I have a question combined with a suggestion.
How is rider "rigidity" modeled? By rigidity, I mean mostly his arms and legs.
Watching replays and riding quite a lot, I notice that the rider is very, very rigid. He stays put most of the time, and when he moves, he tries to get back in position at lightning speed. I'm asuming the rider has a weight, and that the bike feels this weight and its position.
When riding SX, the whoops are a good example: the rider shakes around a lot while IRL, the torso of the rider barely moves. On steep / short jump faces, the rider will "tilt" with the bike, upsetting the balance. etc. A side effect is that the bike suspensions have to work twice as hard: while IRL the rider will "soak up" a good deal of a shock, virtually making himself lighter for the bike, it's not the case here. So bikes are bottoming out a lot (leading to other issues like bouncing, even on VERY hard settings).
In order to test this properly, would it be possible to have access in the next beta to some "rider settings"? Something along the lines of the fork settigns would be ideal - "spring" (or "strength"), bump and rebound. Maybe even for arms and legs separately. If not, could you try making his arms and legs more loose? Poor guy is cramping up 24/7.
A full on +1 from me..
I've mentioned it before but not had an answer / response, and definitely love the idea of having some rider controls like you've requested.
I can be wrong but I was under the impression that the rider is rigid (i.e. articulations are kinematics only, no dynamics, just for animations).
That would definitely explain why the modern and classic riding styles in gpb don't affect the physics..
I guess if that is the case though, there must be some physics at play from the rider or an external force even if it's not connected to the rider model. Perhaps this can be tweaked, and maybe the animations could be looser to reflect.
Quote from: TFC on January 04, 2017, 09:49:36 AM
That would definitely explain why the modern and classic riding styles in gpb don't affect the physics..
Hmm I think that's a different problem. In GPB the riding style has only a visual impact, no impact on physics. I.e. when the rider leans, physically it leans in the same way even if visually it may lean differently in classic style than in modern style. Rider style is (at the moment at least) only eye-candy.
Quote from: TFC on January 04, 2017, 09:49:36 AM
I guess if that is the case though, there must be some physics at play from the rider or an external force even if it's not connected to the rider model. Perhaps this can be tweaked, and maybe the animations could be looser to reflect.
My understanding is that rider lean has an effect on physics, but it's "simplified", e.g. rider leaning forward --> rider mass position more forward --> different overall inertia matrix and CoG.
What Asdrael is asking is more complex, something like: the rider mass localized somewhere in the torso is linked to the bike with two damped springs for legs (and maybe 2 for arms, legs allowing vertical movements, arms allowing longitudinal movements). It's not terribly difficult, but not trivial either: additional state variables (more complex dynamics to integrate), plus some tuned controller of the rider position (trying to stabilize the rider position in the target position dictated by the rider lean l/r and f/b inputs) ...
The rider movement (visual, not sure how its mass is handled) is already smoothened from your controller input tho. Having dampners for arms and legs is just a fancy way to smoothen visual and hopefully mass movement. I know that the rider mass position is taken into account by the bike already, since moving your rider around moves your bike around too (try leaning forward or backwards on a jump face).
I think there's no vertical mass movement today (except for the sit/stand): the rider's mass is "rigidly" linked to the bike (and not liked to the bike via springs "simulating" the legs/arms).
When you land flat the riders' legs do not bend to "absorbe" part of the hit. This is what you're asking, right ?
Quote from: HornetMaX on January 04, 2017, 01:54:52 PM
I think there's no vertical mass movement today (except for the sit/stand): the rider's mass is "rigidly" linked to the bike (and not liked to the bike via springs "simulating" the legs/arms).
When you land flat the riders' legs do not bend to "absorbe" part of the hit. This is what you're asking, right ?
Yep. Not sure if they bend very slightly or not at all though. Pretty sure we'd need that modelled to move forward thought, especially in SX.
Edit: I guess it could work with having the center of gravity of the rider in the middle of 3 springs (one for each direction X, Y and Z relative to the bike). The rider lean and sit/stand would just act as a force against a spring in the specific direction. Could be the easiest model in that case, at least easier than modelling each arm and leg as independent entities.
Sorry, I was talking generally but was referring to the same setup as Asdrael.
Watching a real life rider take on whoops is a perfect example. The riders body stays roughly the same height off the floor, and alternates between absorbing the impact of each whoop and letting the bike drop away slightly until it hits the next whoop.
