Is there anyway to model correct topology without TransDEM? The reason I ask is that I don't want my routes to be flat with no elevation changes. While the area I am modelling is relatively flat there are some changes in elevation. I don't have TransDEM, I heard of other programs that act like it, but I don't know whether they work like TransDEM. Is there any other way to model topology?
Modelling topology without TransDEM
- Thread starter rs2007
- Start date
K4andT1lover
New member
Hi.
If it were me, I would find instructions for placing images under the route as a "template". Most people use Google Earth images; but with a little work, you could probably find a topographic map image to place instead. Either one would at least give you some indication on where to raise or lower the grid.
Hopefully someone has an even better solution. Good luck!
Ron
If it were me, I would find instructions for placing images under the route as a "template". Most people use Google Earth images; but with a little work, you could probably find a topographic map image to place instead. Either one would at least give you some indication on where to raise or lower the grid.
Hopefully someone has an even better solution. Good luck!
Ron
fant_autentico
Well-known member
For a real route or a fictitious one?
As Ron mentioned, you can use the basemap objects.
<kuid2:119912:60012:2> 1km Basemap I
<kuid2:119912:60004:2> 1km Basemap A
<kuid2:119912:60005:2> 1km Basemap B
<kuid2:119912:60006:2> 1km Basemap C
<kuid2:119912:60007:2> 1km Basemap D
<kuid2:119912:60008:2> 1km Basemap E
<kuid2:119912:60009:2> 1km Basemap F
<kuid2:119912:60011:2> 1km Basemap H
These are simple plain-objects that you replace the basemap name with an image of your own of the same size. It may mean scaling the image to fit because the expected image needs to be 1024 x 1024 pixels.
Now keep in mind this is only an image, you then need to raise the land over it since there is no terrain associated with it. You will definitely need to use layers here to lock your images in place, otherwise, they get in the way of placing assets, and they have to be sunk down a bit into the terrain so that they don't get in the way and work needs to be done in wireframe mode. Once things are in place, the asset(s) can be deleted from the route.
These are great for a small area, or small route because their placement is cumbersome and their use is awkward. This is the method also used to place model railroad plans scanned from books into Trainz, and user ModlerMJ wrote a program called Basemapz to import the plans and generate the baseboards with the tiles in place.
The other method is the old HOG program, or Hand-of-God as it's real name is called. This is really old and requires using MicroDem, a PC-only program if I remember correctly, to generate the underlying topographic information from satellite imagery. Once you do that, you need to import some kind of associated map image in HOG and that will then generate the landscape. This method is not accurate and it's very low resolution because the imagery doesn't always align with the underlying terrain and weird things happen.
There are some already generated routes like this that can be downloaded from the DLS that you can take a look at. Look for those by fishlipsatwork, for example. They are great for freelance work, but not good for prototypical modeling due to the inaccuracy of the terrain and maps. These were generated using the method above and the images are from the old TIGER-grant maps. The maps can be off by several meters or more putting the rails in rivers, over mountains, and other places where they shouldn't be. The same is with rivers ending up not where they're supposed to be as well.
In the end, even for fictional work, Trans DEM is the best option even with its learning curve. These other methods aren't accurate, and cumbersome. Even if the route is relatively flat, Trans Dem will combine all the steps into one ready to build route, and even place track if you want it to. Your job then is to fix up the track and place your objects. Using Trans DEM, you can take a real landscape and maps, and modify the generated route to suit your own fictional needs. I have done this numerous times with surprisingly realistic results. This is great for generating what-if scenarios, as you the route-builder figures out a possible route through various towns, or between cities. Using this, you can sometimes see why a railroad chose the route it did rather than one that you have explored.
<kuid2:119912:60012:2> 1km Basemap I
<kuid2:119912:60004:2> 1km Basemap A
<kuid2:119912:60005:2> 1km Basemap B
<kuid2:119912:60006:2> 1km Basemap C
<kuid2:119912:60007:2> 1km Basemap D
<kuid2:119912:60008:2> 1km Basemap E
<kuid2:119912:60009:2> 1km Basemap F
<kuid2:119912:60011:2> 1km Basemap H
These are simple plain-objects that you replace the basemap name with an image of your own of the same size. It may mean scaling the image to fit because the expected image needs to be 1024 x 1024 pixels.
