Trainz Plus Plant Factory Release Candidate

Interesting, after the Update a notification appeared from my AV program to inform me that TRS22.exe (aka Trainz Plus) had now been removed from its Real Time exclusion list.

I ran all the recommended checks to ensure that the updated TRS22.exe was not infected - it wasn't.

Added TRS22.exe back into the exclusion list.
 
Interesting, after the Update a notification appeared from my AV program to inform me that TRS22.exe (aka Trainz Plus) had now been removed from its Real Time exclusion list.

I ran all the recommended checks to ensure that the updated TRS22.exe was not infected - it wasn't.

Added TRS22.exe back into the exclusion list.
That's good to know. Thanks
 
Updated build 129343 to build 131950 with no problems. Update went very quickly. Ran several of my old routes and so far no issues.

The real question is what to do next. How do you go about replacing all the old treez with the new pf trees. Some of my routes have thousands of the treez and shrubs. Many of which do not have counterparts in the new format. It seems like this process will takes months to convert. Is there a reasonably plan on how to approach this process?
 
Updated build 129343 to build 131950 with no problems. Update went very quickly. Ran several of my old routes and so far no issues.

The real question is what to do next. How do you go about replacing all the old treez with the new pf trees. Some of my routes have thousands of the treez and shrubs. Many of which do not have counterparts in the new format. It seems like this process will takes months to convert. Is there a reasonably plan on how to approach this process?
You can use the bulk update utility to replace the trees, but you'll have to do this multiple times as new trees are created.
 
Is there a reasonably plan on how to approach this process?
As John posted above, you can do this in stages. In my case I use a very restricted set of trees and bushes in my routes - mostly a subset of the SpeedTrees Eucalyptus trees. I will be bulk replacing them with selected Plant Factory Eucalyptus trees that I have just installed.
 
Has anyone having installed the update experienced improved performance?
On some routes yes, but on my very large route I have not. That route has issues to begin with and now things are worse. Splines and tracks draw in chunks with bits of track remaining missing. Baseboard sides are missing making trees in the distance appear to float above the ground. Once the terrain grows in, everything is fine.

In other places, the scenery pops up in chunks or appears and then disappears as the program is attempting to figure out what to do with it.

I've removed some splines and replaced others with some good results but the thing is, I don't use a lot of them other than fences, walls, roads, telegraph poles, and powerlines. I removed the grass splines ages ago when I started upgrading to the TurfFX grass.

I had this issue with the beta versions but didn't report it because I'm not sure this is a bug since my other routes load in fine.
 
I have not noticed any improvement in the three largest routes I have. Maybe that is because I have not replaced my old trees with the plant factory trees.
The good news is that I have not noticed any issues with any splines, track, road, or grass. I am running my session with everything maxed out. The distant views appear the same.
Is there supposed to be any improvement in frame rates or GPU usage if the new trees are not installed? I have reread all the posts regarding the imposter system and it's not clear to me what assets other than the plant factory that are affected.
 
I have replaced all my SpeedTrees Eucalyptus Trees with their Plant Factory equivalents in two large routes using the S20 Bulk Replace feature. Apart from a few errors of my own making the replacement went very well. I have not noticed any improvements in performance but I still have many other SpeedTrees vegetation to replace for which there is not (yet) any Plant Factory direct equivalents. Jacaranda Tree PF anyone?

The main benefit to me, so far, is in removing the ancient licenced SpeedTrees technology, which has always had some limitations, with modern non-licensed equivalents.
 
Interesting as the imposter system was promoted as a general performance enhancer: “Trainz automatically generates camera facing billboards for each asset's lowest LOD, in real-time! This helps improve performance, not just for trees but most other scenery and even rolling stock!”

(From Trainz+ website)
 
“Trainz automatically generates camera facing billboards for each asset's lowest LOD, in real-time! This helps improve performance, not just for trees but most other scenery and even rolling stock!”
Which makes perfect sense. But how many assets do we have installed and are using in scenery that do not have LOD? Some of the latest assets from well known creators still do not include LOD.

Without LOD in an asset then the only thing that the Trainz rendering engine can do is to create the full asset at its original resolution no matter how far away it is located in the distance. That must have a performance penalty.
 
I have reread all the posts regarding the imposter system and it's not clear to me what assets other than the plant factory that are affected.
The imposter system works with any 3D tree. I've been testing it on the KCS 2 route which has no Plant Factory assets installed. Performance is about the same in my tests as well but that comes down to tuning the system which is why we are testing it in the beta. Once the team from the sweet spot where the transition from 3D tree to 2D tree is mostly unnoticeable and performance is at least as good as SpeedTrees if not better then the system will be rolled out.

The concept of the imposter system is that when the player's POV is close to a 3D tree, you see the tree as is. But as the tree gets farther from the players POV then first it hopefully goes through several LOD levels and then the imposter system converts it to a 2D image of the tree which is always facing towards the camera. Basically, the system takes a picture of the tree on the fly and creates a 2D billboard tree that replaces the 3D tree. It is at this point that some performance improvement can be seen especially when draw distance is set very high. At the same time, distance mountains sides will have 2D trees with a realistic profiles on them rather than very low poly spikes of green that look nothing like the profiles of the original trees.

Plant Factory is just a program for making 3D plants. These plants are polygon based so if a creator goes way over the suggested limit of 5000 polygons for trees then they can have a negative impact on performance. So it is a balancing act of appearance vs performance when creating assets with PF.
 
The imposter system works with any 3D tree. I've been testing it on the KCS 2 route which has no Plant Factory assets installed. Performance is about the same in my tests as well but that comes down to tuning the system which is why we are testing it in the beta. Once the team from the sweet spot where the transition from 3D tree to 2D tree is mostly unnoticeable and performance is at least as good as SpeedTrees if not better then the system will be rolled out.

The concept of the imposter system is that when the player's POV is close to a 3D tree, you see the tree as is. But as the tree gets farther from the players POV then first it hopefully goes through several LOD levels and then the imposter system converts it to a 2D image of the tree which is always facing towards the camera. Basically, the system takes a picture of the tree on the fly and creates a 2D billboard tree that replaces the 3D tree. It is at this point that some performance improvement can be seen especially when draw distance is set very high. At the same time, distance mountains sides will have 2D trees with a realistic profiles on them rather than very low poly spikes of green that look nothing like the profiles of the original trees.

Plant Factory is just a program for making 3D plants. These plants are polygon based so if a creator goes way over the suggested limit of 5000 polygons for trees then they can have a negative impact on performance. So it is a balancing act of appearance vs performance when creating assets with PF.
Animatek did this over 20 years ago with World Builder. They implemented a Z-buffer system that could turn polygon trees and plants into flipboards that turned to face the camera.

From what I understand, N3V plans to do this with other assets as well such as buildings. If this is successful, this will be an awesome improvement for us.
 
Back
Top