RGE Inaccurate

Right, I've found out the long and lat coordanates for my start point.My Start point is Watford Junction which is in Hertfordshire, just north of London for those that don't know.Coordatnates are 51:40:11N 0:23:47WWhen I look for them in the RGE, I'm hovering over an area that is more like Surrey, which for those that don't know is South west of London and about 100 miles away !!!Am I suposed to ignore the map, or the coordanatesThanks Steve

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

RE: RGE Inaccurate, LawrenceKellie, Sep-12-02, 03:40 PM, (1) RE: RGE Inaccurate, sph70, Sep-12-02, 03:51 PM, (2) RE: RGE Inaccurate, dcm, Sep-12-02, 03:55 PM, (3) RE: RGE Inaccurate, sph70, Sep-12-02, 04:13 PM, (4) RE: RGE Inaccurate, dcm, Sep-12-02, 04:18 PM, (5) RE: RGE Inaccurate, 3footer, Sep-12-02, 05:43 PM, (6) RE: RGE Inaccurate, LawrenceKellie, Sep-12-02, 05:52 PM, (7) RE: RGE Inaccurate, sph70, Sep-12-02, 06:02 PM, (8) RE: RGE Inaccurate, rjdiii, Sep-12-02, 06:20 PM, (9) RE: RGE Inaccurate, timken, Sep-12-02, 07:59 PM, (10) RE: RGE Inaccurate, davidbeach, Sep-12-02, 09:04 PM, (11) RE: RGE Inaccurate, dcm, Sep-13-02, 07:39 AM, (12) RE: RGE Inaccurate, davidbeach, Sep-13-02, 03:45 PM, (13) RE: RGE Inaccurate, dcm, Sep-16-02, 09:30 AM, (14) RE: RGE Inaccurate, rdamurphy, Sep-16-02, 03:05 PM, (15) RE: RGE Inaccurate, davidbeach, Sep-16-02, 04:09 PM, (16)

Conferences | Topics | Previous Topic | Next Topic

Welcome to the world of the RGE.First, each tile that you select is approximately 2x2km. Therefore, there is some room for lat/long variations. Second, you can always add or delete tiles at a later date. So if the first time you choose wrong, you can add/delete as needed.Lawrence

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

But Lawrence, it's way out......
What should I select, the coordanates or just roughly where i think it is on the map.
It makes me wonder what all these people are doing with dem data.
100(ish) miles is a long way, so if you use dem data the terrain aint gonna be in the right place is it.(?)
Are the lat and long coordinates really neccessary..?Steve

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

Hi Steve,The geographical outlines in the RGE are not that accurate (as you have seen). But you have raised a good question. For a long time I was a proponent of the 'It must be spot on geographically accurate' argument - which is one of the reasons why I built so much accuracy into TsTools. However, my answer to your question these days would be ignore both.My reasoning is based on the projection system used in MSTS (I won't bore with the details here, I have done so at length in other forums). The bottom line is that in order to represent a curved surface in a flat way, you have to 'project' it. To visualise that, imagine the world as a large spherical lamp shade with the continents drawn on it. When you switch the light on, the image of the continents is projected out onto the walls of the room - you have a projection onto a flat surface - but you will notice that the shapes and proportions do not resemble those as seen on the original curved surface.There are a huge number of projection algorithms around, each has its pros and cons, and each is a compromise of the true shape of what is being represented. Depending upon the task at hand, and the extent of the area being mapped, you may need to choose a system that preserves distance relationships at the cost of directional relationships, or preserves shapes at the expense of distances - etc. etc.Basically, imho, the projection system adopted in MSTS gives us the worst of all worlds for the task it is required to perform. If you were to accurately plot the course of a track in geographical co-ordinates and then view it in MSTS (projeced in a Goode projection) and compare it with a map projected in Transverse Mercator, the angles and curves and distances would look obviously (too obviously for my taste) different.Although, as I said, I fly the flag for accuracy, I do not believe that you achieve anything like a real world experience by building MSTS routes based upon real world Lat/Lon coords.During chats with TsTools users in another forum, I came to realise that what simmers really want are routes that 'feel' the right shape -and it doesn't matter whether the lat/lons match the real world or not. I have implemented in TsTools (and I believe that Zorcan also has done this in TSTF) the ability to place your route so that it will appear in MSTS in the same projection as the map you are working on. Basically, you position yourself in roughly the correct starting place but work with a TM projected map to trace the path of your route.At the end of the day it doesn't really matter if Watford ends up somewhere off the M23 (in a true lat/lon sense) since you will be building the landscape around it and the user driving the route won't be checking his in cab GPS system for his location (I hope!).Don't know if that helps - but it gave me chance to get it off my chest CheersDave C

