I’m 49 now, living in Munich and my real job is constructing 3D models for car makers such as BMW. This might sound an ideal starting point for creating models for Trainz, but the software is so different you can’t really compare them.
J50/1 tank LNER locomotive.
I’ve always been interested in railways and have dabbled with making physical models, but have hardly ever finished anything, mainly because I always had that nagging feeling that the model wasn’t quite as good as I could make it. Still, the result was that I accumulated a large collection of drawings and information on most aspects of railways in the UK, which would prove useful later on.
Another interest of mine around 10 years ago was computer programming, exclusively for the BBC and Archimedes computers produced by a company called Acorn in England. I was very active in those days, writing a complete programming library system (EvntShell it was called) in BASIC, contributing articles to magazines and creating free and shareware. The programming experience would also prove useful later, especially as the BASIC language for this particular range of computers was heavily dependant on passing messages back and forth between different programs – the same principle it turned out as Trainz script uses.
Starting out, Trainz CE
It was almost inevitable that I would create models for a train simulator one day, given that my job involves 3D models and my long standing interest in railways and computers. So it was that one day in August 2001 I discovered the Auran site and the Forum where a new model railway simulator was being discussed. The fact that it would be a model railway simulator especially interested me, as I was not (and am still not...) terribly interested in a ‘real’ railway simulator. I suppose it depends very much on one’s personal background, and in my case researching real railways was always with the idea of one day creating a model version.
Anyway, the first version of Trainz (CE) was ordered as soon as possible and arrived the day I had to travel back to England for the Christmas holidays. I had about an hour to play with it before I had to leave! Anyway, I was pretty impressed with the first version – much better than I had expected, and I knew that it had the potential to get even better. Coming from the ‘model railway’ direction, though, I suspect my expectations were different to that of many users.
Soon after the CE release, however, Auran announced that a gmax game pack would be available, which really was the icing on the cake for me.
First models
I’d played around with some other simple CAD programs before, building locos without any real hope of ever seeing them move, but gmax was a new challenge. I’ve never seen a 3D program yet that was easy to use; all are completely different to each other and to any other software and the user interface is usually, well, weird.
Given that, I wasn’t too surprised that the learning curve for gmax would prove to be steep. Instead of working my way through the tutorials (which would have been the sensible thing to do) I plunged in and started to build a loco. I hadn’t totally taken leave of my senses though, as I chose a diesel tram loco that didn’t need modelled wheels or animation. That took about four months to finish, most of the problems turned out to be with the texturing.
Getting animations working in Trainz was another ‘interesting’ problem that I eventually solved with much help from the Auran Community and video tutorials from Meletou. My first animated loco was a version of the tram loco without the side skirts and appeared in April 2002, on the same day that User: RO released his Flying Scotsman model. My little shunter got rather less attention than the Scotsman, but I was pretty pleased to have cracked the problem at last.
There followed a series of similar diesel shunters plus a few steam locos, which I still haven’t finished, but I was gaining experience all the time and each new model was getting easier.
TRS2004
My involvement with TRS2004 is documented with every copy sold, so I’ll keep this bit short. I got involved with TRS2004 as I suspected that Auran were about to move several goal posts with the introduction of ‘steam support’ and I was right. Initially, I offered to make some private owner wagons available (nice and easy, plus they were almost finished anyway), but as the project progressed it became clear that Auran would like a British steam loco as well, so the Flying Scotsman, which I was making at the time got included.
I didn’t have enough time to finish it properly, so Auran finished it in-house. The cab (from Rob Shaw) was and is superb, but the skin of the loco and tender (artist unknown, at least to me) was very disappointing, bearing little resemblance to the real thing.
After TRS2004 was released, I finished the Scotsman off the way I intended it to be and, unsurprisingly, I’m much happier with my model.
Rail-sim.co.uk and TRS2006
Jumping now to mid-2004, I was a bit jaded with Trainz by now, having produced a few diesel shunters and a couple of steam locos, which weren’t really an advance on anything else available at the time. Pikkabird (David Dallaston) had released a stunning class 37 diesel, which used specular lighting (dynamically changing highlights on the model, which avoids the cartoon-like appearance of block colours) and I couldn’t get it to work on my own models, despite generous help from David. I was a bit depressed as up to now everything I’d tried to do had worked – it might have taken me months of spare time work to do it, but at least I got there in the end.
I’d also vowed not to get involved with Auran’s beta testing program ever again after TRS2004 and then I got a mail from ChileanLlama (Paul Franklin). He was setting up a group to recreate the Settle and Carlisle route and was a big fan of my work (at least he said he was), which immediately weakened my defences.
