TransDEM issues

Well, the nice thing about this route is that it could be compressed via portals or design because there are long, long stretches of nothing. Even highlighting key parts will be a challenge. The Copper River can (and has) change drastically from year to year. The bridge I am starting with was taken out every year by break-up and then replaced. They designed a raised steel bridge but it was determined to not be cost effective. Today, there is a raise road bridge. You can street view it, and there are plenty of photos. It is just a mile or so northeast of Chitina. It is obviously more cut in and higher than the series of railroad trestles.

If I want to prototype, I really need to find decent maps from that era. But, like I said, I'm new to route-building, so I need to work on texturing and general Trainz skills too. Also, I want to build custom assets in Blender - at least some buildings and whatnot. So, this is a long-term hobby, just like model railroads. My initial goal will be to focus on a just a small section of track leading into and out of Chitina - 5 or 10 miles at most. So far, I only have a few low-res photos of Chitina. The survey is from 1909, and rail was not completed until 1911, so I have no guide to the yard or anything. I have more books on the way. Like I said, this could be years.

The great thing about all of this: I don't have to do the remaining 185 miles to accomplish anything. If I can pull this off, I should get some amazing screenshots and short diorama-style videos, and that really would be something since so little remains of this railroad in any media. And, if I do get the hang of route building, I can attempt Cordova or the Million Dollar Bridge or the mud flats or Kennecott or so on. The only sad thing is: no tunnels. There was one, but bits kept crumbling, so they just made it a cut-through. Maybe I'll make a fictional branch in to Bering Lake or west to Valdez. I'll find a tunnel somewhere.

To be honest, precise rail data might be wasted on my current route-building talents. However, learning to sculpt realistic terrain for the area is a challenge and task all by itself, so TransDEM will be a big help. Since I don't need typical railroad, street, and building map data, I am learning how to leverage the map data that TransDEM can put on ground textures so that I can repaint nature in the right areas, if that makes any sense. If/when I do get readable railroad data, it will be after TransDEM has advanced to putting that on ground textures. And now I am much more informed about geo-referencing maps images.

By the way, I am so looking forward to destroying that trestle bridge once I finally build it!

It's good to see everyone enjoying TransDem! It's been awhile for me, but with a little help, I learned the basics and built my route.

Cascade; Don't ever try to take the "If I can do it.." title away from me! Truly, if a moron like me can figure it out, anyone can! I knew you would get onto it in no time. I think my problem at first was "over thinking". Never a good thing with my limited brain power!

Cheers...Rick

Amassing 21 DEM maps covering all across Pennsylvania, I have moved all my Transdem data files to a 1TB external hard drive, as my disc was filling up

cascaderailroad said:

Amassing 21 DEM maps covering all across Pennsylvania, I have moved all my Transdem data files to a 1TB external hard drive, as my disc was filling up

Click to expand...

Congratulations on the birth of your TD Data files. Happy route building!

John

I expect that my finished size of Transdem data will be @ 30GB, (and growing), on my external 1TB hard drive.

And I want to again apologize for my past comments about Transdem, as PC problems on my end were complicating the running of a fine program ... not top mention the utter frustration of troubled family life that exists in my home, and the hustle and bustle of the holiday scavenger hunt (which it is not supposed to be what the event is all about anyway).

Yikes!! I'm modeling just a little corner of northeastern Massachusetts, with dash of Vermont and New York thrown in. Cascaderairoad is modeling Pennsylvania!! And Cheechako is modeling Alaska!!!!!!!.......now thats gonna take some serious hard drive space........

Yes it does, I have NED and raster data from Philadelphia to Pittsburg and Harrisburg to Erie. I have settled on doing the Pennsy from Mount Union to Johnstown. I did do a map from Sunbury to Erie but the Trainz engine would not handle it.

.

steamboateng said:

Yikes!! I'm modeling just a little corner of northeastern Massachusetts, with dash of Vermont and New York thrown in. Cascaderairoad is modeling Pennsylvania!! And Cheechako is modeling Alaska!!!!!!!.......now thats gonna take some serious hard drive space........

Click to expand...

Nah, just a tiny little part of Alaska - just one percent of one route. Speaking of which, I was experimenting with geo-referencing historical 1951 topos and exporting with the custom WebColors. This method seems to have a lot of potential, especially because UTM can be such a pain in deep mountain valleys.

