Graphics are great, the details are superb. The physics though really lack unfortunately. Both of the EMUs should have camshaft controllers (well, at least their counterparts in the real world), but they don't behave like that at all. It's especially noticeable in the class 5000 EMU (the 2 car EMU with the manually lapped brakes), because you have an ampere meter there. The ampere just stay at the same position the whole time while accelerating, you can even go a few notches down and back up and the meter reacts to that. This isn't exactly camshaft controller behaviour. The diesel train also starts accelerating way too fast. Still, the game is fun and looks really great, but at the moment not really a rival to TSW.
Just tell the boss it's broken, maybe these shutdowns are meant to be. It's a sign from the train gods that it's time for a new PC.
I don't know what you mean. While I don't know how in-depth the simulation actually is, the overall performance seems pretty realistic to me when compared to real life videos. Here is a cab ride video of a 115 series for example. I'd say the driving characteristics - acceleration and delay between handle movement and actual acceleration/braking - looks pretty similar. Here is a video of a KiHa85 as well, pulling out of a station. I'm not sure it's much different in game? Also, the 5000 series does not have lapped brakes in the game. Only the 2-car 115 series (called 1100 series in game) does.
Sorry, that was my fault with the 5000/1100 series. mixed the brakes up, but I meant the train with the ampere meter, which is the 5000 series. And yeah, you are right, the performance kind of meets the prototype. Still, the technical and physical side behind it seems not to be simulated. That's what I meant to say. Well, in TSW on the other side, there are trains where technically and physically everything seems to be simulated but the driving performance does not feel right at all, but that's another problem. Still a great game, but not a TSW killer. It shows, what's possible in terms of graphics, if you try to, though.
The trains I tried up until now feel similar to similar trains in JR East Simulator (which is used by the real JR East for driver training). So I doubt the physics are that far off.
I'll have to correct myself. Previously I said it does in fact has real suspension physics, but since then I saw the dev confirming it that it is not actual suspension physics. From his words:
Well, the oldest trains in JR East Train Simulator are the 205 series (prototypes built from 1984) and the 211 series (built from 1985). Both the 115 series and the Keio 5000 series, the both prototypes for the RTS trains, were built from 1963 onwards. There are around twenty years of technological progress between them. As I said, the performance itself doesn't feel off, but the way it achieves it's performance. The trains don't feel like a rheostatic controlled train from the 1960s.
Right, exactly what i found out but haven't communicated here yet since no one wants to hear that. Nice to see the dev itself told the truth about that. The game indeed uses the UE5s chaos physics system, what is basically what Nvdia PhysX was before but a tad better. All the extra movement besides the repeating pattern from the uneven track is just impulse triggers for the used physics asset. So, means that on the junctions there is a trigger placed what causes the rig to dip down a tad by applied force. A neat, old, trick. And the stopping wobble is also just a preprogrammed event thats being called when the trains stops, no matter how you stop. Also seems the whole physics system is disabled on all AI and when using the 4 camera, what makes some sense for saving performance.
I don't care if its physics or some default engine tricks - it looks kinda nice, suspension is moving, cab is swinging. Much better than nothing.
Do you have a source for that? Exactly. Even "real" game physics is just a bunch of mathematical rules that approximate real physics in a "good enough" way.
Is a video of the controller system of a 115 series enough of a source for you? Edit: Please don't misunderstand me. I quite like the game, already have a few hours play time in it (more than TSW in the last few weeks altogether) and it certainly is a well done game. It just puts the weight on totally different things than TSW, and the implementation of the correct driving mechanics (which in fact a pity for me, who loves to see how different trains behave) doesn't seem to be one of the things. So I would not count it as a TSW killer or even some kind of rival to TSW at all. That's all I wanted to say in my first post. And imagine the backlash if the German BR 111 in TSW had the same driving mechanics as a BR 146, just with the acceleration characteristics close to the BR 111.
If I remember correctly, one of the early documentation mentioned things about suspension, dampers and whatnot. So I wonder if that got scrapped at some point. I remember he had issues of the train yeeting off the rails in very rare and random occasions. Maybe that was the reason for a much simpler physics model. I don't know. Either way, I think it's pretty nicely done. But I agree that sideways movement could be a bit more.
Right, that is one major problem when trying to use pure physics systems to create track followers. It's not working well because the solver precision can't be set to a value that is needed for such accurate positioning. You always tend to leave the tracks. Setting the value as high as needed ends in a 1fps slideshow. And even then the precision is not enough to hold the trains on the rails. For that reason most train sims/games use a dual system of pure Math positioning and a physics engine. The track follower system calculates the position of each axle of a formation on the tracks system, with the same curve Math the tracks being laid with (this can be really complicated like it is in MSTS, oBVE, TSC and TSW and others). After getting the idea where each axle, bogie and train body has to be at that specific point in sim time, it pushes these transform position data to the physics bodies the trains is consisting of (whatever physics is being used, it always is rigid body physics), then the physics solver iterates over all these positions with its logic and calculates the results of the asked positions and tries to reach them by applying the data to the bodies. After that there is a sort of sync process that again reposition the followers back to the nearest position on track as given from the physics calculation results. This, if not precise enough, can lead into wrong positions and derailing (where that term is misleading since it is a threshold where the train being stated as derailed and the game does its specific stuff when that happens, activating simple body physics for example like in TSC and TSW). MSTS didn't had physics, it was a pure spline track follower system, as most RTS games use for simplicity when doing trains. Running trains seems to do dual system with UE5 Chaos Physics as the solver. I'm sure i found the track system being used on FAB. It was extended by the dev for the game (no, i will not tell the name of it). Really simplified explanation how that works. This stuff is massively complicated, especially when doing real life track laying with correct curves and clothoids and such, like TSC and TSW does, but not the spline laying like in Trainz.
