[Weter]

We all know two japanese routes, one scottish and many australian.
There are two routes, known to me: on in Moscow itself, and another is 125 km to the North.
But it's unknown to me, how precise authors followed coordinates of the real-word trackwork there.

My personal route-builder's observations about distortion are reported in topic about training project.

[Siebren]

I've tested some routes and took a look at the maps. My findings:

* Chicago - Seattle: lat-lon fix make it definitely better. Placing in Everett is in the station, without fix in the city itself

* Bernina: same, much better, quite usable

* Edinburgh - Glasgow: better, but not as much as the above two

* Chiltern (London): unusable with/without correction

* OR_CPV (Portugal Algarve): unusable with/without correction

* Ruebelandbahn (Germany): usage good with correction

Don't know with which editor these routes are created. I don't see any relation between the location of the routes and the usability. When I started coding I used the Bernina bahn in Zwitserland - Italië for testing. And by then I was quite surprised with the usability. It also depends on the complexity of the railplans. The Bernina bahn is more or less one long stretch. London is much more complicated. For flightgear, where I found this map, this usability is not an issue. As the route between two airfields is not hard coded like with rail tracks. As long as the airfields are on the correct location everything is fine.

Can anyone tell me the name of a TSRE created route?

And I'll have a look if an idea I have can be implemented. Dragging the locomotive icon or double clicking on the map where the actual location is. And store the correction calculated that way somewhere related to the route. But technically that would mean feeding back data into the OR webserver. Done that before but not for Open Rails. Seen some API forum messages where that was happening, feeding data from the webpage back into OR.

[cjakeman]

Siebren, on 04 December 2022 - 03:06 AM, said:

* Chiltern (London): unusable with/without correction

* OR_CPV (Portugal Algarve): unusable with/without correction

And I'll have a look if an idea I have can be implemented. Dragging the locomotive icon or double clicking on the map where the actual location is. And store the correction calculated that way somewhere related to the route. But technically that would mean feeding back data into the OR webserver.

I'm guessing that "unusable with/without correction" means that the Lat-Lon is so far away that your animated map is useless.

Can you also tell us whether the error is larger with your proposal or smaller in these two cases?

Working with Edinburgh, I tried using your proposed constants and also adding an offset to the Lat-Lon to position the train in the right platform at Waverley. I guess this is equivalent to your idea of dragging the loco icon to where the actual location is. Then, when the train travels westward, it very quickly deviates from the track on the map.

That was disappointing, and I assume it is because Scotland has some distortion as shown by the grid on the Goode projection above. Instead of offsetting the Lat-Lon, perhaps we should be adjusting the constants to achieve the right Lat-Lon but I haven't tried that yet.


I am coming to the view that each region of the globe should have its own constants and perhaps these can be provided as part of the route.

[Weter]

Some authors turn their "maps" of real-based routes 90⁰ due to unknown for me reason.

[BillC]

cjakeman, on 04 December 2022 - 11:41 AM, said:

I'm guessing that "unusable with/without correction" means that the Lat-Lon is so far away that your animated map is useless.

Can you also tell us whether the error is larger with your proposal or smaller in these two cases?

Working with Edinburgh, I tried using your proposed constants and also adding an offset to the Lat-Lon to position the train in the right platform at Waverley. I guess this is equivalent to your idea of dragging the loco icon to where the actual location is. Then, when the train travels westward, it very quickly deviates from the track on the map.

That was disappointing, and I assume it is because Scotland has some distortion as shown by the grid on the Goode projection above. Instead of offsetting the Lat-Lon, perhaps we should be adjusting the constants to achieve the right Lat-Lon but I haven't tried that yet.


I am coming to the view that each region of the globe should have its own constants and perhaps these can be provided as part of the route.

Hi Chris
I am not going into the horros of the GH projection, this has been discussed in many prior posts. The poblem discussed above has to do re-projecting GH on Web Meracator projection from GPS coordinates used with Open Street Maps, from which OpenRaiway map is a derivitive.

For back ground look at some of the folloing web links:

GISsurfer Map with Military Grid Reference Sysytem (MGRS) Coordinates Learn MGRS with color-coded coordinates

In the reference pdf GISsurfer USNG and MGRS Coordinates section 3. 911 dispatching and responding discusses problems using GPS coordinages from there phones, UTM coordinates, and an address.

A Guide to Coordinate Systems in Great Britain goes into highly technical detail. However, section 1.2 A few myths about coordiante systems, is worth a read.

This document MAP.ARMY discusses MGRS system in detail.

IMHO I agree with your conclusion that each route has to be tuned to it's own constants, taking there "poetic license" in account.

Bill

[Aldarion]

Siebren, on 04 December 2022 - 03:06 AM, said:

I've tested some routes and took a look at the maps. My findings:
(...)
* OR_CPV (Portugal Algarve): unusable with/without correction
(...)

This one was creatd with MSTS RE in 2003.

I have a new route layed just this year with TSRE usings Darwins's new route tutorial on UKTS.

The route is called PT2030 and I have it in a private github repo along with other routes. Can you please provide a username so i can give you permission to access the repo?

[Siebren]

my github username is:

sweiland-openrails

[Aldarion]

Invitation sent.

[eric from trainsim]

I tried it today, and it's useful for approximation purposes.

I'm seeing in Chicago that the cursor is between 4-10m to the West of where the location is on the route itself. I've built my routes using Google Maps as the Terrtex, so I was expecting a little more accuracy.

If anyone has the LS&I, I'd expect that to have some variance as well thanks to the projection being distorted above 45N.

I'll test that and one of my own routes later tonight.