If this could be simulated using what Asdrael suggests, riding while standing would put a lot less pressure on the bike when hitting small lumps and bumps on a track as the riders springy limbs could absorb these impacts instead of being part of them.
Quote from: HornetMaX on January 04, 2017, 01:54:52 PM
I think there's no vertical mass movement today (except for the sit/stand): the rider's mass is "rigidly" linked to the bike (and not liked to the bike via springs "simulating" the legs/arms).
When you land flat the riders' legs do not bend to "absorbe" part of the hit. This is what you're asking, right ?
(https://koenig-media.raywenderlich.com/uploads/2012/06/Wait%E2%80%A6What.jpg)
The rider is a separate body with its own mass and inertia.
There are multiple springs to link it to the chassis, to simulate the arms and legs.
It's easy to see the "legs" in action with the rider standing: when landing from a jump the rider will almost sit to soak the bump.
Is it possible for us to adjust these springs?
I had a look in the rider pkz but couldn't see anything obvious.
Futhermore, is there a possibility you or Snappe might revisit the riders contribution at some point?
Doh, all is already in place then. Just a matter of tuning ?
Quote from: PiBoSo on January 04, 2017, 02:18:23 PM
The rider is a separate body with its own mass and inertia.
There are multiple springs to link it to the chassis, to simulate the arms and legs.
It's easy to see the "legs" in action with the rider standing: when landing from a jump the rider will almost sit to soak the bump.
Great. Then, I think this could be (significantly) improved rather easily by changing the settings - or even allowing us to change them. The arms are currently way too rigid, and I think overall both arms and legs are too linear in their behaviour - they should be softer the first half of the movement and same as now / if not harder for legs the later half. Also, they come back to "neutral" way too fast. But I'm glad to have confirmation the model is already in in an advanced form, just tuning it is necessary.
Quote from: HornetMaX on January 04, 2017, 02:22:34 PM
Doh, all is already in place then. Just a matter of tuning ?
Hopefully only tuning is needed.
However, maybe the springs or dampers need to be more advanced, for example with separate damping for slow and fast bumps... :-\
Quote from: PiBoSo on January 04, 2017, 02:34:12 PM
Quote from: HornetMaX on January 04, 2017, 02:22:34 PM
Doh, all is already in place then. Just a matter of tuning ?
Hopefully only tuning is needed.
However, maybe the springs or dampers need to be more advanced, for example with separate damping for slow and fast bumps... :-\
Could help. Is the basic problem like: I want the rider legs to be stiff when climbing a jump ramp and to be soft when on whoops ?
Well thanks. I'll have a look into it and see if I can make sense out of the settings, and then find values I like. I had no idea it was accessible in rider.cfg, I thought it had to do with how the rider camera was bound.
Tested it some more (yes this is bugging me quite a bit :p).
It appears that with Automatic rider F/B lean OFF, the "arms" become rigid and lose any dampening properties. Not sure if intentional. With it ON, the arms do soak up (not ideally but hey at least they move). It has no influence on leg movement though, it function both ON and OFF.
It also appears (but this is harder to judge), that with Automatic Rider F/B lean ON, if you give even a minor lean F/B input, the arms rigidify (hands - shoulder distance doesn't change).
I'm not sure if both occurrences are intentional or a bug. I would think the auto lean would act on actual leaning but not on shock absorption (I know shock absorption is leaning in relation to the bike, but I think the difference is stil clear).
Quote from: geofanatec on January 04, 2017, 05:34:46 PM
Kinda cool to, you can adjust the leanspeed (which someone thought it might be slow I think).
That was me! Nice find Geo. I'll play with it.
Quote from: geofanatec on January 04, 2017, 05:34:46 PM
In rider.cfg
You can adjust the stability (auto f/b l/r movements) and experiment with different factors and rates.
I was adjusting values like mad trying to figure out why it didn't work. I guess you need Auto Lean on :p Thanks!
I'm getting some pretty hilarious results, but haven't found the settings I am looking for so far. I can land quitupule backflips tho :p If anyone figures out how to soften the rider's arm on bumps, let me know.
Anyhow, the lean stick should not prevent the arms of the rider from absorbing shocks. Nor should having auto lean off.
Well damn.. I spent a good 40 minutes trying different values, not getting results and figuring the settings weren't what I thought they were.
Problem is, I don't want to ride with auto lean :(
Maybe this belongs in the bug reports?