RW Feature Update

[theokus]

I agree with Vern, some sort of autogen scenary facility needs to be developed. I mean, even if we just had countryside and cityscape, it would be staggering the amount of routes that appear because to be honest, I pretty much enjoyed laying all the track, signals etc. for my Midland Mainline route but scenary is just a sticking point which will probably just mean the route stays on my PC as I wouldn't really be that happy releasing it with no scenary apart from platforms and bridges. It's the miles of countryside I just can't really lay scenary for because it's seemingly so random, perfect for an autogen sort of facility . So if autogen could be used for the majority of a route and then hand made scenary for important features then that would be a pretty acceptable route!

Richard,

I am cautious about this system of autogen.
It can never properly reflect the actual situation.
If have the same feelings about this as most people have, talking about the accurate operation of the signals or the AI.
Placing scenery is a lot of work....

Now my question (again): do you use decals or do you paint all the enviroment?
(again= I asked this in other topics)

[Neptune50006]

because it's seemingly so random

It may look random, but believe me, if you just plonk trees and bushes any old where it just looks wrong. Producing convincing country side can take a very long time IMO.

[phat2003uk]

Yeah sure, I definitely agree with you both. Unfortunately though, at the moment it's either autogen scenary or not really any at all for me anyway .

[theokus]

Yeah sure, I definitely agree with you both. Unfortunately though, at the moment it's either autogen scenary or not really any at all for me anyway .

I understand.
You can use clusters he phat.
Look at the foliage pack.

In my route I have a small village called Rotem (Belgium)
Y'll find there the old plant of a sinc factory with a lot of woods around.
Those woods (trees) I did by hand...
And when I go back about 500 à 600 meter and look at it...it is as real as it can gets

[karma99]

I think the best option for scenery would be for copied groups to still align themselves with the ground individually when placed, and a "copy and paste pallet" which is part of your install of the game/editor.

So you create an area, with hedges and fields for example. You select and copy it and permanantly store it in your pallet, and can now place that into any route you edit. When it's placed, and for any copied groups of items, each item aligns with the ground based on where it's placed so you don't get any floating and the painted ground is applied as well.
It will still need a little tweaking but for those repetitive ploughed fields surrounded by hedgerows you could whack out miles of it pretty quicky with this option.

[theokus]

I think the best option for scenery would be for copied groups to still align themselves with the ground individually when placed, and a "copy and paste pallet" which is part of your install of the game/editor.

So you create an area, with hedges and fields for example. You select and copy it and permanantly store it in your pallet, and can now place that into any route you edit. When it's placed, and for any copied groups of items, each item aligns with the ground based on where it's placed so you don't get any floating and the painted ground is applied as well.
It will still need a little tweaking but for those repetitive ploughed fields surrounded by hedgerows you could whack out miles of it pretty quicky with this option.

Karma... how many spots in the soil, ground etc. are realy flat?
I have here a few 100th. prints of Google , and nothing is the same
It's very clear now that we (everybody ?) is looking for to speed up route creation.

In the same line: I wonder how Microsoft solved that W.O.R.
And the correct gradients...

[karma99]

I think the best option for scenery would be for copied groups to still align themselves with the ground individually when placed, and a "copy and paste pallet" which is part of your install of the game/editor.

So you create an area, with hedges and fields for example. You select and copy it and permanantly store it in your pallet, and can now place that into any route you edit. When it's placed, and for any copied groups of items, each item aligns with the ground based on where it's placed so you don't get any floating and the painted ground is applied as well.
It will still need a little tweaking but for those repetitive ploughed fields surrounded by hedgerows you could whack out miles of it pretty quicky with this option.

Karma... how many spots in the soil, ground etc. are realy flat?
I have here a few 100th. prints of Google , and nothing is the same
It's very clear now that we (everybody ?) is looking for to speed up route creation.

In the same line: I wonder how Microsoft solved that W.O.R.
And the correct gradients...

I think you misunderstood me Theokus, the ground would be DEM'd as normal, but the scenery dropped onto it in groups would align with it correctly and the painted area would be the same. The last thing we want is anything being flat

I'm just suggesting that speeding up doing it manually is probably a better option than anything automated.

