• Welcome to PiBoSo Official Forum. Please login or sign up.
 
February 27, 2024, 05:22:18 PM

News:

GP Bikes beta21c available! :)


MaxHUD plugin

Started by HornetMaX, September 26, 2013, 04:34:50 PM

Previous topic - Next topic

HornetMaX

Quote from: Vini on August 11, 2021, 03:05:42 PM
Quote from: HornetMaX on August 11, 2021, 11:46:17 AMWhat do you mean ? Rumble if front and rear do not have the same wheelspeed ?!
Yes and make scale adjustable. So just RumblePercent=100*Scale*(1-FSpeed/RSpeed)
Don't worry about wheelie and gyro and yaw and blablabla.
That would be stuff for later if the feedback turns out to be of use.
Sounds like a terrible plan to me: suspect idea to start with + clear problems in many situations you're just asking to ignore (at least for the moment).

I'm not even sure what is the target in principle here: lateral slip ? longitudianl slip ? both ?

Myst1cPrun3

Quote from: HornetMaX on August 11, 2021, 11:46:17 AMBad feedback is not better than no feedback.


While I would like to see some feedback for tyre behaviour (As I said I look to rumble for a "Fill In" for the SoTP feel that is lost) I completely agree, that bad feedback is less useless than no feedback.

If there's one person that knows about this stuff here its you mate, so if it ain't possible in a good way then its a shame but that's the way it is. :(

Quote from: Vini on August 11, 2021, 03:05:42 PMOnly one way to find out.

If I had to guess, he sounds like he's already tried it... Or something similar.

Vini

Doesn't sound to me like he tried it.
I tried it through SimHub and it works fairly well. Not perfect but infinitely better than nothing. Problem is I can't redirect it to the gamepad.

HornetMaX

Quote from: Vini on August 15, 2021, 08:48:25 AMDoesn't sound to me like he tried it.
OK.

Quote from: Vini on August 15, 2021, 08:48:25 AMI tried it through SimHub and it works fairly well. Not perfect but infinitely better than nothing. Problem is I can't redirect it to the gamepad.
Well, if you already tried you may want to share exactly what you did.

I still think having something that rumbles when the ratio of front and rear wheelspeed changes is bad more often than good.

One thing you can do: do a decent lap, save the telemetry with MaxTM and have a look at front / rear wheelspeed graph (*): I did and I don't see anything that looks usable.

(*) I know, I should improve MaxTM to allow plotting multiple lines on the same axis. That would be useful (in our case and in general). I may do a temp hack and add a couple of new signals (like front/rear wheelspeed difference and ratio), just to ease looking into that. Will have to wait when Im back from holiday: if a new release of GPB is out I'd need to update all plugins anyway.

Vini

August 20, 2021, 09:48:51 AM #919 Last Edit: August 20, 2021, 06:17:33 PM by Vini
Quote from: HornetMaX on August 19, 2021, 04:52:04 PMOne thing you can do: do a decent lap, save the telemetry with MaxTM and have a look at front / rear wheelspeed graph (*): I did and I don't see anything that looks usable.
Well, are you actually sliding the rear on a regular basis?
For me, it's part of my riding style so I induce a little slide almost every corner (sometimes bigger ones).
The problem is judging when it's getting too much and reacting in time before the bike highsides, which happens to me a lot if I'm not careful. I know other guys with high TC values that basically never loose the rear or highside, so in that case you are probably right in that the wheelspeed rumble wouldn't provide much useful feedback.

The SimHub test was just a very crude one and I had to use my subwoofer since I don't have a real rumble motor.

I just loaded a good lap of mine into MaxTM but since there is no way to overlay two curves, it is impossible to tell if the Rspeed-to-Fspeed ratio could be of any use...


I don't understand why you can't just simply add an experimental rumble mode without any advanced tweaks or algorithms and then this whole discussion can be settled in one minute.
We already spent more time discussing this than it would take to code this alternative mode.

Vini

August 20, 2021, 10:10:11 AM #920 Last Edit: August 20, 2021, 10:23:55 AM by Vini
I just checked a few places in the graphs and manually calculated the R/F ratios.
The result strongly correlates with slide angle and I can easily tell where the corners are that I ride clean and where the ones are that I tend to induce a slide.
It is clearly visible.

