.
Trainz: A New Era Pre-Orders Have Now Arrived! - Visit our Information Page to find out more.
Results 1 to 14 of 14

Thread: It's (almost) 2014: digi-railfans deserve a "World of Rails"

  1. #1
    Join Date
    Nov 2006
    Location
    United States of America, Georgia, Atlanta
    Posts
    97

    Default It's (almost) 2014: digi-railfans deserve a "World of Rails"

    OK, first of all, I'm in. Senior Engineer/Early Bird, though I would have certainly liked to buy a seat on the Pioneer Council or the "Enthusiast" group. This December 17th is just not a great time for a much bigger investment from me, though.


    Honestly, it'd be easier to be a cynic and not to invest, and here are my personal reasons why:

    • Installing SP3 was so terrible on my performance - constant stutter - that I finally removed Trainz entirely after spending dozens of hours trying every piece of advice on the forums
    • Trainz is getting its tail kicked in both realism and smoothness of gameplay by a certain other railroad simulator
    • N3V is just using Kickstarter cash to build the graphics engine they would have built anyway, then sell it in larger volume to the non-backers
    • The features in the new game engine won't improve my enjoyment more than I could get from the aforementioned competition AND, based on living with stutter throughout my Trainz experience, have inherent risk of crippling performance

    But, I put the majority of my cynicism aside. I've always had a soft spot for Trainz, and I always want to see it get better.


    So, where can we look for the next real differentiation, rather than just geeking out on graphics and detail? In my opinion, Trainz still has a chance to beat the competition to the "World of Rails" that Microsoft proposed in "TS2" around '07.


    I want to see Trainz be the first big train sim to have built-in tools for procedural world generation. We should be able to specify an area of the planet, have Trainz generate the rails based on public domain maps, and then have auto-gen buildings and foliage that match the approximate environment. This should be user friendly and nearly automatic. Seven years or so after Microsoft proposed it, we digi-railfans should be able to drive anything anywhere in the world with a few clicks.

    This doesn't have to harm anybody's custom content or ability to create routes completely separate from the "World of Rails." Everyone would still want to dive into the microcosm of specific routes.


    I know that's possible. Some company WILL do it eventually, and I'd rather it be N3V than the competition. That's why my stretch goal vote will not only be for # 8...


    Landscape Genie - a configurable tool to auto-create and paint terrain and populate with selected objects
    Import DEM data directly with the Terrain Importer
    Google maps overlay or similar system


    ...but also to not give up on Trainz until it can live the "World of Rails" vision. Unless someone else does it first.

  2. #2
    Join Date
    Nov 2006
    Location
    United States of America, NY, Albany
    Posts
    763

    Default

    Quote Originally Posted by jayinatlanta View Post
    • N3V is just using Kickstarter cash to build the graphics engine they would have built anyway, then sell it in larger volume to the non-backers
    I think you under estimate how much a up to date game engine cost these day's. The $165,000 goal is just a fraction of the cost that other's have spent on game engine development.




    Quote Originally Posted by jayinatlanta View Post

    • The features in the new game engine won't improve my enjoyment more than I could get from the aforementioned competition AND, based on living with stutter throughout my Trainz experience, have inherent risk of crippling performance
    That's the very reason to get it moved onto a new platform (game engine) to begin with, better system resource utilization to avoid the performance issues that plagued past versions of Trainz. A new game engine means a lot more than just eye candy (graphics).

  3. #3
    Join Date
    Nov 2006
    Location
    On The Edge Of Forever
    Posts
    2,513

    Default

    It's a good idea but I personally have the feeling MS bit off more than they could chew with the idea and, like Flight Sim, outside of the detailed areas they chose to represent would simply have amounted to very basic infrastructure traversing generic scenery. So while I agree in principle, I don't think it's quite what we're going to see with TANE, however better handling of real world data (projection, native DEM import, Google Earth overlay) to put it on a level playing field with TS2014/Railworks would be a start.

  4. #4

    Default

    Quote Originally Posted by jayinatlanta View Post
    Trainz is getting its tail kicked in both realism and smoothness of gameplay by a certain other railroad simulator
    The other railway simulator is very enjoyable, I use it myself, but it is only a driver simulation. If you want to simulate a railway, Trainz wins hands down.

  5. #5
    Join Date
    Nov 2006
    Location
    Australia, Victoria, Tecoma
    Posts
    3,161

    Default

    Personal words here!!

