Why adjusting control sensitivity is not the answer

[norman_99 209]

There has been a lot of talk here about the flight dynamics, and specifically reducing control sensitivity to make this more realistic. I’ve been thinking about this out this for a few hours and can’t agree with the concept for a number of reasons. I’ll try and keep the hyperbole to a minimum, as emotions are currently running higher than normal, but if you do think I’m exaggerating for dramatic effect, please understand that is not my intent.

I’ll preface these reasons by stating I think great strides have been made with the airflow modelling. Effects such as up/downdrafts from terrain, thermals over differ ground surfaces, smooth air over water and many more are improvements never before seen in software of this level. Unfortunately these improvements are severely hindered by the over sensitive nature of the actual aircraft dynamics.

Firstly, Almost all consumer control devices have a relatively similar range of motion, which by default produces a liner input response. Most aircraft also use near to liner control surface deflections (obviously FBW is different). The aircraft flight dynamics should therefore be designed to work with these input controls without drastic modifications. To think otherwise does not acknowledge the majority of users and their system setups. When a 50% roll input from a joystick or yoke is commanded, it should represent a similar input to the aircraft, and relevant control surfaces. When large reductions in sensitivity are set, such as -80%, a physical input of 50% now only commands a simulated control input of 10%. Logically this is simply putting a bandaid over the problem. The reverse is even more illustrative of the issue. If you are satisfied with an 80% reduction in sensitivity, without it you can effectively command a control 5 times stronger than designed. This should not be possible. Now before people start claiming we wait for 3rd party devs to produce high fidelity flight models, I’m not talking about recreating exact roll rate figures with specific control inputs etc. I simply want my 172 to fly and handle like the stable, docile, GA training aircraft it is, and not have the roll/pitch rates of an F16.

 

The second issue I have, is if we do change joystick sensitivity settings, where do we set it? It becomes at best an educated guess. As soon as we move away from a known baseline, we are effectively become our own flight dynamics designer. Any work put in by the developer (however good or bad) is now lost as control feel becomes arbitrarily set by the end user. As a community we would end up with control sensitivity’s spread out over a huge range, set to everyone’s individual taste. Everyone would therefore get a completely different handling experience. This seems completely counter to the philosophy of a simulator, which is designed to reproduce accurately the real world. Effectively, anyone can make a 152 respond like a 747 if they reduce sensitivity enough. Even with my real word experience, I couldn’t tell you -65% is better than -70%. Even if I could find a figure I’m satisfied with in the 172, I’ve never flown a Citation, so how do I know it’s appropriate for that aircraft too?

This leads me to another point. What happens when A2A and PMDG start releasing aircraft in 6 months time. I can already see the complaints of “why does my 182 handle like a 747”. And “why does my 747 handle like the Titanic!” Well, with control inputs reduction by up to 80%, what would you expect? Developers have no idea where each individual has there control sensitivity set. They too would have to start from a known baseline, namely 100% sensitivity and flat, linear curves. They then do there best to build an accurate simulation from there.
 

Obviously there will be those users that will tinker with these setting, and some may be justified due to hardware limitations ( larger null zones for noisy pots, etc), but these should be relatively minor. Huge reductions like the -70% and -80% recommendations I keep seeing should not be required to make aircraft flyable, and are indicators that something is drastically wrong with the current flight dynamics.

 

I do feel this sim has some huge potential, and as stated at the start, there are some really impressive airflow simulations already included. But currently the interaction between the users physical controls and the aircraft handling is surprisingly poor. Fingers crossed this gets fixed soon.

[rtodepart 761]

I’ll give it some more time and I’m sure some of if not all this issues will ironed out.. feels just any other new release to me . 

Edited August 19, 2020 by rtodepart

[MattNischan 3,778]

4 minutes ago, norman_99 said:

I’m not talking about recreating exact roll rate figures with specific control inputs etc

In many cases, though, that's exactly what was done here. If you look at some perf charts (for aircraft in which one can actually find them), roll/pitch rates at x deflection are actually pretty close.

The problem is, that's not really quite how airplanes feel, even if it is accurate. So then you have to make compromises and that becomes more of a subjective matter. Third party FDE devs have fought with this same push and pull for years. It isn't an exact science.

I think one thing is that the aircraft do have a tendency to return to their previous attitude too quickly. I'm working through the SDK now to see if that's model pervasive or more of a per aircraft configuration thing.

