Cab tilting in rel. 1569
#1
Posted 16 April 2013 - 12:00 PM
Jtang introduced cab tilting on superelevation to release 1569, see picture (superelevation to maximum, and moreover tilting train):
Thank you Jtang! However in my opinion in the picture it would be better if it was not the cab that tilts, but instead the landscape, because the viewer also tilts because of centrifugal force.
What do you all think about this?
Thank you Jtang! However in my opinion in the picture it would be better if it was not the cab that tilts, but instead the landscape, because the viewer also tilts because of centrifugal force.
What do you all think about this?
#2
Posted 16 April 2013 - 12:05 PM
G-R-E-A-T!
I had asked many times when this would be introduced - finally it was, can´t wait for saturday to come!!!! :)
But I agree, that it would be better to turn the viewport with the cab texture - first because in reality the engineer will "tilt" with the cab, so his view of the inside of the loco doesn´t change, and secondly with controls, dials and other cab instruments moving out of view, it will become sowhat difficult to drive a train on routes with lots of curves
But again, great, and can´t wait for the next experimental to be released! :D
I had asked many times when this would be introduced - finally it was, can´t wait for saturday to come!!!! :)
But I agree, that it would be better to turn the viewport with the cab texture - first because in reality the engineer will "tilt" with the cab, so his view of the inside of the loco doesn´t change, and secondly with controls, dials and other cab instruments moving out of view, it will become sowhat difficult to drive a train on routes with lots of curves
But again, great, and can´t wait for the next experimental to be released! :D
#3
Posted 16 April 2013 - 03:26 PM
Tried it out and am very impressed, cab vibration is also a bonus but still needs tweaking in that.
But one thing I found out is the controls and digital guages bounce off the screen a bit and the view resets to a different position every view change getting back into the cab.
In other words, great job and addition team :D
But one thing I found out is the controls and digital guages bounce off the screen a bit and the view resets to a different position every view change getting back into the cab.
In other words, great job and addition team :D
#4
Posted 17 April 2013 - 04:42 AM
Yeah, I also noticed the different-reset-position thingy, but related it to pressing the 1 key multiple times without "going outside" in between - which is not the intended "use" of this function, so I thought, why should this matter?
With going out in between I did not experience it, as I never got to run OR all too long during the last days...
With going out in between I did not experience it, as I never got to run OR all too long during the last days...
#5
Posted 17 April 2013 - 06:26 AM
Oh no, I was afraid it might look like this... I can't see why the engineer would tilt his head in the opposite direction to keep his head horizontal. I really hope this is changed in the future so the landscape tilts instead! Great work though! :pleasantry:
#6
Posted 17 April 2013 - 09:16 AM
Think about this, if the cab controls and instruments jiggle around a bit while "leaning into the curve" and "straightening back up again", how about a whole outside environment, distant mountains and all?
That is hundreds, possibly thousands of assets, to keep track of regarding their positions, sightlines, culling, all while the train rounds the curve, you are able to view around, etc. etc. in real time without hickups, flutter or stutter.
I think OR has come very far already with superelevation on existing routes. A fully tilting world in realtime will be a huge achievement.
No, I think more track profiles should come first, so those routes which do not use ACleanTrack and ACleanRails ace files look nice. Then there is the task of having only mainline track and tracks with moving AI trains having active superelevation, while yards, sidings and spurs remain level. I'd like these challenges to be solved first.
That is hundreds, possibly thousands of assets, to keep track of regarding their positions, sightlines, culling, all while the train rounds the curve, you are able to view around, etc. etc. in real time without hickups, flutter or stutter.
I think OR has come very far already with superelevation on existing routes. A fully tilting world in realtime will be a huge achievement.
No, I think more track profiles should come first, so those routes which do not use ACleanTrack and ACleanRails ace files look nice. Then there is the task of having only mainline track and tracks with moving AI trains having active superelevation, while yards, sidings and spurs remain level. I'd like these challenges to be solved first.
#7
Posted 17 April 2013 - 09:30 AM
_o_OOOO_oo-Kanawha, on 17 April 2013 - 09:16 AM, said:
Think about this, if the cab controls and instruments jiggle around a bit while "leaning into the curve" and "straightening back up again", how about a whole outside environment, distant mountains and all?
Actually I think the solution to this is quite simple, as of what I understadn: The cab in MSTS is a picture laid over what can be seen thru the in-cab viewport (or camera). In the MSTS cabview definition files (.cvf) for each and every individual cabview there is a different camera positioning defined, in the file. This also contains the turning / tilting angle of the camera around the longitudinal axis of the train, i.e. the axis that goes alongside the moving direction. This parameters NEED to be processed by OR as to correctly show the cabs like they do in MSTS (haven´t taken a look in the code, was simply logical deduction). IMO when going into curves, it would just be neede to turn the longitudinal angle of the camera, and depending on how OR renders the cab-overlay-image, this should automatically "turn" with it, if it is done like in MSTS: it´s just laid over the screen output. So, with all that said, the conclusion is: no positions of outside objectss would have to be change, the visible changes would in every rendering cycle be done automatically by the existing code, as just the viewport is chahnged a bit, like it´s already implemented with the outside cameras.
Those, indeed are a good example for what I mean: If you move an outside camery (pan, turn, and whatsoever), you just change the position of the viewport, the rest is done automatically. With the above solution, the same would apply for the in-cab one: only the viewport is changed, the environment stays untouched and is just viewed from another angle. The only difference will be: the changes to the incab viewport needs to be done by the proram, whereas the changes to the outside one can be done by the user (mouse, arrow-keys and whatever).
Relating to the actual implementation seen above in the screenie: not the overlay pic, but the viewport needs to be repositioned (angle...)
Maybe this helps, regards
Markus
PS: For me this solution seems simple - maybe so simple that nobody had thought about. What i wanna say with this is: that´s why i wondered why cab tilting on SE´d curves hadn´t been implemented yet. It´s just letting the program change one variable (the longitudinal angle) in relation to the superelevation parameter!
#8
Posted 17 April 2013 - 11:09 AM
Csantucci, on 16 April 2013 - 12:00 PM, said:
Jtang introduced cab tilting on superelevation to release 1569, see picture (superelevation to maximum, and moreover tilting train):
Cabtilting.jpg
Thank you Jtang! However in my opinion in the picture it would be better if it was not the cab that tilts, but instead the landscape, because the viewer also tilts because of centrifugal force.
Cabtilting.jpg
Thank you Jtang! However in my opinion in the picture it would be better if it was not the cab that tilts, but instead the landscape, because the viewer also tilts because of centrifugal force.
That dramatic image above is the view that the driver will have when his train is stationary or nearly so. According to Wikipedia, tilting trains at full speed only provide about 80% of the centrifugal force, so a driver would still need to lean into the curve a little.
So the angle should surely be speed-dependant and, at full speed, might look more like this:
#9
Posted 17 April 2013 - 04:57 PM
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.