• Welcome to PiBoSo Official Forum. Please login or sign up.
 
March 28, 2024, 06:22:56 PM

News:

World Racing Series beta14 available! :)


Realism in game

Started by matty0l215, August 30, 2016, 12:38:26 PM

Previous topic - Next topic

Do you care if the in game bikes are realistic to real life?

Yes, In game bikes should use real world data as much as possible.
45 (95.7%)
No, the bike should be easy to ride whatever it takes.
2 (4.3%)

Total Members Voted: 45

Hawk

Quote from: h106frp on August 30, 2016, 07:46:52 PM
More consistent handling requires the 'direct lean' issue to be resolved, It really needs to be removed so that bikes can only react within their real life correct physical limits whatever the controller type.

That is not possible. Direct lean is there for people to be able to scale the lean depending on the length of the controller axis they are using.... I mean if Pib set it to only work properly for DD's controller then you guys with an Xbox Controller would be right up the good old swany! Hehe!  ;D

Hawk.

h106frp

August 30, 2016, 08:46:51 PM #16 Last Edit: August 30, 2016, 08:52:40 PM by h106frp
Problem is that we have a controller setting that (can be used to) effects bike simulated handling.

You should be able to scale controller throw rates but it should not feed back into the way the simulation behaves. For a given geometry and mass distribution a bike should have the same roll rate no matter what the controller used.

Maybe that's the problem, at the moment we seem to have a direct 'steering' input maybe what we should be asking for is how much steer input the virtual rider can provide in a given situation

HornetMaX

I don't get what you mean h106: at the moment we have a direct lean input (if you set it to 100%), but it acts within the limitations of the virtual rider (max applicable steering torque).

What's the "direct lean issue" ?

h106frp

Maybe i just do not understand GPB control philosophy, why does altering the value of direct lean alter bike behavior and why is it really needed? Surely full controller deflection should be 'maximum steer input' no matter what the controller design.

It almost seems we have 2 things trying to happen at once, in one instance we have a system that controls steering input demand via the virtual rider and the other requirement appears to be for a control where the 'stick' angle matches the on screen lean as required by DD's system. The direct lean configuration slider does not seem a good solution for either approach.



Vini

August 31, 2016, 02:28:10 AM #19 Last Edit: August 31, 2016, 03:55:06 AM by vin97
Quote from: h106frp on August 30, 2016, 08:46:51 PMFor a given geometry and mass distribution a bike should have the same roll rate no matter what the controller used.
And the direct lean setting is simply an aid that limits the maximum "roll rate" by telling the virtual rider to take it a bit slower, giving you more stability. It does not affect how the suspension reacts to  forces or anything else physics related.

Quote from: h106frp on August 30, 2016, 10:20:19 PM
Surely full controller deflection should be 'maximum steer input' no matter what the controller design.
You mean 'maximum lean angle input', right? Otherwise you are talking about direct-steer, for which we currently have no proper controller.
I would describe it as this: You are telling the virtual rider where you want to go with your controller and the direct-lean value (nothing to do with direct-steer) decides how quickly he makes the bike go there (of course if lean angle input is not maximum, the process is even slower). In case of direct-lean 100%, there is no more 'buffer' and the 'lean angle input' is directly related to the current bike lean angle, meaning that you decide how quickly the bike goes there. This also means that you can input roll rates that are higher than what is possible for a motorcycle, which then leads to crashes.

I really don't see anything wrong with that, in fact, I'm pretty sure that a lot of people would have big problems if it weren't for that direct-lean slider and everybody was forced to ride with 100%.

HornetMaX

Quote from: h106frp on August 30, 2016, 10:20:19 PM
Maybe i just do not understand GPB control philosophy, why does altering the value of direct lean alter bike behavior and why is it really needed? Surely full controller deflection should be 'maximum steer input' no matter what the controller design.

It almost seems we have 2 things trying to happen at once, in one instance we have a system that controls steering input demand via the virtual rider and the other requirement appears to be for a control where the 'stick' angle matches the on screen lean as required by DD's system. The direct lean configuration slider does not seem a good solution for either approach.
Unless you're using DST or DSA, your input is the target lean angle.