[cjakeman]

cjakeman, on 04 December 2022 - 11:41 AM, said:

Working with Edinburgh, I tried using your proposed constants and also adding an offset to the Lat-Lon to position the train in the right platform at Waverley. I guess this is equivalent to your idea of dragging the loco icon to where the actual location is. Then, when the train travels westward, it very quickly deviates from the track on the map.

I've now tried adjusting the constants instead of adding an offset to the Lat-Lon to position the train in the right platform at Waverley, Edinburgh.

Travelling westward, the train deviates from the track on the map just as before.

It seems that a good solution will be more complicated than just setting a couple of constants - shame.

[Siebren]

hi Aldarion,

thanks for sharing this Portugal route, I like Portugal.

I tested the map option with your CP2030 route. More puzzles... When I start your route in Santa Apolonia, the map and the compass shows Lat: 38,719480 and Lon: -8,591463. Which is not in Lisbon, but somewhere east near Canha, some 50 km's difference! However when I open the route in TSRE5 and via the Navi Window jump to Santa Apolonia I see Lat: 39.5288 and Lon: -8.46892. Which is precisely in the station of Santa Apolonia.

Then I created a route of my own with the station Faro. In TSRE5 the lat/lon are not correct, however when running OR on this route the map and compass are correct. So just the other way around.

[Siebren]

hi Chris,

I added some code to move the icon on the map to the correct starting point. I share your conclusion that it does not help. When moving it deviates unacceptable from the track for the Edinburgh route. And there are more...

However there are also acceptable routes: Ruebelandbahn, Chicago - Seattle (BNSF), Bernia, ZigZag Australia.

And with acceptable I mean that you can see around where you are. Which bridges you cross. Which tunnels you drive into. But not accurate enough to see on which platform you are, which switches you cross.

I still think it's a valuable enhancement. But with some sort big disclaimer in the manual and/or the menu where you start the map.

I could also have a look at the TSRE5 code. Cause somehow TSRE5 knows the correct position. Which I saw in both Aldarion route and my little try. But the TSRE5 code looks not that easy.

[Aldarion]

Siebren, on 08 December 2022 - 02:14 AM, said:

hi Aldarion,

thanks for sharing this Portugal route, I like Portugal.

I tested the map option with your CP2030 route. More puzzles... When I start your route in Santa Apolonia, the map and the compass shows Lat: 38,719480 and Lon: -8,591463. Which is not in Lisbon, but somewhere east near Canha, some 50 km's difference! However when I open the route in TSRE5 and via the Navi Window jump to Santa Apolonia I see Lat: 39.5288 and Lon: -8.46892. Which is precisely in the station of Santa Apolonia.

Then I created a route of my own with the station Faro. In TSRE5 the lat/lon are not correct, however when running OR on this route the map and compass are correct. So just the other way around.

Hi

Probably the ony point of coincidence is the geographycal cener of portugal. Acording to Darwins's manual when starting a new route we should select the geo center of the area (in my case the entire country) and then make the necessary calculations to enter the appropriate TsreGeoProjection line in the .trk file.

One that was done i started using TSRE to import the heightmaps and map layers... since those two were coinciding and after some tests I reallized that distances were maintained I started laying track paying no attention to actual coordinates.

So with all this, my question is... how do I lay a new Route where i can import heightmaps, map layers that keeps distances ( right now, in PT2030 i have a +0,1566% distance variation. Quite acceptable.)

[Laci1959]

Quote

So with all this, my question is... how do I lay a new Route where i can import heightmaps, map layers that keeps distances ( right now, in PT2030 i have a +0,1566% distance variation. Quite acceptable.)

Hello.

I use the GPS Visualiser program for this purpose. I navigate to the railway station where I want to start the construction of the track. I will pick up 2, 3 markers. I went to gpx format and downloaded it.
After that I do the new route with the TSRE program. I don't care where it will be. I accept what TSRE offers.
I'll correct it later.
I've gone the new route, I'm closing TSRE. I copy the downloaded gpx file. I open the new course, select the GPX marker list in the Navi Window. I select the appropriate marker from the list and jump there.
I will make a new tile, download the topography and, if necessary, the satellite map.
Based on the Navi Window, I correct the data of the trk file. Starting tile and coordinate.

https://kephost.net/p/2022/49/8066_266f7e7484de.png

https://kephost.net/p/2022/48/9279_611162ab9091.png

https://kephost.net/p/2022/48/9696_a402afb07fbd.png

The accuracy can be seen in the last two pictures. Half of the building is the difference, but that could also be my fault. I wasn't aiming accurately enough.
Sorry for the Hungarian captions of the pictures, I didn't think to post them here.

Sincerely, Laci 1959

[Aldarion]

Ok. Tried that

Here are the results:
https://i.ibb.co/KFdJ0SQ/Image1.jpg

Has you can see the difeerence betwen the GPs Visualizer and TSRE Geo lenght is quite good. just 5m difference. It means that in the totality of the 320km of the route I will have 1.6 km more in lenght -> +0,5% variation. I belive that in timetable mode I wont have any problem in adhering to timetable.

But now, looking at game lenght i get quite a shock: the real 1691m (average) are actually 2081m. So, in 320km I will gain 124,8km... +39% variation in track lenght. I can't adhere to timetables. My trains will always arrive late...
In that regard, I cannot give value to such a fine addition to the OR web interface... it's just aesthetics.
Unless there's a relatively simple way to deform the geo heigh data and map layers to insert ont TSRE5 or maybe an editro to transform maps that will appear in the web interface ... wich quite frankly jsut sound too complicated for me... as a Route developer I will choose not to pay attention because route lenghts are more important to my routes objectives than coordinates.