[WaltN]

Post-TSRE5 v0.631: OR Fails with Null LevelCrossing (???)

This was my first time using TSRE5 (v0.631), so I expected something to go wrong, and it did.

But, first, let me say that my first impressions were very favorable. I read about the select, copy, and paste godsend, and I decided to try that. Then, I saw that the main window was resizeable. That about blew my mind -- a welcome feature, particularly when you're running a 2560x1440 Desktop.

In TSRE5, I selected, copied (Ctrl+C), and pasted (Ctrl+V) a single tree into a new position. I rotated that into a new orientation, de-selected the tree, and saved the route.

I then exited the editor and launched Open Rails. Proceeding from the main menu, I launched an activity in the route. Before loading was complete, I got a null reference exception in LevelCrossing.cs (line 51). I should mention here that I'm running the public 1.1 version of OR. I should also mention that there are NO level crossings yet in this route. (The first time I ran OR against this TSRE5-edited route, I picked an activity that was 40 miles from the tree added. So I ran it again in or adjacent to the tree's tile. OR failed the same way.)

I've attached the OR log for the first run. The route is not a new one, but it is still under construction. It is approximately 350 MB.

OR runs this route successfully when the route backup is restored.

Attached File(s)

•  OpenRailsLog - POST-TSRE5.txt (10.94K)
  Number of downloads: 7

[Goku]

WaltN, on 29 September 2016 - 04:20 AM, said:

Regarding the last two posts, Open Rails does not even bother to use the MSTS _n.raw terrain tile files (apparently because no one has been able to figure out the representation). Open Rails generates the normals from the height map at tile load time.

My app does the same - realtime, even during editing terrain:
https://www.youtube....h?v=hgyk30NIb-4

[Goku]

Quote

Post-TSRE5 v0.631: OR Fails with Null LevelCrossing (???)

Have you any roads in your route? If no - place one and add it to rdb (Z key).

edit: I see what is on LevelCrossing.cs (line 51) and it could be something much worse than I though ...

[Csantucci]

Rohit R Krishnan, on 28 September 2016 - 07:17 PM, said:

Please see the captures enclosed...Issues with signals and staggered traction lines. Traction lines are staggered at this area only.

Do you get staggered traction lines also if you run the route with MSTS?

[Rohit]

Csantucci, on 29 September 2016 - 07:04 AM, said:

Do you get staggered traction lines also if you run the route with MSTS?

I cannot run the route in MSTS...Since 2012, I am using only Open Rails...

[Goku]

Csantucci, on 29 September 2016 - 07:04 AM, said:

Do you get staggered traction lines also if you run the route with MSTS?

Go here: http://www.elvastowe...traction-lines/

[wacampbell]

Goku, on 28 September 2016 - 03:50 PM, said:

But I have one question: how to determine if shape uses normals or not? For example 2D trees and grass looks ugly with normals enabled.

This is controlled by one of the parameters of vtx_state.

vtx_state ( 00000000 0 -9 0 00000002 )

The parameters are:

dword,flags :uint,MatrixIdx :sint,LightMatIdx :uint,LightCfgIdx :dword,LightFlags

The third parameter, LightMatIdx configures lighting ( -9 in the above example ). Values from -8 thru -12 ignore normals.

-1     	light_material specular0       not used
-2	light_material specular25       "
-3      light_material specular750      "
-4      light_material Fullbright       "
-5	light_material OptSpecular0    Default setting - diffused lighting dependent on surface normal.
-6      light_material OptSpecular25   very shiney specular reflection - spec power = 25
-7	light_material OptSpecular750  slightly shiney specular reflection - spec power = 750
-8      light_material OptFullbright   all surfaces lit full regardless of normal
-9      light_material Cruciform       for plants etc
-10     light_material CruciformLong   also for plants etc not sure on the difference from -9
-11     light_material OptHalfbright   all surface lit medium regardless of normal
-12     light_material DarkShade       all surfaces are shadowed regardless of normal

I think this is what you are asking...


Wayne

[Goku]

Thanks. Do you know what these values mean? Cruciform?
I was thinking it will be something related to shader values.

[wacampbell]

Goku, on 29 September 2016 - 07:49 AM, said:

Thanks. Do you know what these values mean? Cruciform?
I was thinking it will be something related to shader values.

Yes, use LightMatIdx to configure your shader's diffuse light, ambient light, specular light and specular power inputs.

For the default lighting config, -5, pixel brightness depends on the vertex normal. Sun facing surfaces are brighter than those that face away.

By cruciform, they are referring to the cross-shaped billboard models used for vegetation models. The brightness of a pixel depends on the brightness and color of the sun, but vertex normals are ignored to disguise the geometry.

The other three configs, DarkShade, OptHalfbright, OptFullbright also ignore vertex normals, however their lighting also ignores sun brightness. So OptFullBright surface are lit at night.

You can see this behavior in the Open Rails shader code attached below. The shader techniques relate as follows:

lighting config -> shader technique
-5 -> ImagePS3
-6 -> ImagePS3
-7 -> ImagePS3
-8 -> FullBrightPS3
-9 -> VegetationPS3 ( cruciform )
-10 -> VegetationPS3 ( cruciform ) we are not sure the difference in MSTS
-11 -> HalfBrightPS3
-12 -> DarkShadePS3

Its DX shader code, but it may help understand the techniques we used.


Wayne

Attached File(s)

•  OpenRailsShader.zip (4.83K)
  Number of downloads: 14

[jamesc25313]

Goku, on 29 September 2016 - 03:30 AM, said:

Everything you want is not compatibile with MSTS.

Yeah, I know, but either I misunderstood your post, or I thought you were talking about implementing normals somehow. I was presuming OR.