    Considering the amount of work I've had to do to correct the track locations based off fire service maps which are overlayed onto my DEM's, let alone general maps, and 'world of rails' system is going to be extremely basic... You'll also find that, unless you're going to spend a LOT of money with the rail companies, you aren't going to have much more than single or double tracks with basic junctions. You'll have no gradient data as well (99.99% of the DEM's available don't have the resolution to give accurate track heights, you need to do this yourself), so your track will just float across the top of the terrain...

    Add to this the fact that DEM's for a lot of areas are either unavailable or are very low resolution, and you'll have a lot of errors. You also only need a 5m shift in the map to make everything look stupid (e.g. that 10m deep cutting could become 30-40m deep or need to be changed to a tunnel!). I've had to have DEM's re-done (a fellow creator makes the ones I use for me ) because they're out, because the track suddenly ends up higher (or lower) on a mountain.

    So you'll essentially get spaghetti across mountains, with a bit of basic scenery (based on what the area has...) over low-res terrain that hasn't had track formations created (or if auto-created, wouldn't be what you actually want).

    Note, this is with the available information right now. US modelers may be lucky with higher res DEM's that include the track formation, and 3D tiger/vector data (last I checked, this doesn't exist ). But from my own experience, building Australian routes, automating it simple isn't viable. You need to put in the hard yards to get a useable route.

    Regards


    Zecrail DLS Content
    All comments are my own unless otherwise stated.

  6. #6
    Join Date
    May 2011
    Location
    United Kingdom
    Posts
    1,213

    Default

    I tend to agree with Zec on a lot of what he has said. Just taking my own town for example. The UK does not have the best of scale satellite data. The track through our town is not complex now compared to what it used to be (its just a double track splitting at the junction). But even using the data supplied by OS leaves a lot wanting. We have track going up when it should be going down and visa versa. Without the gradient data its near impossible to get it right. If/when the stretch goal which includes the DEM data is done then it will be a starting point. Though I doubt at this point how much - if any - of an improvement it will be over Transdem. Being able to pull in the google earth overlays will be great but again it will be a starting point not the finished product. As for the buildings etc? For me Id prefer to add those myself - particularly as what I want to do is create a map of my local area as it was say 50 years ago not its current layout. So adding the buildings in automatically for me would be a waste. I guess my 6 year old grandson would approve of it though

  7. #7

    Default

    Quote Originally Posted by S301 View Post
    Considering the amount of work I've had to do to correct the track locations based off fire service maps which are overlayed onto my DEM's, let alone general maps, and 'world of rails' system is going to be extremely basic ...
    Zec said it all.

    I am dealing with this geo data stuff for over 12 years now. Many things have improved in this time, and data in general has become more detailed and easier to access. But accurate track geometry as 3D vector maps are assets quite unlikely to ever appear in the public domain. Where it exists, it is private and confidential and you have to sign NDAs if you are allowed to use them within a project contract.

    Adding to the problem of accessibility is the problem of the data format. GIS and digital cartography started in the late 80s/early 90s with a vast number of parallel approaches, every approach striving to reinvent the wheel. Today we seem to have a Babylon of data formats, with individual usage often stretching the limits which makes it even worse.

    Then, as we are talking cartography here, there is also the map projection issue (round globe vs. flat map).

    From my experience this means that three different railway companies will have three different map projections and three different data formats. Well, some of the features you see in TransDEM today are the result of that frustrating experience.

    In my humble opinion, a "world of rails" deserving the name is reaching far beyond the scope of a railway simulation game.

  8. #8
    Join Date
    Nov 2006
    Location
    Australia, Qld
    Posts
    1,035

    Default

    Quote Originally Posted by geophil View Post
    In my humble opinion, a "world of rails" deserving the name is reaching far beyond the scope of a railway simulation game.
    Good to hear since we chose this name for our mor ecasual upcoming title on Steam.

  9. #9

    Default

    Quote Originally Posted by Tony_Hilliam View Post
    Good to hear since we chose this name for our mor ecasual upcoming title on Steam.
    ... giving it a slightly different meaning from what the OP had in mind.

  10. #10
    Join Date
    Nov 2006
    Location
    United States of America, Georgia, Atlanta
    Posts
    97

    Default

    I appreciate the thoughtful replies.



    I should also mention that Xplane did trains in Xplane 10.


    And again, I want to be realistic here: those are two excellent simulation developers (Microsoft & Laminar Research) which at least started down the road of the "world of rails" concept.


    I don't expect to see this in T2, but I put up my money and my votes anyway, because I think transport simulation developers need to move TOWARDS it!
    Y'all are right.
    - It's a big bite to chew.
    - The rail networks would be spaghetti in many places.
    - The detail would be insignificant or wrong.
    - Six year olds would find it a whole lot more enjoyable than the rest of us, who would quickly want to take what the sim generated and expand it (like the FSX community does even from a 1000 foot in the air view).
    - Tracks wouldn't be available in a lot of places and you'd have to add them, and add or modify the autogen scenery and foliage.


