• Welcome to PiBoSo Official Forum. Please login or sign up.
 
March 28, 2024, 10:44:24 AM

News:

GP Bikes beta21c available! :)


Physics observations....

Started by CapeDoctor, July 13, 2016, 05:41:38 AM

Previous topic - Next topic

CapeDoctor

July 13, 2016, 05:41:38 AM Last Edit: July 13, 2016, 07:29:35 AM by CapeDoctor
well guys, been playing b8 since it's release, both the stock and modded bikes and tracks.
generally it's been a pleasant experience, with most of the serious issues of b7 sorted.
i figured it's time to discuss some of the current physics issues, purely in an attempt to try and help with improvements - this is not a thread for just pointing out problems, but will hopefully lead to positive discussion, from which we can work forward through some of the apparent physics issues.
my first and foremost issue with the bike physics in it's current state is how it handles weight shift/transfer, and more specifically, how it handles the bike going 'light' or unloading weight, in certain situations.
while it's taken me some while to initially get used to the new physics, and generally the feeling is good (way better than anything else i've experienced with a pc, it still seems to be the most common 'unforced error' type of crash for  me, which is the thing that shatters my illusion of being on a bike.
usually it happens when i flick the bike one way and then another quickly through a set of turns - the first rapid flick, say, left to tight, works beautifully - it's when i flick it back the other way - tight to left, in this example, that it loses all sense of cohesion - the bike seems to just lose any weight at all, and virtually does a 180 roll - so while i am quickly moving it in one direction, it just decides to be really helpful to me, and flips the entire bike over - something which is physically impossible in the real world.
this is by far the most illusion-breaking issue for me - it doesn't always happen, either, btw - only on occasions when i really flick it back and forth very quickly using a lot of force, that it happens.
at other times, and less so, i have had the bike suddenly do strange things when i make a move that allows a side of the bike to go unloaded, weight-wise, and things go pear shaped.
for me, the common denominator with these issues is how the sim interprets the bike 'losing weight' like this in certain situations.
i believe physics wise, this is for me, one of the biggest issues with the physics.
let's discuss - i'm always willing to accept that i might be at fault, lol    :D

HornetMaX

When you see the problem, do you go left-right and then right-left (the problem happening on the last flick, right-left) or do you go vertical-right and then right-left ?

CapeDoctor

mmm...... i'll have to test some more to know exactly what you mean, Max, but as far as i can tell so far, it happens when i do a quick left-right, then right-left flick in quick succession (or vice versa, of course)
i'm pretty sure it doesn't happen if i go left-right, then vertical right, then right left.
lol, this could get confusing.
i can probably post up a vid if i find some time, to better illustrate.
it's like on one flick too many in quick succession, the sim kind of loses the plot, and the bike seems to become weightless, allowing it to almost just roll completely over on its own axis - which is physically impossible in the scenarios i'm referring to.(or any, for that matter)

h106frp


CapeDoctor

my direct lean is on 18 percent

h106frp

You would think that at 18% the rate of change would pretty much be under the control of the VR system

CapeDoctor

July 13, 2016, 08:31:41 AM #6 Last Edit: July 13, 2016, 08:37:52 AM by CapeDoctor
the problem is that this doesn't happen often, but those times when it does happen, it entirely shatters any sense of immersion/simulation.
i should maybe not have made a blanket statement of all bikes earlier - it's possible this only happens with some bikes - for sure, there are some bikes i know it's never happened (at least to me) and others where it's happened more than once, on bikes that otherwise perform very nicely.
as i said, i'll have to investigate more, and be sure to make replays when i see it again.
i might have to trawl through my old replays, as well - i'm sure there are one or two instances on record, so to speak...  ;)


ps note - i only use direct lean for long corners, otherwise i'd be doing full direct lean myself - this, again, is simply down to a controller issue, which can only really be cured by using something like DD's system. it's simply not doable for me to keep a little toggle stick in the right place constantly during a long sweeping corner between my thumb and forefinger - and i find myself constantly having to adjust this stupid fukkin' toggle thing all the way through the corner, when in reality this would be an almost natural process.
a gamepad may be better than a keyboard for GPB, but it's still a sucky compromise, imo.
this sim would be in a world of its own if we all had decent controller systems.   8)

HornetMaX

In real life, when you flick the bike very fast you do lose some load on the wheels, that's normal.

I'm under the impression that you're just flicking the joystick too quickly from left to right.
But I fear that we can't dig this more unless you show a video.