There seems to quite a few Trainzers modeling Pennsylvania railroads. We, at Yesterdayz Trainz, are considering a route encompassing the Pennsylvania anthracite country along the Susquehanna River, just south of Scranton, through Wilkes-Barre to Nanticoke. Lots of Trainz operational possibilities in a relatively small area. We have all the maps and DEM data downloaded; with preliminary Trainz maps already made.

cheechako said:

I was experimenting with geo-referencing historical 1951 topos and exporting with the custom WebColors.

Click to expand...

Usually, the default TransDEM ground textures will not miss detail with any topo map because of the limited number of colours. The idea of those textures was to use as few colours as possible - since the number of ground textures is limited to 250 and you will want to save enough room for your final textures. The default textures were created by computing the average for numerous topo maps from different agencies around the world and then using an octree to find an optimal subset of colours. The default textures also feature the 10 m grid lines.

For a 1:24k map the most important aspect is to generate 5m terrain, not 10m, to be able to resolve map details spatially.

geophil said:

The idea of those textures was to use as few colours as possible - since the number of ground textures is limited to 250 and you will want to save enough room for your final textures.

Click to expand...

Good to know. Of course, I'm still just messing around with TransDEM and seeing what I can do with it. The historical topos I geo-referenced seem to be shifted to the east (north or south too?) ever so slightly. In any case, I really don't want 1951 data so I won't continue with that data. The topos might help as a visual reference outside of Trainz, so it was worth the download.

Question, though: As I said, I want to just attempt a small slice of this railroad. I probably only need one DEM. However, it is a DEM made up from three others because I am right in a corner. I cannot seem to add a 5m DEM no matter what - it eventually aborts. I can merge the 3 10m DEMs without any problem. I don't want to export it all to Trainz - I can just draw a route and then set a radius for export. If I use the 5m DEMs, must I do 3 Trainz routes and merge there?

I have a Win8 64-bit i5 with 8GB if that matters in the merging of DEMs. Like I said, I don't want 3 DEMs in Trainz - I need smaller L-section out of the DEMs.

Question 2: Do I really benefit from 5m DEMs, or do I just want to import 10m and export to Trainz as 5m? Keep in mind that I have no accurate track data, so I will have to sculpt the terrain a lot.

Edit: I only need the 62w145n DEM, and only the lower half or so, but I also need a narrow extension into the DEM to the east. I was thinking three DEMs because I was playing with three topo maps. I really have no plans to take on Alaska just yet! 2/3 of one DEM + 1/3 of another DEM should do me for quite a while.

cheechako said:

Good to know. Of course, I'm still just messing around with TransDEM and seeing what I can do with it. The historical topos I geo-referenced seem to be shifted to the east (north or south too?) ever so slightly. In any case, I really don't want 1951 data so I won't continue with that data. The topos might help as a visual reference outside of Trainz, so it was worth the download.

Question, though: As I said, I want to just attempt a small slice of this railroad. I probably only need one DEM. However, it is a DEM made up from three others because I am right in a corner. I cannot seem to add a 5m DEM no matter what - it eventually aborts. I can merge the 3 10m DEMs without any problem. I don't want to export it all to Trainz - I can just draw a route and then set a radius for export. If I use the 5m DEMs, must I do 3 Trainz routes and merge there?

I have a Win8 64-bit i5 with 8GB if that matters in the merging of DEMs. Like I said, I don't want 3 DEMs in Trainz - I need smaller L-section out of the DEMs.

Question 2: Do I really benefit from 5m DEMs, or do I just want to import 10m and export to Trainz as 5m? Keep in mind that I have no accurate track data, so I will have to sculpt the terrain a lot.

Edit: I only need the 62w145n DEM, and only the lower half or so, but I also need a narrow extension into the DEM to the east. I was thinking three DEMs because I was playing with three topo maps. I really have no plans to take on Alaska just yet! 2/3 of one DEM + 1/3 of another DEM should do me for quite a while.

Click to expand...

Isn't Roland awesome! He explains stuff so well.

Regarding the size issues and merging. The computer runs out of memory (not disk space) while TrandDEM is processing the merge. There's not much you can do with this, although I can handle, sometimes I may add, two maps merged together by changing the grid to 10m on import.

Notice the "sometimes".

The way around this is to get the maps first then bring in one DEM. Trim that down to the size of your map image, then bring in the second DEM and repeat that and merge the two together. Repeat this as many times as needed to make the complete DEM.

