archaeotect
-
Posts
42 -
Joined
-
Last visited
-
Days Won
2
Content Type
Profiles
Forums
Events
Plants Library
Posts posted by archaeotect
-
-
I am trying to imitate how branches radiate from a tree trunk in more or less random directions, but as you get to the level of twigs they will (at least in many species) fan out horizontally. The transition from vertical growth (to reach height) to a horizontal spreading (to maximize leaf surface area) can be quite subtle.
So far I've used circular distribution for the primary branches and angular distribution for the lesser branches and twigs and apply the Vector direction modifier with a global vector. But the lesser branches angular spread is influenced by the parent branch´s axis (which may have rotated).
Do I need to pick another vector, another node or use the vector orientation modifier or other path distributor options to ensure that twigs spread horizontally from the parent path, no matter what the parent axis may be?
I guess I'm stlll a little confused about the vector direction modifier (and it's orientation for leaves checkbox) and the vector orientation modifer. Any help or examples would be appreciated!
-
Thank you very much for your replies! I'm tending towards ForestPack right now.
MultiScatter explicitly mentions being fully multi-threaded on its website. I'm assuming the same for ForestPack (Itoo doesn't say) since you mentioned that it's fast.
Eduard, I already saw that offer - very tempting. Do you know if it's only available for a limited time? I might still need some time to make up my mind...
-
Hi,
I guess I'll need a dedicated scattering solution soon. Basically I want to convert GrowFX plants to VRay Proxies and populate huge scenes with them. From what I've seen both ForestPack Pro and MultiScatter look capable and very similar. Does anybody have any experiences to share regarding these two plugins (or others) and the differences between them? I know there's also CarbonScatter, but I haven't been to happy with E-On in the past. Any help is appreciated.
archaeotect
-
Thanks for this update. Looks very promising!
-
Good to hear that, I'm looking forward to the new version!
-
That's great news! I'm really looking forward to the future of GrowFX
And you're right of course, developing this as a general profile mesh is even better than just for the base of the cylinder, there will be many more applications!
-
Here's a (probably crazy) idea on how to improve the base of tree trunks created with the cylinder mesh builder. Having a perfectly round cylinder appear out of the ground often doesn't look very convincing. I'd like to give the mesh a more irregular base simulating roots. I know I can create root paths and use the Meta mesh builder to combine them with the trunk, but maybe there could be a quicker way.
My suggestion is to create a custom (closed) spline as the base of the tree trunk and have a mesh builder to blend this with the simple cylinder (in a meta mesh-like fashion?). I've attached a sketch to illustrate the idea. Basically you pick a closed spline to appear at height zero, enter a height value for when a perfectly cylindrical mesh should be reached and (optionally) adjust the blending with a curve.
Would something like this be possible? Maybe it would be a quick solution for trees that are neither hero trees with articulated roots nor far enough to simple pop out of the ground. And since we are talking about a custom spline there'd probably be other uses as well. What do you think?
-
Hi Eduard,
is there any news regarding my suggestions? Will you be able to implement any of them in the near future? Currently I have to create a forest of massive trees, and especially an Optimize Mesh or Optimize Steps option would be very helpful to reduce my polycount.
Please don't get me wrong, you are doing a great job with GrowFX, I'm just asking
Cheers
archaeotect
-
Sorry, I must have been blind
You are right of course - thanks for the tip! Feel free to delete this topic.
-
Hi Eduard, I've got another suggestion:
Would it be possible for the Deviation Modifier to deviate from higher levels of its own path? If I'm not wrong, right now it can only deviate from other paths distributed from it. That means if you want multiple levels of branches with deviations you cannot use the time-saving levels value in the Distributor, but have to manually create new paths for every level. A "Self-Deviation" option would be a great time-saver in that case.
-
While working on my trees I had some ideas of how to easily add more natural randomness to them. I thought about starting 3 threads, but I hope you don't mind if I put them together in this one.
1.) Distributors > Path distributor
Would it be possible to add a "count" value to the Starting path group? So that new paths only grew from randomly chosen x % of the path instances of the Starting path list? For example if you added leaves with a Path distributor and a starting path "branches", you could let them grow from only 90 % of the actual "branches" generated by the parent path.That would be a very easy way to create random dead branches. Or to let fruits/flowers grow only from some branches (without the need to manually create a seperate set of branches). If Affects could be directed at this value you could influence the distribution of the "dead" vs the normal starting paths, which would be great.