Now, granted even with zero slide on a straight the rear is still a little bit faster. But the difference to an actual slide (and I don't mean a huge one close to highsiding) is more than ten fold.
So in addition to a variable scale (that determines sensitivty and therefore when the 100% rumble value is reached) you could add an offset setting for the 0% value.

I will uplaod my lap and the data I collected shortly.

Vini

August 20, 2021, 11:51:22 AM #921 Last Edit: August 20, 2021, 03:52:20 PM by Vini
I tested the new MotoGP bikes at Silverstone yesterday. This data is taken from my fastest lap (2.03.4) so I was pushing the rear even when not intentionally sliding.

Here are a few rear/front ratios I calculated:
T1 apex (max lean, minimum corner speed, no throttle): +1,8%
End of straight upright (full throttle): +1,7%
Exit onto backstraight (clean): +3%
T10 exit (right at the slip limit): +9%
Exit onto straight (shortly breaks traction): +12,7%
Turn 12 exit (intentional slide): +19%

Video of the 4 exits in the same order: https://streamable.com/ivspar

Now as you can see, even the last slide was relatively subtle. Not even close to a "stunt level" powerslide. Just about what every MotoGP rider does to get that extra bit of rotation.
But even then you can clearly see the correlation and the sensitivity perfectly scales with the intensity of the slide.

The data is clearly there and if translated to rumble it would provide great feedback over the limit of the tyre.
This would not just be an on/off switch that triggers randomly or when it's already too late.


I don't know if this ratio based rumble could even help detect underrotation/slip of the front under braking (maybe Fspeed/BikeSpeed ratio) but it's not that critical there anyway since the current chassis based rumble already works pretty well for detecting the limit under braking by feeling the front chattering.
So adding the R/F ratio rumble on top of that would be the perfect combination.

Myst1cPrun3

Quote from: Vini on August 20, 2021, 09:48:51 AMI don't understand why you can't just simply add an experimental rumble mode without any advanced tweaks or algorithms and then this whole discussion can be settled in one minute.
We already spent more time discussing this than it would take to code this alternative mode.


If it's so simple and quick why don't you give it a go yourself? You seem to want it bad enough.

Vini


Tom HWK

Whilst we are on about shite ideas, would it be possible to add a light that you can enable in MaxHUD for tyres overheating? since there is a tyre temp sensor. Either be able to set the temp the light comes on yourself or any way of it pulling straight from the tyre values? Just wondering as its something MotoGP bikes have and its easier to tell when you are burning your tyres up than trying to read the numbers at race pace.

Vini

If you had the rumble mode I suggest, you wouldn't need a light cause you'd feel it ;D

HornetMaX

Quote from: Tom HWK on August 20, 2021, 05:07:10 PMWhilst we are on about shite ideas, would it be possible to add a light that you can enable in MaxHUD for tyres overheating? since there is a tyre temp sensor. Either be able to set the temp the light comes on yourself or any way of it pulling straight from the tyre values? Just wondering as its something MotoGP bikes have and its easier to tell when you are burning your tyres up than trying to read the numbers at race pace.
You mean a light on the bike dashboard ? That woul dbe for bike modders (I'm not not sure if it is doable).

In MaxHUD you have the tyre temps (front/rear x left/middle/right) in the ECU widget.

HornetMaX

Quote from: Vini on August 20, 2021, 09:48:51 AMI don't understand why you can't just simply add an experimental rumble mode without any advanced tweaks or algorithms and then this whole discussion can be settled in one minute.
We already spent more time discussing this than it would take to code this alternative mode.
I don't understand plenty of things. That's life.

Quote from: Vini on August 20, 2021, 11:51:22 AMVideo of the 4 exits in the same order: https://streamable.com/ivspar

Now as you can see, even the last slide was relatively subtle. Not even close to a "stunt level" powerslide. Just about what every MotoGP rider does to get that extra bit of rotation.
But even then you can clearly see the correlation and the sensitivity perfectly scales with the intensity of the slide.

The data is clearly there and if translated to rumble it would provide great feedback over the limit of the tyre.
This would not just be an on/off switch that triggers randomly or when it's already too late.

Thanks for looking into this, but I'd say that some data is there, that's far from proving anything. For example:

How stable are your ratios along the slide ? If the ratio signal is very 'noisy', translating it into rumble may lose all useful information.
What happens in case of mild wheeling on a straight or exting a turn ?

That aside, can you upload the MaxTM data files (the two of them) of your lap/session somewhere so that we can use the same data ?