I’d heard of the Auran Trainz Partnership Scheme (TPS) before this and I must say I wasn’t too enthusiastic. The idea of the TPS members getting advance information on TRS2006 and putting others at a disadvantage seemed very unfair to me. Still, for TRS2006 this was the only option available. Cautiously I agreed, mindful of the fact that working as a team we’d achieve more than we could as individuals and the project would give me a new impetus to get something finished.
It was agreed I’d do some typical rolling stock from the 1960s, nothing too spectacular, but with extensive scripting to provide features that hadn’t been seen before (at least on British models). During brainstorming sessions, details were worked out, Wulf_9 and Bloodnok (James Moody) providing help in the scripting department. Once I had some working scripts to look at I found the principles weren’t so very different to the old BBC BASIC programs I’d written years before and I gradually started modifying and extending the scripts myself.
Once I’d learned a bit about scripting it became obvious how powerful it could be. Using texture swapping, the same model could be made to look different, and by using sub-meshes I could assemble the wagons in the config file, re-using parts from other models. This was put to practical use for the van roofs (!), which have their own script to randomly select a texture when placing the vehicle, adding variety to a train made up of nominally the same models.
Incidentally most of the Rail-sim.co.uk rolling stock use scans of real models in kit form that I painted myself. Final weathering and lettering was added in Paint Shop Pro.
Pikkabird was also in the group in the early days, but it slowly became obvious that he was drifting away from Trainz, which meant we would have to come up with at least one new loco for the TRS2006 package and the only thing I’d started that was appropriate was the 4F, so that finally got finished after spending about two years lying around on my hard disk(s).
The problem with getting specular lighting to work had also been solved by this time with the help of Mike10 (Mike Banfield), although ironically, I’d discovered that a good skin is a lot more important than the specular lighting, and a dirty loco needs hardly any.
My involvement with Rail-sim.co.uk continues to this day, and further models will be released under this label, most importantly a number left to the group by Pikkabird. Of these the Jinty tank has already been released, although my input was limited to adding lamps, steam effects and numbering. I must say it was very instructive to see how Pikkabird had built the Jinty – it included more detail (working inside motion and a shovel on the tank top) yet used around 15% less polys than my equivalent models. Needless to say the lessons have been learned for the future.
Having spent the last few pages talking to myself as it were, I’ll try and answer a few questions that I think people would ask if they had the chance.
Do you actually use TRS?
Well, no. Only as a means of testing my content, as anybody who has seen me demonstrate my prowess in Surveyor and Driver at a TotR show can tell you.
How long does it take to complete a loco?
I spend at least an hour a day in gmax, more at weekends (too long my wife would say). It’s obviously heavily dependant on what the loco is and how much of a previous model I can reuse. My latest model is a J15 modified from the 4F, only the smokebox, chimney, dome, cab and mainframes are completely new, the rest is rescaled and recycled from the older model. About 20 hours went into the loco, 10 hours for the tender and mapping both models and texturing and bump mapping another 10 hours.
Getting the model ready for release takes me a long time though. I always start out really fast and slow down as I get to the end of a project, usually because I’m already thinking about the next one. Also, the final slog really is hard work – is the shadow file ok? Is everything working as expected? I find it’s best to create a checklist in Excel to keep track of a model’s status, just to make sure nothing gets forgotten.
I also try to incorporate something new into each model, and this can take a long time to get right. I lost count of how long I spent on getting the headcode lamps working properly on the 4F because it involved passing messages between the loco and each lamp, and this was something I hadn’t done before. Once it was working though I found I could add the lamps to other models very quickly.
Why make only UK rolling stock?
Simply because that’s what I know best. It’s hard to get interested in making a model when you don’t have any feel for when and where the prototype was used. Although I’ve lived in Germany for seventeen years now I still don’t know much about the railways here. I’ve loads of drawing though!
What about making models for the new simulators coming out in 2006?
I am not involved in the development of the Kuju sim or TMTS, so I will have to wait until they are released to see what they are like and what would be involved to port my models over. I certainly wouldn’t rule out supporting one or both of the new sims, but from what I’ve read about them they are going more in the direction of real railway simulators, which doesn’t hold as much interest for me.
I’d also doubt that either sim will have an equivalent to Trainz script that will allow the flexibility of Trainz, which is I think a major advantage. Even if they did, it would be different and I don’t know if I’m up to learning yet another programming language.
I’m keeping an open mind about alternatives, though it doesn’t look as though Trainz will be developed any further, and sooner or later it will get left behind by a new generation sims with much better graphics. As I see creating models for simulators as a hobby, I can’t see me giving it up any time soon