2.) Direction Modifiers > Optimize steps
The Optimize Modifier makes most sense at the bottom of Direction Modifier stack, but it does not work correctly below a Deviation Modifier. Right now I'm putting the Optimize second to last, above Deviation, but this does not always look good (sometimes there are not enough steps left to deviate smoothly). So I wanted to ask if it would be possible to create an AfterMesh Optimize Modifier that optimizes the final (cylinder) mesh (by deleting edge loops?)?
3.) Direction Modifiers > Start/End Slice
Would it be possible to add a random range (min/max) option to the slice length? So that instead of specifying a fixed Start/End Slice value in units you could also say that X % of all paths should be sliced randomly somewhere between Y % and Z % of their length (from either beginning or end).
Many trees contain branches that just broke off at some random point (mostly due to wind or snow, I'd guess). That could be simulated if this modifier could slice at random path positions.Of course I can't tell how difficult it would be to implement these changes and how important they are to other users, but maybe they could improve the GrowFX workflow.
-
Hi. I'm having a problem: Basically the path position of primary branches can no longer be affected by the trunk's path length after putting a deviation modifier on the trunk. Full description:
I was setting up a tree for growth animation (not so much for animation, but so that I could get its every age out of a single tree preset). I followed the video tutorial you hosted but made some additions. So I created a path length affect for the trunk and connected it to the starting position of the primary branches. The point was that a small tree would branch out very soon, but the large tree should have a very long trunk before any branches (like a tree in a dense forest that has to grow high and strait and has a much smaller leaf area than a freestanding tree).The branches are basically pushed upwards during growth. The affect was set up like this: Trunk: Parameter: Length => For length between 0 and 30 m the primary branches would be affected by 0,2 to 1,0, meaning branches would start from position 10 to 50 %. This worked fine and as expected.
But as soon as I put a deviation modifier on the trunk the affect curve ceases to have any effect. Only the first value (0,2 in this case) has any effect, the rest of the curve is completely ignored.
Is this the expected behaviour or am I missing anything? Any help would be appreciated!
cheers,
archeotect
-
Thank you very much for your efforts. I knew you probably already had considered everything a beginner like me could think of -and much more
-
I quickly ran into that problem, too. ptrain03's suggestions certainly help.
About implementing multithreading in metamesh creation, eduard said in another thread:
"Yes, we have not added yet support for multi-threading because Metamesh algorithm works sequentially, ie there each new face depends on previous faces."I understand that it might be difficult to have several cores working on one metamesh "welding point". But if a path object consists of seperate paths that never meet (like many branches being distributed along a trunk path) it should be possible to calculate these sub-paths simultaneously instead of one after the other? It might not be considered fully multithreaded, but I think it would be satisfactory if it could calculate several independent (sub-)paths at the same time.
Metameshes look so good that it's a real temptation to use them on the whole tree...
-
Thank you very much for your replies. I've just made the purchase. Looking forward to create some vegetation!
-
Hi,
I'm thinking about buying GrowFX. The trial version really convinced me. GrowFX seems so complex and flexible, but I was surprised how quickly you can achieve good-looking results. Still, I have two questions before purchasing:
- In another thread a developer mentioned on July 25th that a new version will soon be finished and released. What is the upgrade policy if I purchase now? Should I wait for the new version so I won't have to pay for an upgrade right after my initial purchase or will it be free? What is the general upgrade/update policy?
- In another thread in this subforum a member asked about multi-threading metamesh creation and was told that you are trying to find a method to implement this. The metamesh transitions between meshes look great but I also experienced significant lags (and 5 idle cores), so I wanted to ask how multi-threading is coming along (or if it's even possible)?
Thanks in advance!
archaeotect
horizontal branches
in For technical questions
Posted
Thank you very much. That's exactly what I meant.
I also learned why I couldn't get the same results so far: The parent and child paths had random direction modifiers, which seem to affect the angle axis. I replaced them with Noise direction modifiers to get the same kind of irregularity, and everything looks as it's supposed to.