skipper1945's loco's in TRS19

I had a brief look at the m.reflect materials on a couple of Skipper1945's loco's;

BR ex GWR 55xx <kuid2:116296:1225:4>
GWR 2251 class Collett Goods <kuid2:116296:262490:5>

Do either of these locos have texture problems in TRS19?

In both cases, I could not see anything wrong with the materials setup. They have the required Diffuse and Reflection maps in the correct slots of the Material Editor, and the image files are right there in the folder, as they should be. If these loco's are not rendering properly in TSR19, it would indicate a problem in the program rather than just unassigned texture maps (ie. a creator error).


.
 
Last edited:
I had a brief look at the m.reflect materials on Skipper1945's BR ex GWR 55xx loco, which is <kuid2:116296:1225:4>. Is this one of the locos with texture problems in TRS19?

In this case, I could not see anything wrong with the material setup. They have the required Diffuse and Reflection maps in the correct slots of the Material Editor, and the image files are right there in the folder, as they should be. If this loco isn't rendering properly in TSR19, it would indicate a problem in the program rather than just an unassigned texture map (ie. a creator error).


.

Yes afraid that is one of the affected, also appears to have flipped normals on BR Logo and numberplate on one side.
 
"And another reason to skip 19 for the forseeable future."

From a person who doesn't have TRS19?

There is only one valid reason not to migrate to TRS19 -- once you have it and run a TRS19 route, you can never ever go back to T:ANE. The improvements are that significant.

I've posted several TRS19 videos showing how good it looks. This is one of my favourites:



There are some teething problems. At least N3V are now aware of them.

Phil
 
Sadly though Phil, if changes in the sim are borking content from some of the most respected and careful content creators, there is reason to be worried. I'm not giving up yet but in the face of these new revelations it is becoming very hard to remain motivated, certainly when looking at big projects. These underlying issues were apparent to some of us when we first starting playing TRS2019 but we were told it was environmental settings that needed toning down or cat hair in the monitor. It is only now the real cause has come to light.

It would be nice to have an assurance from N3V the issue will be fixed otherwise route building becomes a rather haphazard exercise not knowing if objects or textures are affected and that when we come to drive a train that will look wrong to. It's only the fact TRS2019 doesn't have much competition for route building that keeps me clinging on, but at this stage I am beginning to wish there was a way to get my $70 back...
 
Hi Guys - I am currently running my Tristyn Route in TRS19 and, while acknowledging these issues, there really is so much potential for a lifelike experience.

I use AI operations and tend to ride the train from the outside of either the first or second carriage, which effectively "hides" this problem. "Hiding" is of course not a satisfactory solution, but I would suggest that it should not be difficult for N3V to resolve. That being said from a purely speculative position, I have noted that many other assets (vehicles and buildings) suffer from similar effects. Those specific circumstances would appear to be the result of little more than poor balance between "in shadow" and "out of shadow" orientation. i.e. It would be interesting to see the results of lightening the "in shadow" orientation.

It is disconcerting when, at 10:30am, a double decker bus sides in shadow are virtually black, as is the "in shadow" side of a GWR coach. At the risk of being simplistic, perhaps the lighting really does just need tweaking? Regards. Colin.
 
But Vern...

This isn't even the full product yet. It's still in testing, aka beta/early release, which means we're going to find stuff, report it, and hope that it gets fixed before the full product is released. Unlike T:ANE this product is in far better shape to begin with. Tony even said that they're looking into a solution for this issue so there's hope that it can be repaired in-code rather than needing to repair a gazillion-plus assets.

We have to all keep in mind too that unlike TS12 and its predecessors, and unlike T:ANE, TRS19 has a lot more going on under the hood, making the older stuff not work quite as expected. Whenever changes are made to the texturing engine, lighting, or other big components in a program like this, surely there's going to be something that's not going to come out in the wash. In many cases, we have no way of repairing some things such as this and it's up to the developers to come up with a solution. It seems that in most cases they are good at that. Take the uniform texture issue, and the shared texture issues that hit us in T:ANE SP1, for example.

Things get broken in software when things are changed due to how everything is connected. It's like pulling a thread in a Cardigan sweater. The next thing you know you've unraveled the whole thing. This is what happens when base code is upgraded or changed. A simple thing like improving the graphics can snowball and affect many other things. This is why we need to report errors as they occur rather than sit there and lose hope. Find something amiss, report it. Simple enough. The more reporting the better.
 
I agree that TRS2019 is in a slightly better state than TANE was on first release and at least it isn't burning out our GPU's (yet). When the updated graphics engine does its job properly then the results are very good.