Dave Charles
developer, TsTools
http://ts-tools.com/

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

Thanks Dave, I sort of got most of that.
So basically lat and long are not needed, you could build a route from London to Reading using tiles selected from north east Scotland.
If that's the case, what is the point of the Route Geometry Extractor..?Thanks again Steve

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

gives you that nice warm feeling that life, the universe and microsoft actually have some bearing on our perception of reality sorry - i think i need some more coffee.Cheers,Dave

Dave Charles
developer, TsTools
http://ts-tools.com/

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

>If that's the case, what is the point of the Route Geometry
>Extractor..? Presumably it was useful for the game developer in extracting DEM data (they weren't using "demex"). I use it to add/delete tiles from my route and set the start tile. After getting the tiles added in the general area, I used demex to load the DEM terrain. Then the lat/lon went out the window, and I just correlated (by eye) the terrain to my topographical maps (and memory). It helps to have a mountain railroad with recognizable features. A couple of times I ran out of terrain so went back to the RGE to add a few more tiles and demex'd them.
ROW and elevation maps are a BIG help.

3footer

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

>If that's the case, what is the point of the Route Geometry
>Extractor..? Try creating the tiles without it..... Lawrence

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

Yeah point taken
What i realy meant was why don't they just give you a blank grid to select tiles from.
Don't need a world map.steve

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

What I still don't understand is why the projection thing is such a big deal anymore anyway. I can understand it for making flat maps but there is no more a such thing as flat than there is spherical inside a computer. From a computer's perspective Lat, Long and elevation (or whatever your reference dujour is) are only points in 3d virtual space and I don't understand why it is so much more difficult to describe them in spherical terms than flat, cubical, trapezoidal, whatever 3d space. I would imagine that now and henceforth this data will be generated/collected from satellites in either circular or eliptical orbits or high flying aircraft which are themselves flying in an arc to maintain a steady absolute altitude and that translating them to flat transforms should only be done by end use applications. Cartographers should quite trying to fix a problem that computers and their virtual space have made a little irrelevant.Richard

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

Strange, I have the DEM topography for my Sumpter Valley line, and I can get decimal long. lat. data from from Streets USA. If I use the data from where the highway enters the Powder R. canyon, it puts me spot on, with exactly the view I'd expect. I live 3 miles from here, so I know it's right. I can also plug in the data for the base of the Mason dam, and end up looking up at it. Maybe I just got lucky.

Steve T.

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

Steve, et. al.,If DEM data is used to generate the terrain, the lat/long has to be exact. If you are going to extract tiles and hand sculpt the terrain, you could do the Swiss Alps in the middle of the Pacific Ocean.When the terrain for the Sumpter Valley was created, I knew where it needed to be and set the tiles up there, don't know if the RGE map shows it to be in exactly the right spot or not, don't even care. I just needed the tiles that included the lat/long boundaries of the area to be modeled.Once the terrain was created, of course you will find everything exactly where you expect it, because the DEM files and the maps you are using all trace back to the base, raw, USGS data. There are some distinct advantages to living (or at least modeling) in the United States.What you won't find though is a way to run a north heading along a line of constant longitude. Try it in the RE, you can either move the camera north (angle 0) or you can move it north (constant longitude) but you can't do both at the same time. So what is north?David Beach.Palouse & Lewiston