Now keep in mind this is only an image, you then need to raise the land over it since there is no terrain associated with it. You will definitely need to use layers here to lock your images in place, otherwise, they get in the way of placing assets, and they have to be sunk down a bit into the terrain so that they don't get in the way and work needs to be done in wireframe mode. Once things are in place, the asset(s) can be deleted from the route.
These are great for a small area, or small route because their placement is cumbersome and their use is awkward. This is the method also used to place model railroad plans scanned from books into Trainz, and user ModlerMJ wrote a program called Basemapz to import the plans and generate the baseboards with the tiles in place.
The other method is the old HOG program, or Hand-of-God as it's real name is called. This is really old and requires using MicroDem, a PC-only program if I remember correctly, to generate the underlying topographic information from satellite imagery. Once you do that, you need to import some kind of associated map image in HOG and that will then generate the landscape. This method is not accurate and it's very low resolution because the imagery doesn't always align with the underlying terrain and weird things happen.
There are some already generated routes like this that can be downloaded from the DLS that you can take a look at. Look for those by fishlipsatwork, for example. They are great for freelance work, but not good for prototypical modeling due to the inaccuracy of the terrain and maps. These were generated using the method above and the images are from the old TIGER-grant maps. The maps can be off by several meters or more putting the rails in rivers, over mountains, and other places where they shouldn't be. The same is with rivers ending up not where they're supposed to be as well.
In the end, even for fictional work, Trans DEM is the best option even with its learning curve. These other methods aren't accurate, and cumbersome. Even if the route is relatively flat, Trans Dem will combine all the steps into one ready to build route, and even place track if you want it to. Your job then is to fix up the track and place your objects. Using Trans DEM, you can take a real landscape and maps, and modify the generated route to suit your own fictional needs. I have done this numerous times with surprisingly realistic results. This is great for generating what-if scenarios, as you the route-builder figures out a possible route through various towns, or between cities. Using this, you can sometimes see why a railroad chose the route it did rather than one that you have explored.
A real one.
colorado71
Well-known member
fant_autentico
Well-known member
Look here.............
https://forums.auran.com/trainz/showthread.php?158198-Real-orography-for-rookiesRHKluckhohn
Active member
I would urge you to get and learn TransDEM if you plan a route of any size. I created FDL Fond Du Lac Branch, a 30-mile long route,beginning about Christmas of 2009 and it was mid-2014 before it was finished enough to be presentable. It was 2015 or so before the terrain was complete and to this day I'm finding spots that are not finished, although they are not visible from the cab when running a train, so they don't need to be finished.
I worked from the publicly-available topo maps and roughed in the contours (10'=3.05M for my purposes) using the terrain tools. The track went in first with vertices set according to the map, then the LEVEL TRACK tool raised the terrain. Then I set the contour height in the USE HEIGHT tool and eyeballed the contours in, occasionally using the ruler tool to get the distance from the track, building up from the lowest to the highest. Once the contours were in, they were smoothed by stringing Invisible Track from top to bottom and levelling with the LEVEL TRACK tool.
Another technique used was to create the contour lines with invisible track and raise the terrain that way. It's more accurate and somewhat quicker.
If your landscape is low-relief, just eyeball the elevations in and smooth them as above and it shouldn't be too hard to produce something satisfactory. I found the best invisitrack spacing to be 15 meters to prevent gullying, although in some spots you want that effect.
Just remember, topo maps average heights out and do not necessarily show the minor sags and humps between elevations, so don't worry about being too particular. Work from a 10-foot contour if available.
I still kick myself for not starting with TransDEM!
:B~)
Its most of southern Ontario.
How much of southern Ontario? And there is a site for Canada where you could get DEM from. The other mapping sites only cover the US. You might get DEM overlap into Canada but TIGER data ends at the border if you look at a Niagara Falls DEM from the DLS you will see that road, rail, and river data ends at the border and the Ontario side is blank.
schweitzerdude
Active member
The SRTM data found at https://dds.cr.usgs.gov/srtm/ covers most of the entire Earth.