However I don't share your confidence in N3V's ability to put it right. This is basic stuff in a sim which was trumpeted as retaining backward compatibility. If you're programming a new (!) lighting and graphics system, then you make darn sure you test a good cross section of that content in the game environment and you know what code changes might do to certain types of shader. Instead, as mentioned above, we got blame the users, blame the creators and an admission that the people (or is it person these days) working on the code don't even play the game.

This should not have been let anywhere near even early access until such fundamental issues were sorted.
 
Last edited:
I believe the main issue here is just the shader for m.reflect. Clearly as the forums is showing that material needs the most work, and I'm sure N3V will not let assets look like this for the official release. Where is there claims of users and creators being blamed? The only I can think of is the issue regarding the green texture, but that is the fault of the artists.

Clearly the amount of feedback the community is giving regarding this issue will get us some results, you just have to be patient. I'm not sure why people are not realizing it's still in early access.
 
Where do i start... short answer, if it was easy, it would all be done already.

Longer answer:

* We have a totally new material structure to ensure that new content can take advantage of modern techniques.
* We want backwards compatibility to ensure that 15+ years of route building isn't thrown out the window.
* We took a wide cross section of the 500,000 Trainz assets and worked on creating legacy shaders that would mimic the old look in the new engine
* One of the delays in releasing the original closed beta was working through these issue (if you are concerned now, you should have seen the build a few months ago!)
* We then got feedback from a wide cross section of the community during months of external beta testing and improved the look of a wide variety of legacy shaders as a result
* As we received problem assets, we investigated why these assets would look different to other, apparently similar, assets.
* Most often there is a different technique used somewhere in the pipeline (such as an inverted normal map) that went undiscovered in the old engine but become apparent with the new one.
* We rinsed and repeated these steps a number of times to work out "trade-offs" where missing data would be "correctly" interpolated
* Afaik, no-one reported this selection of assets until recently
* We are still working on improvements
* This is just one of the reasons we have not released the new routes yet
This should not have been let anywhere near even early access until such fundamental issues were sorted.
If it wasn't for Early Access, we wouldn't have a sufficiently wide network of users providing us with the additional feedback required.

Then having written the above, apparently the isue identified in the Skipper content is nothing to do with shaders at all, but rather a content issue. Note, saying it is a content issue does not mean "it's the content creators fault". It means that previous versions of Trainz didn't assist the content creator doing it the right way and it didn't matter if it was done the wrong way.

The problem in more detail is as follows;

* the material is double sided (i.e. the materials that appear "black")
* the normals are (as specified by the creator) the wrong way around.
* that means that the side you think is going to be brightly lit is dark, and vice versa
* in T:ANE, double-sided materials were handled incorrectly with both sides being lit the same.
* in TRS19, we handle it correctly:
** seen from the front, it gets the standard lighting applied
** seen from the back, the game automatically reverses the normals and shows you the correct lighting.


From a code perspective, there is no bug in TRS19. However, we also realise that content creators were not aware of the problem in previous versions, so we're investigating (and will implementing) a code workaround which would result in older assets being handled in a different manner that attempts to correct the normals at load time.
 
Last edited:
..apparently the issue identified in the Skipper content is nothing to do with shaders at all, but rather a content issue. Note, saying it is a content issue does not mean "it's the content creators fault". It means that previous versions of Trainz didn't assist the content creator doing it the right way and it didn't matter if it was done the wrong way.

The problem in more detail is as follows;

* the material is double sided (i.e. the materials that appear "black")
* the normals are (as specified by the creator) the wrong way around.
* that means that the side you think is going to be brightly lit is dark, and vice versa
* in T:ANE, double-sided materials were handled incorrectly with both sides being lit the same.
* in TRS19, we handle it correctly:
** seen from the front, it gets the standard lighting applied
** seen from the back, the game automatically reverses the normals and shows you the correct lighting.


From a code perspective, there is no bug in TRS19. However, we also realise that content creators were not aware of the problem in previous versions, so we're investigating (and will implementing) a code workaround which would result in older assets being handled in a different manner that attempts to correct the normals at load time.

Thanks so much for that explanation Tony, it makes perfect sense out of all the speculation and confusion. It gives me (us) hope that an in-house solution can indeed be found to keep the old assets alive, despite their previously undetected flaws. While this thread focussed on Skipper1945 content, the same problems could affect any creator's assets, including my own. I guess I won't find out how bad my stuff is until I get a copy of TRS19...
 
If it hadn't been for early release this potential problem wouldn't have been spotted as although N3V have carried out tests most of the developers do not "play" trains.
I have confidence that N3V having detected the problem will find a workaround or if necessary simply apply the old method for assets up to 4.5 and the new method for assets from 4.6 up as it is impratical for creators, most of whom no longer create, to amend their creations.
i would like to think that when I buy my copy of TRS19 next year that most of the basic prblems are sorted out.

