[PerryPlatypus]

For the record, this has been a bug for a long time. It appears in Stable 1.4, in Monogame 1.4-55 Rev. 110.4, and the latest Unstable versions.

The distant mountain tiles in the below screenshots were created in DEMEX using 1-arc-second USGS DEM files, and were created all at one time (rather than in different stages from multiple DEM files). Mosaic was used to apply photorealistic terrain textures. This route is an in-development payware route I am working on, so I cannot share it. However, this same phenomenon is noticeable in several locations on TrainSimulations' Mullan Pass route (try starting from Missoula going east, then at the west switch of Bonner look straight west), and I would guess it occurs in other routes as well?

Under certain circumstances, ORTS draws the row of mesh on the edge of Distant Mountain tiles incorrectly, as shown in the screenshots below, resulting in gaps/voids between where two distant mountain tiles touch. What is *very* odd is that when I start an ORTS session from about 20 miles to the south of this location (at the south end of the route) and drive north, the distant mountains draw correctly for the entire route, but when I start a session to the north of the DM tile boundary, OR when I save the game in this area and then re-load that save, the problem suddenly appears on all distant mountain tiles.

Note that TSRE does not ever display these gaps in the distant mountain tiles.

Screenshots below to show this phenomenon...

1. This is a view looking towards the southwest. In the first screenshot, I started this run about 30 miles to the south (start point was south of the distant mountain tile border in question), driving the train north. In this screenshot, the distant mountains are displaying correctly, no problems.
https://live.staticflickr.com/65535/51886525519_ef3acbd7f3_o.pngOpen Rails 2022-02-16 03-08-46 by Sean Kelly, on Flickr

2. However, as soon as I save the game, exit, and re-load that save, OR if I start a brand new run in this area, then the tile edges go out of whack.
https://live.staticflickr.com/65535/51886201091_f55ba3aaff_o.pngOpen Rails 2022-02-16 03-12-29 by Sean Kelly, on Flickr

3. A zoomed in view showing what is going on. The "seam" is right along an east-west boundary between distant mountain tiles. In the distance, a north-south boundary with a similar issues can be seen. As far as I can tell, when it gets triggered, this problem happens to ALL distant mountain tiles in the route - but it is most noticeable in areas of steep/mountainous terrain. And when it is not triggered, it does not happen to any DM tiles across the entire route.
https://live.staticflickr.com/65535/51886998182_80e4216b74_o.pngOpen Rails 2022-02-17 02-30-07 by Sean Kelly, on Flickr

4. An even more zoomed in view. This view is looking almost straight west. Here we can see that the problem is the very left edge of the distant mountain tile at right is not drawing at the correct elevations. Instead, it is matching the elevation of the next "mesh line" to the right.
https://live.staticflickr.com/65535/51888046468_061dd8d95a_o.pngOpen Rails 2022-02-17 02-33-15 by Sean Kelly, on Flickr

It is incredibly odd that this problem only occurs *sometimes*, and it repeatedly does not occur when I begin a drive from the extreme south end of the route driving northward, but it always occurs if I start at any point north of the southernmost row of distant mountain tiles...

[Genma Saotome]

In most cases the void occurs where near terrain and distant terrain meet. The cause is near terrain is built on an 8m grid and distant is 16m. On every 16m segment, if the corresponding 8m middle point is not on the 16m line there will be a gap. When this is not the case it can sometimes be a different resolution in the DEM data -- all accuracy, whatever is cited, can apply in all three directions and so if the DM image was taken at a different date or used a different resolution that the near terrain this built in error in the data can become visible.

I have taken to using 5 tiles of near terrain to the sides of every tile that has track to minimize this problem but it will never go away entirely so long as both terrain scales are used. Because of this one solution is to use Demex to drop the altitude of DM terrain by a couple of meters. That usually works.

[PerryPlatypus]

Hi Dave,

I do fully understand that, but all of these seams are visible even 20-30+ miles away. It is nothing to do with the near terrain tiles. All of the seams in the screenshots above are between adjacent distant mountain tiles, NOT between a distant mountain tile and a regular tile. These visible gaps are occurring in areas that are many, many miles away from any "regular"/near-terrain tiles.

In the screenshots at the bridge, those gaps are around 10 miles away - WAY beyond the 3,500 meter visibility I have my OR settings at.

My last screenshot is a telephoto (using the mouse scroll wheel) looking at DM tiles about 9-10 miles away, across the valley.

I extracted all of the distant mountain tiles at the same time using a single DEM file.

The fact that sometimes the DM tiles display correctly, and then simply re-loading a save causes them to suddenly display incorrectly means there is a bug. The distant mountains tiles clearly contain the correct mesh geometry - otherwise they would display incorrect 100% of the time.

[scottb613]

PerryPlatypus, on 17 February 2022 - 05:44 PM, said:

Hi Dave,

I do fully understand that, but all of these seams are visible even 20-30+ miles away. It is nothing to do with the near terrain tiles. All of the seams in the screenshots above are between adjacent distant mountain tiles, NOT between a distant mountain tile and a regular tile. These visible gaps are occurring in areas that are many, many miles away from any "regular"/near-terrain tiles.

In the screenshots at the bridge, those gaps are around 10 miles away - WAY beyond the 3,500 meter visibility I have my OR settings at.

My last screenshot is a telephoto (using the mouse scroll wheel) looking at DM tiles about 9-10 miles away, across the valley.

I extracted all of the distant mountain tiles at the same time using a single DEM file.

The fact that sometimes the DM tiles display correctly, and then simply re-loading a save causes them to suddenly display incorrectly means there is a bug. The distant mountains tiles clearly contain the correct mesh geometry - otherwise they would display incorrect 100% of the time.

Hi Sean,

I've been creating and using DM since the early days - I haven't come across that issue. I did just test with "Saves" and "Restores" - I can't recreate your issue. Perhaps you might try creating the DM in TSRE just to see if you get different results. I used TSRE for the first time to do this a few weeks ago. It doesn't take long to do and the process is the same as it is for creating normal tiles. Texturing with custom textures is another matter - since you're obviously vested in your DEMEX mountains. Just a thought.

We had this issue in Flight Sims - as we used multiple LOD resolutions on the terrain - we used to have to load a "Buffer Terrain Mesh" to plug holes.

Regards,
Scott

[charland]

I've had a lesser version of this crack show up and what I did was to go along the seam, select F7 , then select the sub-tiles on one side of the blue line and change the error bias to "0", click on the sky and save. Then do the same on the other side of the blue line and that took care of it.

Paul :-)

[PerryPlatypus]

Now that my Shasta route has been released, I'd like to bring this issue up again and see if any OR devs are willing to look into it. (The screenshots from the top of the thread were all taken between Black Butte and Hotlum). Another very common spot to observe this phenomenon is at Azalea looking north.

These gaps do not show up in TSRE, and they did not show up in MSTS. While TSRE created distant mountains apparently don't have this issue, the DEM data that TSRE uses is low res than what DEMEX can handle, and much more importantly, it appears Mosaic cannot handle TSRE created distant mountains.