Simplifying to the extreme (it's surely not so simple), direct lean @ 100% means the output of your stick is not filtered at all before reaching the virtual rider: this means that if you yank instantly your stick from full right to full left, the virtual rider will see a very sudden change in his input. When you put direct lean at 0%, your stick's output will be filtered a lot before reaching the virtual rider, so the virtual rider will see a smoother (and slower) change in its input (the target lean).

The more you filter your joystick output, the more the bike behaves smoothly. But it also become less responsive: if you filter too much, you may not be able to reach the same roll rate (the max one defined by the bike inertias and the virtual rider limitations). Over-simplifying again, you can see direct lean as a low-pass filter: putting it a 0% (0% means max filtering) means smoother, less responsive output (it's slower, there's a lag/delay).

But notice that, no matter how much you filter with the direct lean setting, then it's up to the virtual rider (which more or less is itself a "filter", but in a feedback loop): depending on how you tune the virtual rider you can have very strange behaviors.

Bottom line: when you see something strange, most likely the "fault" is either on you moving the stick too quickly or on the virtual rider doing something wonky. With Direct Lean at 100% you can move your stick slowly and produce the same virtual rider input as when you have Direct Lean at 0% and yank the stick as a madman. But the dual is not true: with Direct lean at 0% you can't do everything you can with it at 100%.

Quote from: vin97 on August 31, 2016, 02:28:10 AM
I really don't see anything wrong with that, in fact, I'm pretty sure that a lot of people would have big problems if it weren't for that direct-lean slider and everybody was forced to ride with 100%.
Or with 0% :)

doubledragoncc

I want to try direct lean at 150% then I think it would be better for my system, but that is why it is needed. Different controllers cause a different reaction in GPB, it is a needed factor no matter what folks say

DD
GPBOC Live Streams: https://www.youtube.com/c/IASystemsComputerControls; i7 12700K 5.1GHz Z690 ASUS Strix Z690-A Mobo 32GB 3600MHz DDR4 RAM ASUS Strix RTX3080 OC 10GB DDR6X ASUS Ryujin 360 AOI Cooler ROG Thor 1200w PSU in ROG Helios Tower Case.

h106frp

And it think this leads up back on topic;
quote; 'This also means that you can input roll rates that are higher than what is possible for a motorcycle, which then leads to crashes.'

I think this ability to force high roll rates leads some people to believe that all the models should be capable of matching their control inputs even if it is beyond what the 'real life values' will allow.

It is not difficult to make models for GPB that can operate beyond realistic limits, a combination of .geom, .cfg and controller fiddling can manufacture bikes that unrealistically capable - verging on RIDE style handling.

Hawk

Personally I think with the short Xbox sticks and lean set at 100% then the roll rate that can be achieved is massively more than could be achieved in reality, but with a controller like DD's, it would not be possible to achieve an over realistic roll rate because of the length of the axis input for leaning the bike results in a naturally slower roll axis rate input. I think this is why DD would like to see an even greater roll rate setting(Direct Lean) so he can have a bit of a buffer to play with and setup rather than being at the limit needed for his controller at the current 100% direct lean setting? Am I correct with this supposition DD?

Hawk.

HornetMaX

Quote from: Hawk on August 31, 2016, 07:58:46 AM
Personally I think with the short Xbox sticks and lean set at 100% then the roll rate that can be achieved is massively more than could be achieved in reality,
Could be, but it means the problem is not in direct lean, but in the virtual rider.

Quote from: Hawk on August 31, 2016, 07:58:46 AM
but with a controller like DD's, it would not be possible to achieve an over realistic roll rate because of the length of the axis input for leaning the bike results in a naturally slower roll axis rate input. I think this is why DD would like to see an even greater roll rate setting(Direct Lean) so he can have a bit of a buffer to play with and setup rather than being at the limit needed for his controller at the current 100% direct lean setting? Am I correct with this supposition DD?
Direct lean at 150% may not make sense: with a passive filter (one that does not "amplify"), you can't filter less than zero (Direct lean 100% = zero filtering).
If after having set Direct lean at 100% you still need something snappier, the only solution is change the virtual rider.

Keep in mind that DD's system looks like a real handlebar but it still dictates to GPB the lean angle (unless you're using DST/DSA) ...