Quote from: Vini on August 20, 2021, 09:48:51 AMI don't know if this ratio based rumble could even help detect underrotation/slip of the front under braking (maybe Fspeed/BikeSpeed ratio) but it's not that critical there anyway since the current chassis based rumble already works pretty well for detecting the limit under braking by feeling the front chattering.
So adding the R/F ratio rumble on top of that would be the perfect combination.
Hmmm, the front (longitudinal) slip detection under braking sounds more reasonable, in fact.
Speculating here, but I suspect this to work fairly well (still, worth to check some data).
I'm still not 100% convinced we actually want this but at least in principle, this should work better.

Last point: you said "adding R/F ratio rumble on top". I'm not sure we can "add on top".
Yes, PC pads have two rumble motors (with different masses hence different freqencies/vibrations) but I'm not sure using the two for two different purposes results in something actually usable.

Tom HWK

Quote from: HornetMaX on August 22, 2021, 02:40:17 PM
Quote from: Tom HWK on August 20, 2021, 05:07:10 PMWhilst we are on about shite ideas, would it be possible to add a light that you can enable in MaxHUD for tyres overheating? since there is a tyre temp sensor. Either be able to set the temp the light comes on yourself or any way of it pulling straight from the tyre values? Just wondering as its something MotoGP bikes have and its easier to tell when you are burning your tyres up than trying to read the numbers at race pace.
You mean a light on the bike dashboard ? That woul dbe for bike modders (I'm not not sure if it is doable).

In MaxHUD you have the tyre temps (front/rear x left/middle/right) in the ECU widget.

I was more on about a light that could light up when the tyre temp is above the tyre max temp that is just in maxhud, like the tc and aw lights, as the dash would be a piboso thing wouldn't it?

Vini

August 22, 2021, 05:37:20 PM #929 Last Edit: August 22, 2021, 06:11:56 PM by Vini
Quote from: HornetMaX on August 22, 2021, 03:14:41 PMThanks for looking into this, but I'd say that some data is there, that's far from proving anything. For example:

How stable are your ratios along the slide ? If the ratio signal is very 'noisy', translating it into rumble may lose all useful information.
What happens in case of mild wheeling on a straight or exting a turn ?
You can see in the "clean" exit there even was a small wheelie. There are multiple solutions for this problem that we already discussed.: Use front suspension travel or that contact surface parameter you were talking about for an if clause or scratch RSpeed entirely and use BikeSpeed.
You are way overthinking this again. It's not like the wheel speed is randomly jumping up and down 20km/h from one frame to the next. After all it's acceleration is governed by physical laws. Obviously it's not perfectly smooth but if it was it would be useless cause the precise feedback would be filtered away or come delayed.
The important thing is that the ratio for a clean exit would never suddenly jump to that of a slide exit sicne the difference in ratio is so high as you can see in the examples I gave. So even if it's noisy, the ratio is sensitive enough as to never give fundamentally wrong feedback.

Quote from: HornetMaX on August 22, 2021, 03:14:41 PMHmmm, the front (longitudinal) slip detection under braking sounds more reasonable, in fact.
Speculating here, but I suspect this to work fairly well (still, worth to check some data).
I'm still not 100% convinced we actually want this but at least in principle, this should work better.
So then why should the rear-speed based rumble not work?
If it's only the wheelie thing, then you could use bike-speed instead of front wheel speed for the ratio there, too.

Quote from: HornetMaX on August 22, 2021, 03:14:41 PMLast point: you said "adding R/F ratio rumble on top". I'm not sure we can "add on top".
Yes, PC pads have two rumble motors (with different masses hence different freqencies/vibrations) but I'm not sure using the two for two different purposes results in something actually usable.
This is all stuff for later and there are different ways one could approach this. But since the two rumble signals would be quite different in nature (short pulses from the chassis accel vs gradually increasing vibration from the wheelslip/speed) one would still get intuitive feedback even if simply layed on top of each other on the same motor.

Here is the link to the telemtry and the replay: https://mega.nz/file/EZkGSYyD#jhYSgKkkSMDx8ihdL_nNsECyffoBsUG1j05zGmT6X-Y

...It's still funny to me how you are willing to tediously look through the telemtry graphs to find an error in the theory when you are basically one line of code away from the functional practical application. All of the bike data and rumble translation code is already there, only the formula needs to be changed.