The problem isn't so much with TransDEM, but sadly due to the US Gov supplying such large data-sets. This was for their convenience rather than for the end-user as they were receiving 20,000 data-download requests daily and their servers couldn't handle it. Instead of expanding the servers, they opted to consolidate the data. In the old days we could download a tiny slice, a tiny square if we wanted, which was perfect for what we need. We don't need 1 arc-second slices of data as we're not universities or big companies using the data for other uses.

John

JCitron said:

Regarding the size issues and merging. The computer runs out of memory (not disk space) while TrandDEM is processing the merge. There's not much you can do with this, although I can handle, sometimes I may add, two maps merged together by changing the grid to 10m on import.

Click to expand...

I figured as much. I was just messing around, though. I can merge at least 3 10m DEMs but fail with any 5m DEMs; perhaps I didn't trim them enough. However, for my real needs, I think 10m is fine, especially for the distant terrain.

"In the old days we could download a tiny slice, a tiny square if we wanted, which was perfect for what we need. We don't need 1 arc-second slices of data as we're not universities or big companies using the data for other uses."

Tempus fugits, John! Why it seems like 'the old days' was barely a year ago.
My approach to making routes which overlap two or more 1dg x 1dg DEM's, is to georeference and overlay the maps relevant to each 1 deg section, then delete the irrelevant DEM points right there in TransDEM, then repeat the same for each 1dg x 1dg DEM. Finally I join them all into a single DEM in Trans DEM, then import to Trainz. If you are making a large route, I recommend the 10m grid setting. A 5m grid creates large files which will drag down your machine. I also recommend the 0 setting for smoothing when processing the Trainz map. You'll get a more accurate representation of the terrain in Trainz.

steamboateng said:

"In the old days we could download a tiny slice, a tiny square if we wanted, which was perfect for what we need. We don't need 1 arc-second slices of data as we're not universities or big companies using the data for other uses."

Tempus fugits, John! Why it seems like 'the old days' was barely a year ago.
My approach to making routes which overlap two or more 1dg x 1dg DEM's, is to georeference and overlay the maps relevant to each 1 deg section, then delete the irrelevant DEM points right there in TransDEM, then repeat the same for each 1dg x 1dg DEM. Finally I join them all into a single DEM in Trans DEM, then import to Trainz. If you are making a large route, I recommend the 10m grid setting. A 5m grid creates large files which will drag down your machine. I also recommend the 0 setting for smoothing when processing the Trainz map. You'll get a more accurate representation of the terrain in Trainz.

Click to expand...

Yup. Only a year ago. It's like everything with technology. A year seems like a lifetime sometimes.

I'll call you later. I talked with Frank earlier.

John

I really feel the need to say this...

There is some great info in this thread. Some of it comes from the author of TransDEM - a "third party tool". This info will probably be here to help others months from now if they search for it. If I build even this tiny slice of Alaska, I'll probably have to raid the DLS to do it and then build more custom stuff on my own, and get more help like this here.

The official forums (on Steam) for the other sim were just reset again. We're not allowed to have discussions like this over there. Even a simple discussion of techniques used to model distant mountains might get purged in a monthly reset. If I want to mix in freeware content (even my own), I cannot hope to ever share my routes there and really shouldn't even discuss it less I confuse official product and paid DLC discussions.

I wish I could afford to bump my pledge even more than it is now. I would like to thank all the helpful and talented people here in this community, as is evident in this thread. And, as far as I can tell, N3V doesn't seem to try to prevent that community from continuing from month to month, so thanks to them too!

Geophil monitors the forums and is the 'go to' man for TransDEM. Other's, like John and myself, have had the program for several years; we know enough to guide 'newbs' around, and are somewhat familiar with its limitations, but that's about it. Geophil uses words with lots of letters, and he understands 'computer geek'. (So does John!) 'Georeference' is my favorite TransDEM word, and it has many more letters than most words in my vocabulary. My own interest in the program lies in its ability to recreate historical routes, using older USGS topo maps.
John is a railroad history buff also. We got to know one another here, on the forums, and finally did meet up last Spring. We found Skype, not long ago, and use that for those long (often rather humorous) Trainz (deep, deep) technical discussions.
Next week we are going out to the northwestern corner of Massachusetts for a week. We'll be tracing the old Boston and Maine RR route from the Hoosac Tunnel to Mechanicville, New York. Any of you Trainzers who live in that area, and may want to meet up, send a pm to John or myself.