware.com/products/solutions/images/ilm2.jpg

http://www.e-onsoftware.com/products/solutions/images/ilm3.jpg

Just in case you are interested, do take a look at this technique: http://www.asilefx.net/catalog/product_info.php?products_id=30


Gerardo, there's no arguing with you. You give suggestions that don't solve the problem. I clearly state the requirements and you come with things that clearly break them. What kind of dicussion does that lead to but some sheer nonsense. Now I'm a Xfrog / Greenworks fan that doesn't want anything else? I just suggested you to name a single app that could make continuous geometry with LW and had a botanic database. You couldn't name any. So, what is there left to say?
You know the options for LW, if you don't want to use it, it's your.. 'thing'.

But again, you don't need to get stuck to a single solution. If you want to use discontinuous geometry, use a nodal solution like PolyMove or PartMove. If you want that someone write a free script for you to solve the effect with SG_Fertilizer, at least specify with examples what exactly you need.


Because they work fine for many purposes. I guess most Xfrog & Onyx users don't habitully stick their camera into a branch connection. And even if they did, if you export with high quality settings, it's hard to see the difference.
Sorry but if you don't want to recognize that discontinuous geometry is a drawback for LW workflow and several other 3D packages, you'll never have continuous geometry within XFrog or Onyx, and I'm aware that it's indeed possible with L-Systems. However what surprise me is that any of them can export a simple guide curve with the structure of the tree yet? At least you should request that as a user.


Now that I've been destined to like 'crappy models', be a fanboy of Xfrog and Greenworks and rigid in my thinking for wanting a botanical database without having to buy maya or another 3D application, I sign off on this hilarious tantrum fest, and go do more useful stuff. Have fun there, Gerardo, and go try to make a realistic asian cherry tree with Treesdesigner.
Well, again, I'd try it by depending on the take, and I'd try also Vue or Xfrog depending on the take too :)



Gerardo