1% Completed

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

Ah, but it is not irrelevant. AS has been said, if you use DEM data the lat/lon will be spot on, but my point is that in converting lat/lon to ground co-ordinates (whether to display on a computer, a map, or wherever) you need to do some kind of conversion to distances rather than degrees. If you simply plotted degrees as your coordinate system you would soon realise that the distance covered by one degree of longitude is very different to that covered by one degree of latitude (unless you happen to be on the equator) and, even more fundamental to this, the distance covered by one degree of latitude decreases the further you are from the equator.So the problem remains, whether we are using computers or paper - we want to know the co-ordinates in terms of distance - we want an x scale where a change in value of 1 is the same distance as a change in value of 1 in the y or z dimension, and we want a y scale where a change in value of 1 is the same distance wherever you are along that scale.If we just use degrees then this is impossible - we HAVE to project in some form or other. Projection is simply the process of trying to convert lat/lon into actual distances on the ground which we can use for our co-ordinate system, whether it be in a computer or on a map.My point was simply that the projection system used in MSTS is not to my taste, I would have preferred something closer to a TM system which gives a very good conversion from degrees to distances over the kind of scale and areas that we are concerned with in MSTS, and thus a more real look to the world in the sim.This is especially important when you are using DEM data, for exactly the reasons stated above - the landscape will all be at exactly the right lat and lon - in a TM system this would (IMO) 'look' more like reality than it does in the Goode system used in MSTS.CheersDave C

Dave Charles
developer, TsTools
http://ts-tools.com/

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

>If you simply plotted degrees as your coordinate system you
>would soon realise that the distance covered by one degree
>of longitude is very different to that covered by one degree
>of latitude (unless you happen to be on the equator) and,
>even more fundamental to this, the distance covered by one
>degree of latitude decreases the further you are from the
>equator. Actually, the distance covered by one degree of latitude remains fairly constant from equator to pole. It's the distance that is covered by one degree on longitude that varies. In extremely rough numbers, assuming 25,000 miles around the earth at the equator, and assuming a spherical earth, at the equator one degree of longitude = one degree of latitude = 69.5 miles. At the north pole, the distance from 89 degrees north to 90 degrees north is still about 69.5 miles, but all the lines of longitude come together, so the distance from one to the next is zero.David Beach.Palouse & Lewiston

1% Completed

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

>Actually, the distance covered by one degree of latitude
>remains fairly constant from equator to pole. It's the
>distance that is covered by one degree on longitude that
>varies. In extremely rough numbers, assuming 25,000 miles
>around the earth at the equator, and assuming a spherical
>earth, at the equator one degree of longitude = one degree
>of latitude = 69.5 miles. At the north pole, the distance
>from 89 degrees north to 90 degrees north is still about
>69.5 miles, but all the lines of longitude come together, so
>the distance from one to the next is zero. Quite right David, that is what I meant to say, just a touch of brain-error creeping in.cheersDave C

Dave Charles
developer, TsTools
http://ts-tools.com/

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

I don't know if anyone will read down here far enough, but you are using "english" lat/long measurements, the RE uses "metric" You must convert from the standard hour/minute/seconds to the equivalent metric measurement.Rob

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

It's not metric as such, just decimal. The RGE uses decimal degrees rather than D:M:S. Much easier for a computer to keep track of and much easier to do computations with. Easiest thing to do is enter the D.MMSS value on your trusty HP scientific calculation and then use the To Hours function. To go from decimal degrees to D.MMSS, just use the To HMS function. Other wise, take the number of seconds, divide by 60, add the number of minutes, divide by 60, add the number of degrees. To go in reverse multiply the decimal portion by 60 to get minutes and then again by 60 to get seconds.David Beach.Palouse & Lewiston

1% Completed

Alert | IP

[ Home | MSTS Forum | MSTS FAQ | File Library ] Copyright © 2002 by Nels Anderson. All Rights Reserved.