[theokus]

I think the best option for scenery would be for copied groups to still align themselves with the ground individually when placed, and a "copy and paste pallet" which is part of your install of the game/editor.

So you create an area, with hedges and fields for example. You select and copy it and permanantly store it in your pallet, and can now place that into any route you edit. When it's placed, and for any copied groups of items, each item aligns with the ground based on where it's placed so you don't get any floating and the painted ground is applied as well.
It will still need a little tweaking but for those repetitive ploughed fields surrounded by hedgerows you could whack out miles of it pretty quicky with this option.

Karma... how many spots in the soil, ground etc. are realy flat?
I have here a few 100th. prints of Google , and nothing is the same
It's very clear now that we (everybody ?) is looking for to speed up route creation.

In the same line: I wonder how Microsoft solved that W.O.R.
And the correct gradients...

I think you misunderstood me Theokus, the ground would be DEM'd as normal, but the scenery dropped onto it in groups would align with it correctly and the painted area would be the same. The last thing we want is anything being flat

I'm just suggesting that speeding up doing it manually is probably a better option than anything automated.

Aha!! Yes indeed, indeed.
That's also why I am asking about those decals.
If I am not wrong but in Trainz those decals are following the ground.... not in KRS he?
And with those decals one sees clearly the different colours, scenery and trakcs too he.
It could speed up things.
I saw in the IOW one used some decals

[NeutronIC]

I went to a developer day for Microsoft ESP some time ago now (ESP 1.0 was the commercial version of Flight Sim, ESP 2.0 was going to be the commercial version of Train Sim 2) and one of the presentations they gave that really stuck in my mind was their very detailed overview of how Autogen scenery works in FS X. It really is impressive how many different data sources they combine, on the fly, to generate both the textures and the type information, and then how that's mapped on to various other bits of data to produce a final texture. The whole talk lasted about an hour or so, and covered the process for just one small tile - and the engine is doing this in real time on your machine, amazing stuff.

It's difficult to explain but I do think that there's merit in the idea of a *clever* auto gen system like this - you can go around the route "describing" it, it can then drop in default auto gen and then you can go and customise it with localised enhancements, key features and buildings etc, but not have to worry quite so much about "random village" and so forth. Obviously nothing can beat doing it all the hard way, the results are going to be far superior, but with the auto gen approach you could have some thing driveable and with scenery in a small fraction of the time, probably as much time again as it took to do the track to get all the scenery described.

That said, the vast majority of Railway Enthusiasts i've spoken to, that want train simulation stuff, are far more interested in things that represent reality so it's a thin line to draw. Cross it, and while making things easier you turn people off, don't cross it and it turns developers off from producing...

That said, John Griffiths is really working his magic with Rail Sim, so it's obviously not impossible to crank out high detail routes

Matt.

[theokus]

Matt.

Goods news indeed.
And thank you for the info.

[Blackpatch]

Some form of auto gen would help an awful lot. Better to amend bits you don't like rather than doing it all from scratch.

BP

[bigvern]

Some form of auto gen would help an awful lot. Better to amend bits you don't like rather than doing it all from scratch.

BP

That is why it is essential to get a constructive dialogue going with RS.com as to what can and can't be achieved in future versions of the sim. Without new routes it will wither and die regardless of how many times the product gets a makeover. The reason MSTS and TRS have kept going strong is because of the routes (and of course the multiplicity of traction and rolling stock). Without blowing my own trumpet, when experienced route builders like myself hang up the spurs and remove RS to go back to TRS (I can't quite bring myself back to MSTS though) then maybe they need to be asking why. I've mentioned the eyestrain which seems to occur much more rapidly working in RS, maybe because you seem to be working more close up to the detail or maybe it is something with the graphics engine. I will not risk my eyesight on the time it takes to get results in KRS just to produce a few freeware routes! Ergo, route production needs to be quicker and streamlined and I think they maybe need to look at how the graphics are more tiring on the eyes than working in any other sim.

Let's have something positive back from RS Towers - we're not expecting miracles but a response other than RS does it differently/learn new techniques/that's your opinion would be a start. Sit down with the people who will ultimately produce the content to ensure your sim's survival and work with them within the constraints of the code and available money to refine and change what will get people producing content on the scale of the other sims. Otherwise I fear it will just become another Eisenbahn.exe (a handful of users trying to sell each other micro payware) as people migrate back to the older sims or leave the scene altogether.

