Wisdom /Guidance needed.

Dawgvet

New member
So, after a few months of using Trainz-10, I have decided to do my first route. I was really inspired by Dermmy's East Kentucky route so I have decided to do my own "small route" - CSX from Chattanooga TN through Atlanta GA and to the port of Savannah GA. I was fortunate enough to have great help from LWVRR (Scott) and he made the HOG map for me when he kindly offered to do so back in early June. I know there is more than one way to skin a cat, but maybe anybody can add words of wisdom on how YOU go about building a route. Do you finish all of the track first then add scenery? Do you pick two points in the map and completely finish them before moving to the next point? Do you finish a small area at a time then move on...What is the general acceptable area beyond the track to detail? Do most of you use cab view or camera view? Again, I know there is no right/wrong way of doing things and everyone approaches it differently but might help out us beginners to learn different approaches.
Thank you in advance,

Miguel
 
Personally, for my Back to the Future 3 route, I laid the track first, built up hills on either side and placed crosings, a station and the portals. I have placed trees and am now moving on to houses, roads, fences and textures.

So, basically, my approach is:

1. Track
2. Topology (hills, valleys)
3. Basic landmark buildings
4. Textures
5. Finish scenery
6. Splines like roads and fences, small bushes.
 
Thanks for the quick reply Captain.
I'm going to ask questions to learn from people not to question their methods; so, please don't take offense at my questioning. Would not the topology affect where the track is laid down and as such do you find yourself re-adjusting the track after the topology is added?
 
Last edited:
It all depends

Some users start with the goal of replicating as best as possible a real rail line from a real city A to a real city B. So actual terrain (maybe using TransDem), then track, is the top priority.

Other users, such as myself, build a route that is faithful to a specific era and geographical area, but not to a specific chunk of geography and rail. In other words, it's building my own world with a vision of what could be, not what is. So I tend to build small chunks (half a dozen baseboards or so) and complete them before moving on. Also, if I plan on an increase in elevation of the railway, I know that I need to build track that has a realistic gradient and then build the track using the gradient tool and then build the terrain around that - a completely opposite approach from the first example.
 
Last edited:
For fictional routes the terrain is normally adjusted to the track, for prototype routes you're normally following the real track so that gets adjusted to the terrain - which at some point in the distant past, real surveyors mapped out the path the graders and tracklayers would follow. Coolest adventure I had was the Port Ogden & Chicago Metro route for railsim/railworks, I did the DEM terrain from a real world area - Devils Lake, Wisconsin - then since it was a fictional route on real terrain, I had to do my own surveying to figure out the gentlest grade to lay track from one point to another. Lot of bridges, tunnels, cuts and fills involved in that.

My first route was the MSTS Port Ogden & Northern, which was an HO scale bookshelf layout 1 foot by 16 feet. Elevations were in inches, so I did a lot of multiplying by 87 and dividing by 12 then dividing by 3.3 and adding a French postage stamp to convert to meters, eventually got something that was a more or less true scale simulation of the Model Railroader "101 Trackplans" #8 route. Then, of course, it occurred to me that unlike a model layout you don't have to bust down walls and build additional rooms on your house to expand, you have all the room in the world - so I expanded! :hehe: PO&N was the first user created route that wasn't completely flat, and I decided to see how high I could go - Seventy-Seven was originally named that because it was 77 meters above sea level, and I did weird things like a mini - Tehachapi Loop and mini Horseshoe Curve to reconnect coming back down from the top to the bottom again. All the terrain was sculpted to fit the track later.

Fun times, now I'm doing a Chicago route that's almost completely flat, occasionally consulting Bing maps but mostly from memory.
 
I get bored easily and fast - which is why there are so many of my routes unfinished LOL.

To keep interest up though I try to keep my evenings varied. I lay a couple of miles of track then try to completely finish everything before moving on to the next couple of miles. Exactly what order things go in depends a bit on the exact situation, but mostly track, roads, fences to get the general layout right; then ground textures, scenery objects (trees, buildings etc); then detail 'tweaking'.

I like copy/paste so often several of those steps get done at once...

Andy ;)
 
The absolute worst thing to do is lay all the track then find yourself faced with miles and miles of empty tiles to populate. So many route projects I started with great promise, went down the drain as a result of doing that. Anyhow, my recipe, which will hopefully bring success...

1. Research. Before you even start, find out as much as you can about the line and the area it runs through. Is it possible to obtain good mapping and information on track plans and the gradient profile. This is going to be your "baby" for the next several months, are you really that keen on the idea to not get distracted by other interesting routes you might discover in the meantime. (Another reason why these days I generally choose routes that can be built in 8 - 12 weeks).

2. If it's a prototype route then the first stage is terrain with overlaid mapping, pretty much essential. I use Transdem which can handle both and with a fairly easy process in an hour or so you can have a fully DEM'ed and map overlaid set of baseboards in Trainz, ready to start work.

3. So far as the workflow is concerned, start with tracklaying, also maybe add roads and rivers as this will help with the placement of bridges, crossings and other features rather than having to come back and hack things about to add these later on. Maybe lay 10 - 15km of track before breaking off and having a go at the scenery placement - terrain painting and 3D objects, etc. Start somewhere simple, out in the country as this is generally much quicker than doing towns or built up areas. Generally, it takes 3 times as long to detail an urban section as it does remote areas.

Then rinse and repeat until all the 3D placement is done. The best thing about working in sections is if you do hit a creative brick wall, you may be able to still salvage a viable route out of it.

4. Final step is to add signalling and markers, cameras etc. Its best to leave these until last and any slight adjustments to the track have been made.

You can always go through and add fine detail, don't forget to include a session or two to get people started and maybe a short text document describing the route.
 
Thank you for the replies.
My biggest problem is that I tend to lean towards the "anal retentive" side of things and want to finish every little detail. It is hard to do a route and not include the fleas on the dog under the front porch. ( My wife says I need therapy). My route is about 300 miles, not including a few side spurs, and is based on actual TransDem data.
When a route is started, do you decide right away if you are going to use compatibility mode? If you are thinking that eventually you want to share the route is it better to use compatibility mode?
What is the acceptable baseboard distance from the track to detail? I think I read somewhere 2 baseboards?
Can one share the parts finished with others to get feedback/ input on your project? and is it possible to just pluck out a section so you are not sending the entire map?

Thanks,

Miguel
 
My route is about 300 miles, not including a few side spurs, and is based on actual TransDem data.

That's huge - seriously huge. Do-able, but huge. It might be worth cutting your teeth on something smaller - even a small part of the 'ultimate' project that can be merged into the final route. That way you can see the end of the tunnel right from the start. And all route builders need therapy :)


When a route is started, do you decide right away if you are going to use compatibility mode? If you are thinking that eventually you want to share the route is it better to use compatibility mode?

I like billboard trees. IMHO a well made billboard with a really good image beats the pants off a speedtree every time. That's the only reason to stay in 'Comp' Mode. But think about this - with TS10 Comp Mode (presumably SP3 Build 44088) you are building a route which will ONLY work on that one build. Yes you can import it into 12 and later versions, but the billboard trees will look like awful and all your effort goes for naught. Is it worth putting years of your life into something which will ultimately only be used by a handful of Trainzers? I hate SpeedTrees, but I am using the damned things in Native Mode because I don't want to build a route only usable in Build 44088. EK3 simply wasn't worth the effort and the complete hash N3V made of it with their mandatory tree replacement in TS12 cured me of the Compatibility Bug.


What is the acceptable baseboard distance from the track to detail? I think I read somewhere 2 baseboards?

Depends on the route, the type of scenery and the quality of the computer. Play around, work out what works for you and go with it. Nobody else can decide that question for you. If you are a flea-on-the-dog type of guy you need to keep the route narrow or you will bring any rig to its knees. There are tricks though and you will work 'em out as you go :)

Can one share the parts finished with others to get feedback/ input on your project? and is it possible to just pluck out a section so you are not sending the entire map?

Sort of, but the only way to make a route smaller is to delete boards. I prefer to start small and work up with sub-routes which can be merged as I go. My Clovis 3 will include all of C1 and C2, but there is no (easy) way to just distribute the C3 part of the route without the C1/C2 bits...

Andy ;)
 
Last edited:
Dermmy wrote:
That's huge - seriously huge. Do-able, but huge. It might be worth cutting your teeth on something smaller
-

Living life at the edge :p

Dermmy wrote:
Depends on the route, the type of scenery

Scenery very similar to your EK3 but with a couple of larger cities.
---

I usually test drive the track to look for problems. Is there a way to mark a spot when you find a problem and all you have is bare baseboard (minimal reference points) and you've been driving for miles.
 