If you decide to use SRTM data, use this website which is an easier-to-use "front-end". http://dwtkns.com/srtm30m/
If you decide to use SRTM data, use this website which is an easier-to-use "front-end". http://dwtkns.com/srtm30m/
Last edited:
schweitzerdude
Active member
Depends on what you are trying to do. Small route, mostly flat, manual terraforming is fine. Large route, or route with varying terrain, TransDEM is the solution. Also if your route is going to be in a mostly level valley, TransDEM can provide background terrain to make things more realistic.
RHKluckhohn
Active member
Google Earth's altitudes are only approximations. If you look closely you'll see that they take jumps when you come to groves of trees, apparently registering the treetops, not the ground.
Find a grade/level crossing somewhere and lay rulers on the road and on the tracks. See if the heights at the crossing match. I've seen them off by ten meters in places in GE. Now Street View is useful for verifying your work and getting a sense of the territory.
If I were starting a new major route, I'd first get and learn TransDEM.
:B~)
schweitzerdude
Active member
Instead of Google Earth and using the method MP242 described just above, I suggest using software that actually shows contour lines. I like this: https://www.mytopo.com/maps/. One of the options, Map/Terrain actually shows contour lines, although they are faint and disappear if you zoom in too far. Another option, Topo Map, shows paper topomap style for USA and Canada (at least) with contour lines as well.
TransDEM is only the converter. Accuracy is solely depending on the geo data it processes.
For the elevations, we have to look at available DEM and their quality. I would leave Google Earth out of here. They do not produce DEMs themselves, they just import them from other sources and don't normally tell you what they used.
We distinguish terrestrial and orbital DEMs. The often cited SRTM data set is orbital, relatively low resolution and elevations subject to significant error in built-up areas and forests. A lot has been done in the last years to make this data better, as it is the only reasonable source for many parts of the world. Nominal horizontal resolution is 3 arc sec, that's about 90 x 60 meters at moderate latitudes, therefore often referred to as 90m DEM. There is also the 1 arc sec (30 x 20m) variant, and I would suggest to use that. Slightly better than 3 arc sec, but not living up its nominal resolution. So, wherever we have the chance of accessing terrestrial DEMs we should go for those. A terrestrial DEM of 50m horizontal resolution will have better accuracy than 1 arc sec SRTM, as they are produced the same way as topographic maps, from terrestrial surveying, without the elevation error for buildings or trees. Counterpart in the US: USGS NED 1 arc sec. Far better than SRTM 1 arc sec. And there is also the USGS 1/3 arc sec, I believe for all CONUS states.
In the last couple of years we see a new technology emerge: LIDAR DEMs. These can go up to 25cm resolution for some, albeit very small areas, due to the exorbitant amount of data. But the entire country of Denmark is available as 50cm data. In the US you find more and more 1m data, but still very limited. Noteworthy, that these LIDAR DEM always have a metric horizontal grid, because data in the Plate Carrée projection ("arc sec") would lose too much detail during the necessary conversion. Vertical accuracy is in the decimeter range. You will be able to make out plough furrows in the fields.
Cuttings and embankments ("high fills") for railway lines? No problem with a LIDAR DEM. But keep in mind that the maximum horizontal resolution in Trainz is 5m.
However, the DEM is only one aspect of geo data, the terrain. The other is the topography on top of the terrain. And again, it's the geo data source that is important.
In the ideal situation we have access to the official railway infrastructure data and track geometry, in analogue or digital form, but in reality that is rarely the case as this kind of data is not public. The German railway simulation Zusi in its variant as a professional driver/engineer training station sees its prototypical routes built with such data. The S-Bahn Munich project for example has the S-Bahn city tunnel modelled with geometry based on 1:1000 or 1:500 plans to the utmost precision. (If you are familiar with the area and have visited Munich, and who hasn't
, at least once as a tourist at the October-fest :wave:, not this year though 
.) For us normal people without access to original track geometry, we have to live with publicly available data. And as we all know, quality varies. Don't take the precision of ortho-imagery for granted, particularly if it comes from one of the global services like Google. If you happen to build with 1m LIDAR DEMs, take that as your reference, not the ortho-image.