[AndiS]

It's difficult to explain but I do think that there's merit in the idea of a *clever* auto gen system like this - you can go around the route "describing" it, it can then drop in default auto gen and then you can go and customise it with localised enhancements, key features and buildings etc, but not have to worry quite so much about "random village" and so forth.

Sounds exactly like what I am dreaming of for a year now. For our case, the most plausible workflow looks like this.

Stream 1 fed by Google & alternatives.

1. You plaster the route with images (semi-automatically, if the coordinates, size and orientation of these images can be stored together with the image).
2. Some moderately clever image processing software pre-marks the structures, like tree lines in a field.
3. You draw lines (splines, most likely) on that, reusing or discarding what the image processing software prepared for you, adding your "intelligent" recognition capability, i.e., marking structures covered, distorted or blurred in the aerial photo.
4. An external batch programme takes these annotations and creates scenery objects which could be large and custom-made with a level of detail relating to the distance from track to improve framerates. It would help itself from a library of trees, buildings, etc. in the process, just recombining building blocks in many different variants (orientation, texture variant, size).
5. You then correct what you don't like and add the near bits. These near bits (e.g., embankment with support walls, bridges) could also be created by maybe a separate process with a different degree of precision, following the above line.

Stream 2 fed by Open Street Map.

1. You annotate their map with building and vegetation types as seen on the route or based on guestimates.
2. Rest like the last 2 steps above.

The strength of OSM lies in the vast amount of vector data already available. But I would mostly recommend it for generating buildings and streets, as streets are clearly the focus of this project. However, for some places, the annotation of type of area and forests are nice, too.

Clearly, the combination of these streams (stream 3 being the DEM data) is the very hard bit. I don't mean to compete with Microsoft here. But I think that the one missing bit is the generator for the objects. The markup in step 3 of stream 1 could be done by placing abstract lofts in World Editor. The KML format already stores useful information about the saved images/markers but I have not found a way to avoid the need of placing a marker per image.

[scefhwil]

...... the people who will ultimately produce the content to ensure your sim's survival .......

You know what Vern, I think that RSDL/Railsimulator.com have a very different view on who those content producers are, ie it's them not us.

RSAdam recently made it clear on the RS Forums that they didn't get any royalties from the sale of RS1. That income was split between Kuju/EA/F4G and RSDL only made reseller income from the direct sales of the sim off their website. Other than that their income has been from sales of the content they have produced. (Does leave you wondering then who stumped up the cash for MK1 and Mk2? Possibly it was an agreement that Kuju would reinvest a small percentage of their royalties in patching/updates? Might explain why RSDL always referred to have to make hard choices in this respect?) So do they really then have any incentive to make things easier for third party providers if it might threaten their own ability to generate sales? I'm convinced they primarily want to sell content and improving the core sim is secondary (isn't that what 15 months of RSDL has demonstarted? May be RS.com will be different?). Sure, they are business people so will, slowly, improve things so their product doesn't prematurly age, but the demise of MSTS2 can only have reduced the business case for spending on improving the core sim. They also seem to me to be wanting to push train simming away from being in a niche market (used by enthusiast with freeware distribution at it's core) to instead be in the realm of traditional mass market gaming. Hence their move over to Steam. Is it really about providing a constant stream of core updates easily? Or is it about finding a new market where the expectation already is for the occational update and with the ready availabilty of "Expansion Packs" being the norm?

But I am looking forward to seeing the new features that RS.com are saying will be in RWs, they are certainly along the lines of what I was hoping for. But, as I have already been told, I am sure I will ultimately be more dissappointed then enthused by it's release.

Stuart

[g0fthick]

You know what Vern, I think that RSDL/Railsimulator.com have a very different view on who those content producers are, ie it's them not us.

From day one RSDL have been lending a hand to content creators in learning the new tools required. They have produced a vast deal of documentation on the subject, in both written and video formats. There's more of that on the horizon with RS Works too.

The phrase don't look a gift horse in the mouth comes to mind.