Atomic Systems

I'm quite sure I can't be the only one who has run into this problem. With RWDecal and DEM data you can create a world that comes very close to the real one. My biggest problem so far however is how accurate I want to lay tracks and switches. I've been working on a railroad yard/station and have spent quite a bit of time tweaking the exact location of the switches, tracks and platforms. If possible I'd like to make it fit the real world 100%, but that would take up far too much time. My question now is, how accurate do you work? There's a difference in a switch being 20 meters out of place or 150 meters. Wouldn't it be easier just to get the overall shape and size right and take a bit of leeway with the exact locations?

Regards,
Tjeerd

It depends ..

The Spa Valley Railway might consider using my route to help with the location of trackside equipment and other features as this route is placed reasonably accurately in the real world and the GPS locations in my route are very near to the actual route so there is a good bonus there from accurate placement.

Generally, however, limitations in RW will not let you create a truly accurate representation of track but in most circumstances it is not really necessary to have that much accuracy as the average user will not know the difference. A general feel is often enough to give the user the illusion of accuracy.

My attitude is "If it looks right, it is right!"

I usually try to aim for 90 percent, sometimes its lower, sometimes its higher. You cannot accurately reconsruct something the size of Temple meads and not make compromises. The most obvious I can illustrate is that its impossible to make slip points on a bend (sometimes not even crossovers) yet the GWR often did. What you then do is do it as close as you can get to straight, render the points, then put the curve in afterwards. That in turn means the bend is slighter than you wanted, which either means you put in a tighter bend to get where the track ought to be, or carry it through and keep it natural. The latter can often take up a lot of space, former looks unrealistic. You have to tread a thin line between the two.

Similar problems occur with carriage sidings. It had a curve, but physically fitting the manual points in as close proximity to reality was impossible. So I slimmed down from something like 21 sidings to 17-18. its difficult to see the difference passing by, and hard to conceive of a situation when 21 sidings would be necessary. If it ever is, and we all end up using quantum light speed PCs, and fully funcitioning AI shunting engines, I can always relay.

There are always compromises like that. As long as you try to get as close to a gradient as you can, paint in the scenary as well as feasible, and try to get signalling as close as you can find details for, I dont think anyone has a real problem. Track is the real limitation here. In real life, you can dense pack junctions to an amazing degree (look at clapham junction). In Railworks, you do have to make compromises because the track splits and signalling. A shame, but I prefer to think of it as a feature by RS.com to keep us sane.

RW decal is a major step forward in terms of accuracy. At least as far as scenery is concerned, it gives you the ablity to get close to 100 percent accuracy. It was a real game changer.

Hi

I'd agree with all that's been said above. I aim to lay track 100% accurately, but it's often not possible.

There have been a couple of occasions with loco/carriage sheds where I've had to leave out a siding so that I can fit in default scenery around it so it looks 'right', as eyore said. Conversely I've also added in some crossovers that shouldn't really be there to make the AI operations easier to set up in scenarios.

Cheers

Bob

hertsbob wrote:Hi

I'd agree with all that's been said above. I aim to lay track 100% accurately, but it's often not possible.

There have been a couple of occasions with loco/carriage sheds where I've had to leave out a siding so that I can fit in default scenery around it so it looks 'right', as eyore said. Conversely I've also added in some crossovers that shouldn't really be there to make the AI operations easier to set up in scenarios.

Cheers

Bob

Yes indeed. In the Spa Valley Railway there is a link to the main line that does not actually exist, but this was necessary to get the main line AI to work.

On a modern route, I try to be as accurate as possible with the track work. It's not easy to determine the exact spot to start a point, and sometimes satellite imagery as distortions, so it is not possible to get it 100%. Gradient changes near pointwork are usually best avoided as it can be difficult to get them to render. On historical routes it can be difficult to obtain complete track work information, and the shear complexity of the old trackwork will lead to more compromises needing to be made.

stuart666 wrote:The most obvious I can illustrate is that its impossible to make slip points on a bend (sometimes not even crossovers) yet the GWR often did. What you then do is do it as close as you can get to straight, render the points, then put the curve in afterwards.

They are not impossible although getting a weld point or correct render can be hit or miss. One trick I have found to get slips to render correctly is to split the track on a curve either side of the crossover, weld the slip and then re-weld the track splits.

Well there IS that way of doing it. But when you have multiple slips going across 4 tracks, or worse, 2 tracks of slips going across 4 tracks on a compound curve, it gets very very difficult to do it like that. In the end you start screaming in dispair, use dead straight track and take a chill pill. Or at least, thats the way It happens for me.

None to RS.com, if there is EVER a way to make it feasible to drop premake slips, crossovers or pointwork which can the be joined onto existing track, please do so. You will be on my Christmas card list forever.

I find getting a large yard to render correctly nigh on impossible say 12 road yard to go into 1 shunting spur in a realistic length you end up with 3 way point and bend/ blending to make the yard LOOK right but not strictly prototypicaly accurate, as mention many time By Vern make it look right and most people unlesss they have driven or walked a route often will not know the difference if a gradient is not quite as it should be indeed it is very hard to make a route exact, As where is your start point Just because it, say starts at a junction on the level, that "level" may not be what your DEM data sees as I'm sure DEm will see building and roads and large groups of trees as land so you will never be 100% the newer OS landform data looks more accurate on 1 bit i have loaded, and compared to the old NASA data widley used in the past, and appears to have more definition. So time will tell, but above all remember its only a Game and have Fun!, oh and if you build it share it and let others play with your train set too

I agree make it look about right rather than 100% accurate. Two routes currently WIP, one in Germany where reference to Sporenplan web site and the satellite imagery shows vast yards and complex pointwork, the like of which only exists in a very few UK locations these days. Modelling that alone would probably use up the 2-3k tile object quota. So there will be compromises, maybe even to the extent of missing out one of the yards altogether.

My other route, currently taking priority, is a little third rail project based on Bognor to Littlehampton. At Barnham west end there is quite a complex junction with several diamonds (no slips). I have been able to get this to render quite wel but compromise is a very short (about 10m) length of 70m radius curve which of course strictly speaking is too tight for UK mainline use. However it shouldn't be an issue driving over in the sim and amidst a sea of pointwork will hopefully go un-noticed. One thing that has improved accurate track laying is the ability to lay offset track. Both Bognor and Littlehampton are on very slight curves. Before it would have been a real PITA to get this looking right, but by progressively offsetting platforms then track it is now very easy compared to doing it all by eye.

If anyone would like to do RWdecal from Kirk Micheal to Ramsey on the Isle of Man and overlay it onto my Manx Northern Line you'll see how inaccurately it follows the old route however I showed my Dad, who know's it like the back of his hand, and he was none wiser.
The way it looks is the key

Thanks for the replies everyone.