Vini

July 13, 2016, 07:01:09 PM #8 Last Edit: July 14, 2016, 03:35:06 AM by vin97
direct lean 100% + manual rider lean makes fast changes of direction very difficult.
i know the "barrel roll" you are talking about very well, cape :D
the roll itself looks unrealistic but i think that it's not really important. fact is, a bike cannot change direction instantly and if you force the virtual rider to try to do so, you will crash. i think that's totally fine.

h106frp

Be interesting to see if you can plot the roll rate from the telemetry, this article suggests 60deg/second is not unreasonable
http://www.insidemotorcycles.com/blogs/item/1020-roll-rate-how-quickly-can-you-turn?.html

I wonder if a GPB plot would look similar.

HornetMaX

Quote from: h106frp on July 13, 2016, 07:11:52 PM
Be interesting to see if you can plot the roll rate from the telemetry, this article suggests 60deg/second is not unreasonable
http://www.insidemotorcycles.com/blogs/item/1020-roll-rate-how-quickly-can-you-turn?.html

I wonder if a GPB plot would look similar.
You have the rates (roll, pitch, yaw) in the telemetry output, so you definitely can check that  8)

If you open the sample telemetry file that comes with MaxTM (a fairly slow lap on Victoria with the 990), you can see roll rates up to 90deg/s (and no crash).


Grooveski

July 13, 2016, 09:21:29 PM #11 Last Edit: July 13, 2016, 09:28:07 PM by Grooveski
I recently upped my direct lean.  Couldn't handle 100% but at 60 it's feeling good.  Took a bit of getting used to but a couple of hundred laps of Suzuka helped readjust my thinking.

Has definately made the rider input more sensitive(or more likely to cause problems if care isn't taken).  To aviod the barrel roll I've taken to forcing a tiny pause in the bike's upright position rather than throwing it from one side to the other.  It's barely noticable, may not even be noticable at all on a replay(i.e. the bike may not have even got to the upright position yet - I'm just waiting on it before doing the lean into the second corner)

More noticably affected was the washout crash where the bike doesn't pick up from the first corner.  (Turn 3-4 at Suz' was torture at first)
With that one it's all about picking the bike up a touch before moving the rider at all(about 10° lift and you  seem to be into safe territory).
...and being smoother with the initial pick up from full lean.

So what was once just pushing both bike and rider from one side to the other is now:
Gentle on the bike lift - delay before moving rider - move rider from one side to other - briefly pause bike at upright - lean bike into second corner.

...which sounds like an awful footer when it's all typed out but with enough practice should all become second nature.

Funnily enough there was an wierd crash just recently(BSB at Oulton maybe?) that had me thinking of GPBs barrel rolling tendancy.  The guy picked up the bike so forcefully that by the upright position both wheels were off the ground then the bike just slammed over the rest of the way onto it's side.  Wasn't exactly the same crash but was very similar.

Vini

good observations, grooveski.
high direct lean values requires more awareness of what's happening (to the bike, the wheels, the suspension, the rider) but of course being able to do quicker maneuvers can be of great advantage.

that bsb crash sounds exactly like what happens in GPB.

CapeDoctor

well, i'm relieved that at least some folk now what i'm talking about - saves me from having to get too in-depth to explain it properly, lol
barrel roll is indeed a more fitting description for what i'm talking about  ;)
anyways, i've titled the name of this thread as 'Physics observations', as it's meant to be just that.
a thread for discussing things we see in the sim that just don't equate to reality.
it's not for finding out if it's the rider's fault/style or bad controller settings or whatever the reasons might be, it's more for seeing something that isn't well represented in the sim - purely for the purposes of seeing it improved upon.
now, i'm going to go and hunt for that BSB incident to see it for myself, but in general, it shouldn't be physically possible for a bike to roll over on itself like that - it'd make a lot more sense if the bike just lowsided itself when turning the other way, but the way it currenetly looks in the sim, is beyond fantastical, lol

and please don't get me wrong, folks, this thread is meant purely for us to discuss things logically, in no way is it meant to be a diss on the sim - as it stands, in my mind, GPB is head and shoulders above anything else in this market, and i have little doubt that one day it will rule supreme - if it can be improved upon through things like us discussing possible issues, and hopefully getting them resolved.
if it looks like unravelling into a bitch-fest at some point, then admins feel free to remove it, it's only meant to be a place in the forum where we can discuss possible issues, without having to start a fresh thread for each possible new issue - real or perceived, lol  ;D

BOBR6 84

July 14, 2016, 04:42:29 PM #14 Last Edit: July 14, 2016, 09:10:37 PM by BOBR6 84
yeah GPB is a detailed beast.. you can't really force a bike to do something if it isn't ''ready'' for it (irl).. suspension looks after the bike and you.. but any extra harsh movements will just ruin everything.. it's a fine art riding bikes  8) the direct lean settings point this out quite well in GPB, in comparison to riding like a tit in real life..