I don't mean to Hijack the thread, but maybe add to the discussion which is I assume building a route from real data(maps, DEMs etc,). Dermmy, love the East Kentucky maps, been playing them for a long time. How did you get the elevation data for the EK maps? Seems like the UK guys can just download stuff from a free repository, whereas we just have google maps, earth, etc here in the US. Are you using transDEM? Or did you use the Basemap method?
 
How did you get the elevation data for the EK maps? Seems like the UK guys can just download stuff from a free repository, whereas we just have google maps, earth, etc here in the US. Are you using transDEM? Or did you use the Basemap method?

All my routes are DEM based via HOG. All required data free. There is a tutorial for US users here. It is a bit out of date, but can still be worked through...

Andy
 
300 miles is definitely pushing it. The longest route I ever built was around 170 miles (in MSTS) and that was a gruelling, punishing task by the end - particularly when it looked as if it might have got corrupted and months of work down the drain. (Luckily I salvaged it).

I would definitely go for building in more modest 50 mile or so stages. 300 miles is probably getting near the limits of what the software can handle in a single route anyway, particularly when it comes to saving and using copy/paste etc. in Surveyor. Build it in sections and power users can always join/merge these to give a contiguous whole - similar process to what the bloke doing all those Australian NSW routes has done.
 
Thank you for all the replies.
Even though I would have liked to keep this route as a single entity, Vern I see your point. I might have to split it in two. Never gave it a thought that the total size of the route could put a strain in some computers. I always thought it would be more what it is loading per frame that would be more important.

Miguel
 
You are right Miguel, the size of the route per se won't put any extra load on the computer. A route can (in theory) be as large as you want, subject only to the largest file Windows can open. The 'load' on the system comes from the content used on the route, or more specifically the amount of content the system needs to load to draw the current screen.

There are however two problems with large routes. The first, already mentioned a couple of times in this thread, is routebuilder burnout and the extent and frquency with which it strikes is pretty much up to you!

The second problem though is the load placed on the system not be the route itself, but by the session. Keeping track of the number of trains running on a large busy route will tax the system more than the graphic load of drawing the current screen. You need to keep in mind the combination of the two. A 300 mile route through a quiet railroad backwater with only a couple of trains will be no problem. My Clovis Sub v2 runs to about 120 miles but features the BNSF TransCon route. Running a single train on the route is no problem, but trying to run a more-or-less accurate 100 trains-per-day session brings my rig to its knees, it is in fact impossible. I plan to build a new computer 'soon' and the 120 mile route will hopefully be able to run to its potential, but when I get to Clovis5 and the entire Clovis Sub is done I will be back to the 'nice route, but can't run it' problem.

There's a lot to think about...

Andy ;)
 
Route Construction

I am presently working on describing the approach I have taken to constructing "The Canadian Pacific's Mountain Subdivision in the mountains of British Columbia, Canada. Its only one approach. I wasn't familiar with TransDEM when I started so took this different approach. I haven't completed the description yet but for what it is worth you can have a look at it here:

http://homepage.mac.com/doug56/MBC/
Tutorial>Route Construction

Cayden
 
Miguel,

My route is fictional, and I use a combination of approaches. There are areas that were done really early in its development and were built around flat baseboards with lifted and lowered terrain - a la model railroad style building. Over the past years, since this route has been with me since late 2003/early 2004, I've replaced baseboards with displacement maps and added in sections made from DEM files. This is what's cool about Trainz. You can do this without losing anything but time and a few electrons because there is no wasted materials. You didn't kit-bash a building for a specific location, and end up trashing it because there's no longer any use for it, or worse spent a ton of money on some turnouts, only to rip them apart when you decide to remove the track and redo the junction again!

Anyway, seriously heed the warning here, don't bite off more than you can chew. I work in little sections. I have one area are from a Dem file that covers Portland Maine to Berlin New Hampshire, which is in the White Mountains. The southern area in Portland only has some track on it at this point. The rest is still TIGER lines and bare baseboards. Currently I'm fiddling around in Portland, put in Rigby Yard, and did some other work. It's not necessarily real, but not unrealistic either. I look at Bing and Google maps, then put in what I want rather than being true to the rivets because lots have changed since the TIGER lines were created, which track that's still in use, and how the waterfront has changed.

The other thing too, as Sniper mentioned, using a real DEM file and then surveying your route. This too for me was a bit of an experiment, and well worth the efforts. I took the DEM file of the Bangor to Bucksport Maine and integrated this into my route. I then built a branch line where none ever existed. The Sandy Point Branch was laid out using eye-level surveying methods.

I got into walk mode and walked the landscape, sighting out places to put in cuts and grades so the route could make it up over the hill and on to Sandy Point. For markers, I used trees or buildings. I'd escape out of walk mode, place a tree or building, then survey some more.

The results are quite amazing and the effort was worth it. The line integrates well with the rest of the route, and no one would ever know, unless I told them, that Sandy Point is a fictional town in Down East Maine.

Remember this is a creative process, which can take a lot from you mentally, and can burn you out physically. So, as mentioned before, don't let yourself get burned out. This can become a grueling, long process, as while you lay tracks and place buildings and trees. By all means don't feel like you have to rush the building process. I stop my construction for months or weeks, then come back and work furiously until I run out of brainpower and steam.

John
 
You are right Miguel, the size of the route per se won't put any extra load on the computer. A route can (in theory) be as large as you want, subject only to the largest file Windows can open. The 'load' on the system comes from the content used on the route, or more specifically the amount of content the system needs to load to draw the current screen.

There are however two problems with large routes. The first, already mentioned a couple of times in this thread, is routebuilder burnout and the extent and frquency with which it strikes is pretty much up to you!

The second problem though is the load placed on the system not be the route itself, but by the session. Keeping track of the number of trains running on a large busy route will tax the system more than the graphic load of drawing the current screen. You need to keep in mind the combination of the two. A 300 mile route through a quiet railroad backwater with only a couple of trains will be no problem. My Clovis Sub v2 runs to about 120 miles but features the BNSF TransCon route. Running a single train on the route is no problem, but trying to run a more-or-less accurate 100 trains-per-day session brings my rig to its knees, it is in fact impossible. I plan to build a new computer 'soon' and the 120 mile route will hopefully be able to run to its potential, but when I get to Clovis5 and the entire Clovis Sub is done I will be back to the 'nice route, but can't run it' problem.

There's a lot to think about...

Andy ;)

Exactly. I've been an armchair model railroader for longer than I've cared to be and the great lure of railroading on a computer versus physical models is all those barriers that restrict real model railroad layouts are gone. Physical models these days are pretty awesome, but they are very expensive and still suffer from problems with track work, dirty track, coupler faults, and heaven help you if a $500+ model hits the floor.

Doing this stuff on the computer is relatively cheap. Even if you budget $2000 every few years to have the latest and greatest hardware it's still a drop it the bucket compared to doing an actual layout.

However, the computer is really not unlimited, its just that the limits are different than classic model railroading. I'm finding that selective compression still applies. This skill is difficult to implement, simplifying the model, but still keeping the prototype feel is tricky. However, I think it is still a "necessary evil".

I'm having a ball with Mojave in TS12, but I'm not using it in stock form because the frame rate hit is just too great. One trick is to shorten the trains. It works surprisingly well. Take the Mojave to Bakersfield session. Start with your train and chop one lead locomotive and take the pusher off the rear. Then cut 25% of the cars. Do the same thing to the tank train and all the other trains in the session. The result is much better frame rates and while the trains are shorter than the real ones they are still quite long by model railroading standards.

One tools that I wish existed was some way to predict the performance of a route before it is built. I suspect that many of the beautiful, but poor performing routes exist because during development they were fine but as they were finished the frame rates tanked with no easy was to salvage the work.

Getting more back on the point that Andy made, the constraints of building one of these routes has shifted from floor space and money for track, bench work, scenery, locomotives, and rolling stock to CPU and GPU rendering power. The bad news is there is still constraints, the good news is the constraints have been pushed back considerably.

I spent a glorious weekend in the nineties in Needles, CA watching the ATSF and haven't felt "there" until TS12 and the Mojave subdivision. I don't think there is a physical model railroad that could do that - and this was with my "short" trains.

-Mike
 
Selective compression is needed for other reasons, main one being time. With a model railroad you trade in the top loading washer for a front loader so you can build a yard over the laundry area, make a removable bridge so you can have access to the furnace filter, move the water heater upstairs to the kitchen and you still don't have enough space to make the C&NW all the way out to Iowa.

With a train simulator you have all the room you need, but do you have the time to run trains all the way from Chicago to Omaha? Even if you managed to build and scenic the whole line, driving it would get boring rather quickly. That's why I always prefer fictional routes, rather than 50 miles of mainline with one industry spur then another 50 miles of mainline I can toss in however many yards and switching areas per mile I want. :cool: It's a game, not a training device, playability takes priority over reality any time reality gets in the way.
 
Back
Top