    But it'd be a significant start, and I'd still love for N3V to be the ones to do it before anyone else.

    And djt, I could have been more clear. Insert the word "help" in the phrase "to build the graphics engine." As in, N3V has already started building this engine. They discuss that they have the base architecture. Early adopters like me/us are providing a chunk of up-front cash they wouldn't ordinarily have until market release. I HOPE you're correct that the new engine won't have the recurring stutter problem I've seen in Jet for years.

    I can only HOPE it's not just more eye candy with performance detriment on their "64 bit DX11" system spec.

  11. #11
    Join Date
    Nov 2006
    Location
    United States of America, Tennessee, Soddy Daisy
    Posts
    1,193

    Default

    Sorry, but I see this as a hinderance to route creation for those of us who model in years much prior to the 21st Century. If you were to go to a single globe system like FSX, it wold be strange to have some areas modeled in 2013 and some in 1941, and others in the 1870s. The current format provides the best option for developers in my opinion. I believe there's a reason that Microsoft did this in FSX and that has much to do with the fact that you should be spending most of your time at 14000 ft and the rest on the runways and taxiways of the airport. The ground detail in between is not so important in other than the few areas occupied by airports. There's so much detail that would be generalized that its unreasonable to even compare, in my opinion.
    The Southern, Serves The South

  12. #12
    Join Date
    Nov 2006
    Location
    United States of America, Georgia, Atlanta
    Posts
    97

    Default

    Quote Originally Posted by SR6900 View Post
    Sorry, but I see this as a hinderance to route creation for those of us who model in years much prior to the 21st Century.

    That's really too bad, though I was afraid of that type of reaction.

    If I'm intensely interested in a certain place or era, would I personally prefer to be able to spend months to model an extremely accurate route (as I have done in the past)? Sure. And to be clear, even with the ton of work required for extreme accuracy, Trainz still makes this infinitely easier than every other transport sim I've tried.

    Would I prefer to continue to pay for high-quality prototypical routes? Sure.


    Would I prefer to continue to pay for high-quality fantasy routes? Sure.


    But I've got to be honest: getting assets from the DLC and third parties to assemble routes, even the free ones that excellent creators provide as a relatively complete CMP, can be a pretty annoying experience.

    I think that when some simulation DOES accomplish what I'm discussing here, the experience of quickly being able to put a train in an approximation of a rail network with a bunch of auto-gen scenery and foliage will be a fun diversion. Being able to build on that (and of course, jump into fantasy and prototypical routes outside the "World of Rails") would be even better.


    Quote Originally Posted by SR6900 View Post
    The current format provides the best option for developers in my opinion.

    I just can't agree. Even the less ambitious auto-world creation tool set that N3V is proposing as stretch goal # 8 will be a better format for those interested in prototypical route creation.

  13. #13
    Join Date
    Nov 2006
    Location
    United States of America, Massachusetts, Haverhill
    Posts
    10,590

    Default

    I suppose this could be used to generate fractal-based terrain which isn't such a bad thing for fictional routes for those that like that stuff. With something like this, we could set random number parameters which will set mountain height, cliff widths, valley depth, and other aspects of the terrain prior to landform generation. With parameters for tree and grass areas, and water height, we would have a ready-made terrain ready for track laying and asset placing. This is not much different than old SC4 where the parameters could be adjusted for the maps, or another program called VistaPro where fractal landscapes could be created.

    John
    Trainz User ID: 124863

  14. #14
    Join Date
    Nov 2006
    Location
    United States of America, Tennessee, Soddy Daisy
    Posts
    1,193

    Default

    For freelancers, what I think John mentions would be nice. Relevant to everybody, something else that could be useful is a tree placement algorithm where the user defines a tree placement area and selects what tree models to use to easily generate forests. Similar things could be done for painting ground textures and shrubbery and grass and etc. would certainly expedite route creation process.

    One thing that I will likely do with my route projects, Jay, is make everything myself, or at minimum with my team so that it's all allowed to be packaged in self contained cdps. I keep this policy for my modeling too, my few releases in the past were headaches in the regards of finding dependencies for most users, something I don't want repeated. But likely the same would be happening with what you proposed, as each person customizes their area of choice, one must hunt for assets and what not. Then you must deal also with distribution of literal chunks of the globe, as each person makes and customizes their own place. But then once you do that, you're back to what we already have, finite route segments distributed individually, only that they must be further processed and place on your globe.
    The Southern, Serves The South

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •