FlyPT Mover

Thanks @pmvcda. I'll give it a try.

However this issue with speed is that... it's just speed. Meaning that when you jump on the throttle from a stop you'll get a weak signal because speed is at 0 whereas the acceleration value actually jumps.

It's the issue with speed that has to be continuous - physically speaking - while acceleration is not and can jump.

I have doubts that any static filter will achieve this effect. Btw, a "static" filter is one that has constant (static) filter parameters.

Instead of asking "Which filter with which parameters do we need?" , maybe we need to ask the question: "How and when would the filter parameters need to change?"

So this is the result of considering only growing values with a basic washout (multiplier with .995):



Not solving the issue... It's not so simple to make it.
That's what I have right now.
Acceleration, followed by releasing the pedal in DirtRally2 with acceleration

If we put a slower washout, then that deceleration is completely ignored...

Here with bigger washout:



Here with some brakes and accelerations. Seems like a spike filter....
Maybe we should try with a spike filter.

Thanks a lot @pmvcda but there is no emergency, don't feel any stress about this you are already providing great value to us all with Mover.

You can take some time to think about it and also maybe do something else with your life meanwhile if you want to.

Anyway, I believe the chart below that I provided earlier could be useful to help.

Btw - sorry, another request - you could maybe add in the wave types in the source loop the wave that would be a square with pauses between positive and negative alternances, i.e. 0 for 250ms, X for 250ms, 0 for 250ms, -X for 250ms and so on.

`

Thx
Alex

You can achieve that wave with the superimposition of two:

Add them in the pose

“Or you could say, it captures the acceleration data only when it growing either in a positive or negative direction.”


I am smelling that this could be an old @yobuddy quote from times when Simtools washout had a dark side. This pumping effect was than corrected so we could interface 6-dof plugins.

If this was an error in math or a specific washout approach from non-kinematic era, he is the one who can provide a response.

This is all wrong as you can see in the explanation above. Nevertheless I am open to search for alternatives.

I think it has less to do with friction. Even in outer space where two blocks are colliding I would expect the accel to return to 0 just after impact.

Yes, the accel. goes back to zero, but doesn't cross zero and go negative. The new speed after the acceleration doesn't change again until another collision or getting to close to a gravity well.

In a simulator that would be the first kick. The stopping at the end of travel is already not realistic anymore, even less the going back to home position. A full cycle in the sim is not one acceleration event, but four:
(Let's call the direction of the first acceleration event positive)

• a1 (positive): First kick, platform accelerates away from position zero
• a2 (negative): Platform slows down until stop at end of travel
• a3 (negative): Platform accelerates back in direction of position zero
• a4 (positive): Platform slows down to full stop at position zero

The space rock would only experience a1 unless it was sitting in a strong gravity well and being pulled back immediately after the collision and collide there again with another object causing it to stop completely and not just oscillate around.
Acceleration going back to zero is not a second force, just the removal of force. Change of speed would be an additional force.

It's not about a2 - a4 existing, they're necessary in a sim, it's about making them as unnoticable as possible. The bigger a1 the bigger a2 - a4 have to be unless the simulator is really large.
That's also the argument for G-Seats as force (simulated acceleration) removed is just that - no more force.

It the data coming from the game is accurate, the only reason the back kick crosses the 0 on the simulator graph, could be that it simply adds the presence of back accel. to the washout which is supposed to act beyond the threshold of perception.

Like I see on the real car example, the back cue is for me a completely natural phenomenon and not a false cue.

a2 and a4 is the washout which acts beyond of threshold of perception. What’s wrong with it ?

a2, a3 and a4 are the washout, yes. They were never the problem. The problem was that the game reporting the acceleration being back at zero amplified a3 and therefore a4 in an unwanted manner raising them beyond the threshold of perception. That's why it was proposed that tying the reported acceleration directly into an axis for translational motion might not be the answer, but to calculate it indirectly from change in velocity.

That's why I had the idea to do it with the derivative of acceleration and use the sign as a way to ignore one way, which was way too complicated considering ∆v is already known ...

It is both ...kinda... Depending on how you want to perceive it. A gear shift itself does not generate an actual measureable backwards acceleration (ignoring aero drag for now). Only human perception interprets it as such.

But you guys see the symmetry here? The sudden absence of acceleration in a car is intuitively interpreted as a "kick back", and the same "kick back" on the sim could intuitively be interpreted just as the sudden absence of acceleration And on an ideal sim (no vibration, no outside visual, no noise, huuuuge operating space,...) those would be indistinguishable.

Here, it is only the fact that the occupant does have other references that keep him from interpreting it as plausible.

I know that this is not directly a solution for @alex928gt. But the realisation that a good sim should try to make it as hard as possible for the occupant to make this distinction is (I believe) one of the core concepts in motion sim building.

I think of it like,... playing a prank on my own brain!

Oh,... I'm soooo going to fool this stupid dude!

I'm gonna leave him no f'!%&§ chance!

I'm-a-gonna strap him to that machine and blindfold him like that chick in 50 Shades....
...show him moving pictures of being in the cockpit 'n stuff!
...gonna take away his hearing!

Hell,.... I'm even gonna rub this rig in kerosene and serve him bad coffey at 03:00 a.m till he don't know who his mama is!!!

Could not agree more. In my case I have a G-seat and I therefore need to feel the first forward kick from the horizontal actuator when starting to accelerate, then the prolonged effect of acceleration from the back of the G-seat. And, when I release the throttle, no backward kick but instead just the release of G-seat pressure in the back due to the removal of force.

Removal of force is not an additional force. Some other forces - braking, frictions... - can however jump in at this point in time and should be acknowledged but this would be reflected by acceleration figure going into negative territory, not just going back to 0.

No. What you feel in this case is not a horizontal backward kick equivalent to the one you feel on acceleration. It is 1/ a pressure release from the back of the seat and 2/ a strong tilt forward due to spring decompression.

Conversely if you actually brake you get an actual backward kick like the seat going back very strong and fast and you wanting to continue forward but being forced to follow the seat backward because of your harness.

Try it in a car with a harness, if you brake you will feel the strong pressure from the harness on your shoulders whereas if you just disengage the clutch (i.e. stop the engine force without generating braking from engine braking) you won't feel this same pressure from the harness.

What seat man ?
The accelerometer was firmly attached to the car dashboard.
What spring ?
Look at the gyro, I don’t see much pitch rotation. (Angular velocity).

Alex, if you have a chance to drive a race car, I recommend you do a phyphox experience with a smartphone attached under your seat. You analyze the data, Than you try to reproduce what you did in your simulator, and with another phyphox experience, this time under the seat of your simulator, you try to tune it as close to what you did in the reality.

I think doing so, you will see exactly where are your false cues and how eventually get rid of them.

Make it sense to you ? What do you think ?