[jtr1962]

JTang, on 17 April 2013 - 04:57 PM, said:

Figured out how to rotate the landscape, I think we can split the rotation: 1/2 to tilt the cab, half to tilt the landscape. The tilting is small anyway, thus rotating the landscape may not be noticeable, unless tilt train is used.

Depending upon the train speed in the curve, the cab might end up tilted a lot or a little relative to the driver's head. Splitting the difference between the landscape and cab is a good compromise. I look forward to seeing this further refinement of superelevation. A interesting idea to carry things even further might be to vary the tilt of both the cab and landscape with centrifugal force (i.e. the higher the centrifugal force, the less the cab tilt and the more the landscape tilt). If you're at a dead stop on a superelevated curve, then only the cab tilts. The only issue is I don't know if Open Rails currently calculates centrifugal force anywhere (that probably needs to be added eventually to account for friction on curves).

Related to superelevation-there are already track profiles for the default track and ScaleRail. I've made a profile for the default track with normal sized rails. I believe there are profiles for UK Finescale. Anyone know if there is a profile for DB tracks (both wood and concrete)?

[ATW]

I haven't been in a very high speed tilting train before, but should there also be something related to grades and cab forward/backward tilt leaning or should that belong in vibrating?

[Csantucci]

JTang, on 17 April 2013 - 04:57 PM, said:

Figured out how to rotate the landscape, I think we can split the rotation: 1/2 to tilt the cab, half to tilt the landscape. The tilting is small anyway, thus rotating the landscape may not be noticeable, unless tilt train is used.

I very much like this principle, that I think would put ORTS above all other train simulators. Widening the concept that Chris has explained, great would be if the proportion between cab tilting and landscape tilting could depend from ratio between speed and line speed, e.g. varying from 65% landscape and 35% cab at line speed to 0% landscape and 100% cab at 0 speed. For tilting trains the 65% could reach an 80%. Let's hope computing time doesn't jump to the stars!

PS: Oops, I see now that jtr1962 expressed a similar concept, but basing on centrifugal force instead of on the ratio between speed and line speed. My proposal is a bit rougher, but it uses variables that are already available in ORTS.

[gpz]

I think

if ( velocity[m/s] ^ 2 / curve_radius[m] == gravity[m/s^2] * tangent(superelevation_angle) )

then the drivers head is not tilted at all relative to the cab. Of course the cab is always tilted as the rail's superelevation.

In other words, dividing the two sides of the equation above by each other will give the tilting ratio of the drivers head relative to the cab.

[Csantucci]

By the way for such small angles the tangent can be approximated by the sine, that is easily calculated as the ratio between track superelevation in mm and track gauge.
For tilting trains it is a bit more complicated because superelevation_angle=track_superelevation_angle+train_tilting_angle.
Let's see what solution JiJun will choose.

[cjakeman]

gpz, on 18 April 2013 - 12:11 AM, said:

then the drivers head is not tilted at all relative to the cab. Of course the cab is always tilted as the rail's superelevation.

and if a conventional train negotiates an unelevated curve at speed, then we should see the line and the cab both tilt away as the driver leans into the bend.

Should be interesting,

[gpz]

Just let's try not to get sick while driving, will be like an airplane. :)

[Csantucci]

I have read someone saying that the center of the cab rotation is at the left top corner instead of being at the center of the monitor. Is this true?

BTW I have the sound impression that in tunnels superelevation has a problem. I found that train oscillate as if the superelevation would vary at every track section change, as it oscillated in the first superelevation versions also in tracks outside tunnels.

[Csantucci]

Hi JiJun,
your first release of the superelevation effect with also tilted landscape leads me to enthusiasm again! Here two pictures with a tilting train:



The effect when entering and exiting a curved tunnel is also particular:



Thank you! Are you already considering speed to define how to divide tilting by landscape and cab? I see that at 0 speed the landscape remains tilted.

Moreover when you go to opposite cab with CTRL-E I have the impression that the cab tilts in the wrong direction.