Ken
 
As a Trainz 'non technical' person and not owning TRS2019 but a prospective buyer, my current feelings are thus:

skpmbhx.jpg


Rob.
 
Except that software is pretty much never "all fixed and working correctly".

In this case, we have software and tools to create user accessible content, supported by an open ended development community creating enormous amounts of additional content that can be used in an even greater variety of ways utilizing a variety of tools with scripts and animations that end users can combine into even more possible combinations on a wide range of hardware with graphics drivers provided by a multitude of hardware vendors, running on operating systems that are continously being updated.

This process is going to result in at least ONE bug creeping into the release! (Possibly more than one ;)).

As for most of the developers not playing Trainz, they tend to sleep in the hours they aren't actively developing Trainz, and for the most part spend probably half their day "using" Trainz in some fashion or other. So let's, before things get a life of their own here, put to bed the notion that the developers have no clue about how it works or what is involved.

I for one, do not "play" Trainz. However, I think I have a pretty good idea how the program works. Chris knows far more about HOW it works than I do. Paul has far more idea about how to make assets than I do. And so on.
 
Last edited:
Excuses? - So be it then, I won't be buying, no great loss to me as far as I can tell!

Rob.

Far from an excuse, and more a fact.

If you are looking for a bug free software program, this is not the place for you. If you are looking for the best, most stable and bug free version of Trainz we've ever released, then stick around :)

I would suggest more your loss not getting to even try it (which you can do for $15 as well as a range of other benefits during that month).
 
Far from an excuse, and more a fact.

If you are looking for a bug free software program, this is not the place for you. If you are looking for the best, most stable and bug free version of Trainz we've ever released, then stick around :)

I would suggest more your loss not getting to even try it (which you can do for $15 as well as a range of other benefits during that month).

No thanks, not at the moment, given all the negative opinions coming from many TRS2019 users at the moment!

One has to ask the question:

Who is Trainz aimed at, the train sim/model railway enthusiast OR the computer boffin with a degree in computer science?

Rob.
 
The flipped normals issue I had / have on some of my stuff again wasn't obvious in TS12 or TANE or in Max / Gmax. Anyway, this may help some creators, is easily fixed in Max or Gmax by running a Reset Xform on the affected parts and then re-exporting, may need to detach any welded bits first and keep a copy of the original mesh just in case! Still got a few more items to fix and re-upload, as I was intending to improve the lod on my stuff anyway it's not such an issue for me.

Why the normals are flipped? Think it's related to cloning parts and rotating etc when creating the mesh (Guilty of doing that here). Going to perhaps be more of an issue with those of us who create a full 3d mesh for the main mesh rather than using normal mapping.

@Tony H
Thinking back a while was there not an issue with double sided materials not working back in early TANE or perhaps TS12? Could the fix for that be the reason why double sided materials were being lit the same on both sides? Just a thought!

There are still issues with m.reflect and specular though even if it's not responsible for the problems with Georges locos.
 
As a Trainz 'non technical' person and not owning TRS2019 but a prospective buyer, my current feelings are thus:

skpmbhx.jpg


Rob.

Nice one Rob!

(Typing this as '19 does yet another 20 min asset rebuild for no apparent reason).

Despite my criticisms - which as a customer paying $70 and even with the early access caveat I feel quite entitled to make, I am prepared to give the benefit of the doubt a little longer in the hope this all gets fixed up. As I've said before, N3V have a perfect opportunity to corner the market share as the competition is struggling. MSTS/OR is ancient by any stretch of the imagination and seems largely the purview of bitter old dudes bickering at TS or Elvas. DTG can't make up their mind whether to keep going with their traditional TS programme which got a 64 bit makeover but still not very good for building routes, or focus on TSW. The latter supposedly now getting a route editor but based on the UE4 editor, so not likely to be hobby level friendly. N3V could really clean up if they so decided.
 
Despite my criticisms - which as a customer paying $70 and even with the early access caveat I feel quite entitled to make, I am prepared to give the benefit of the doubt a little longer in the hope this all gets fixed up. As I've said before, N3V have a perfect opportunity to corner the market share as the competition is struggling. MSTS/OR is ancient by any stretch of the imagination and seems largely the purview of bitter old dudes bickering at TS or Elvas. DTG can't make up their mind whether to keep going with their traditional TS programme which got a 64 bit makeover but still not very good for building routes, or focus on TSW. The latter supposedly now getting a route editor but based on the UE4 editor, so not likely to be hobby level friendly. N3V could really clean up if they so decided.

I'm sure they're trying but small companies pretty much seen to between a rock and a hard place. I love TRS19 and have no desire to go back in time, especially considering this is early access. Maybe I'm just easy to please but this old feller would be lost without Trainz.:)
 
Back
Top