Vini

August 31, 2016, 08:49:38 AM #25 Last Edit: August 31, 2016, 08:56:07 AM by vin97
Quote from: HornetMaX on August 31, 2016, 07:09:53 AM
Quote from: vin97 on August 31, 2016, 02:28:10 AM
I really don't see anything wrong with that, in fact, I'm pretty sure that a lot of people would have big problems if it weren't for that direct-lean slider and everybody was forced to ride with 100%.
Or with 0% :)
not really, you just slam the stick to one side and the vr smoothes everything out for you.
it just feels slow and spongy.

Quote from: h106frp on August 31, 2016, 07:31:57 AMI think this ability to force high roll rates leads some people to believe that all the models should be capable of matching their control inputs even if it is beyond what the 'real life values' will allow
I don't think anyone does that. Getting tankslappers when you lean the bike 20° in 3 seconds on the COTA straight is not realistic, though.

Quote from: HornetMaX on August 31, 2016, 08:19:07 AMDirect lean at 150% may not make sense: with a passive filter (one that does not "amplify"), you can't filter less than zero (Direct lean 100% = zero filtering).
If after having set Direct lean at 100% you still need something snappier, the only solution is change the virtual rider.

Keep in mind that DD's system looks like a real handlebar but it still dictates to GPB the lean angle (unless you're using DST/DSA) ...
+1

I would add to that, that if a controller does not mechanically allow extremely quick movements (or more specifically, rotation), there is not much the game can do. You either have to reduce friction or range of the controller.
The VR is not the limitation because he makes the bike crash if you input extreme roll rates.

HornetMaX

Quote from: vin97 on August 31, 2016, 08:49:38 AM
Quote from: HornetMaX on August 31, 2016, 07:09:53 AM
Quote from: vin97 on August 31, 2016, 02:28:10 AM
I really don't see anything wrong with that, in fact, I'm pretty sure that a lot of people would have big problems if it weren't for that direct-lean slider and everybody was forced to ride with 100%.
Or with 0% :)
not really, you just slam the stick to one side and the vr smoothes everything out for you.
it just feels slow and spongy.
Which would make (at least some) people unhappy, no ?

Vini

yeah, would be pretty boring and not worthy of a "realistic motorcycle sim" IMO


doubledragoncc

Due to the physical action of my system and to allow the user to use it in different positions, ie couch, desk, sit on frame you have to keep in mind I dont let it use a full 60plus degrees each side. It is impossible physically to use and has to be taken into account.

I have been building systems for years and tried it all, but my SPSS system is the only way to be able to ride without actually doing muscle damage as a real bike axis would do on the frickin couch!!!

because I have to reduce the  lean angle, but have a longer axis movement, 150% direct lean would  make my system be faster like a gamepad but still keep the relationship of the lean angle as realistic as possible without over exaggeration of the riders lean speed.

I have been delayed getting Matty's system finished and hopefully this weekend it will be, but once he gets on it, I am sure as he has used a gamepad until now, he can explain better for all concerned and maybe help out on this subject.

DD
GPBOC Live Streams: https://www.youtube.com/c/IASystemsComputerControls; i7 12700K 5.1GHz Z690 ASUS Strix Z690-A Mobo 32GB 3600MHz DDR4 RAM ASUS Strix RTX3080 OC 10GB DDR6X ASUS Ryujin 360 AOI Cooler ROG Thor 1200w PSU in ROG Helios Tower Case.

HornetMaX

Quote from: doubledragoncc on August 31, 2016, 10:33:31 AM
because I have to reduce the  lean angle, but have a longer axis movement, 150% direct lean would  make my system be faster like a gamepad but still keep the relationship of the lean angle as realistic as possible without over exaggeration of the riders lean speed.
But can you describe exactly what do you mean with 150% direct lean ? What would you want it to do ?

Today 0% direct lean means "maximum filtering" of the input, 100% direct lean means "no filtering" of the input. What would 150% direct lean mean !? You can't filter less than nothing.

If your device has 60deg each side, then there's no "magic filter" that can cure that. It has a long throw so it's probably very very accurate for small movements (which is good), but if you want to flick it fast, then you can't ask GPB to flick the bike faster than the time you yourself need to flick your device: it wouldn't make sense.