[sd_flyer 6,047]

I didn't tweak my controls except rudder. I have learnt my hard lesson during alpha to assign control to "axis" instead of others. 172 and few smaller GA doesn't feel like too much far off, yer they are far from perfect. I think rudder authority somewhat deficient in most GA. My attempts to put 152 aerobat into spin resulted a shallow slow roll which is completely unrealistic. Reverse area of command seems poorly modeled. Is there a hope to fix that all? I sincerely hope so !

[robert young 2,897]

6 minutes ago, norman_99 said:

There has been a lot of talk here about the flight dynamics, and specifically reducing control sensitivity to make this more realistic. I’ve been thinking about this out this for a few hours and can’t agree with the concept for a number of reasons. I’ll try and keep the hyperbole to a minimum, as emotions are currently running higher than normal, but if you do think I’m exaggerating for dramatic effect, please understand that is not my intent.

At last! Someone who finally agrees. I've been harping on about this in many posts and way before release, and the average answer is "user error" or "your controls need adjusting". As you have said, this is utter nonsense. A full control deflection is a full deflection. "Sensitivity" settings only distribute the total movement diffferently. They do not reduce or increase the control surface deflection potential. Addressing the flight model twitchiness by lowering your control settings is just masking the underlying issue which is that the aircraft has been designed with far too much control authority.

Setting your stick for low initial response, or moving your stick barely a millimetre to compensate for the enormous twitchiness is not the answer and thinking it is so is a delusion. But still people keep posting that it is all because of your response curves. It's like hitting your head against a brick wall.

Thanks for posting. I hope others might see the light eventually.

[fogboundturtle 2,708]

I disagree. Every controllers have different control sensitivity characteristic. It would only make sense if everyone was using the same controller.  I am sorry Robert and OP but you are wrong.

[hangar 661]

I don't think the issue is that others don't understand the difference as you say. I think the issue is that most people don't care as long as they get their own desired effect without waiting for fixes that will not come 🙂 That after all would be classic behavior in this genre.

As far as how this will effect 3rd party planes...it wont as long as you are able to save different sensitivity profiles for different planes. I asked about this about a week ago and someone said that yes you can have different profiles for different planes. But I still have not purchased so I really don't know.

Edited August 19, 2020 by hangar

[EmaRacing 302]

I'd like also to be able to edit the curve, by pushing the reactivity towards the end of the joystick movement.
As it is now, even if you set sensitivity to -90% you have violent reactions past 30% of the total movement, not good.

[robert young 2,897]

4 minutes ago, MattNischan said:

In many cases, though, that's exactly what was done here. If you look at some perf charts (for aircraft in which one can actually find them), roll/pitch rates at x deflection are actually pretty close.

The problem is, that's not really quite how airplanes feel, even if it is accurate. So then you have to make compromises and that becomes more of a subjective matter. Third party FDE devs have fought with this same push and pull for years. It isn't an exact science.

I think one thing is that the aircraft do have a tendency to return to their previous attitude too quickly. I'm working through the SDK now to see if that's model pervasive or more of a per aircraft configuration thing.

If you look through the flight model configs in the simobjects folder, you'll see clearly three things: 1) The inertias of almost all the aircraft are set at ludicrously low values. 2) The pitch stability params are set far too low 3) The control surface authority params are set way too high. In another post I have observed that the Bonanza maximum roll rate in this sim is equivalent to a high performance aircraft. It is way too strong.

 

[norman_99 209]

23 minutes ago, fogboundturtle said:

I disagree. Every controllers have different control sensitivity characteristic. It would only make sense if everyone was using the same controller.  I am sorry Robert and OP but you are wrong.

Not true. Most controls come out of the box with a linear curve, and 100% sensitivity. Physical control force/resistance is different, and will make each control feel different, but out of the box, “halfway left” is halfway left regardless of what your using. Asking the sim to interpret it as only 10% left is wrong.

Edited August 19, 2020 by norman_99

[FlyBaby 659]

3 minutes ago, fogboundturtle said:

I disagree. Every controllers have different control sensitivity characteristic. It would only make sense if everyone was using the same controller.  I am sorry Robert and OP but you are wrong.

Exactly...

There is a reason why DCS and XP have very detail sensitivity adjustment options....to be set for each plane / heli in the sim.

MSFS does need more adjustment options, but I can work with what they have thus far.

Controller sensitivity matters folks....if you think that you can plug in a random controller and have it work for the random plane that you loaded in sim...without any adjustments / fine tuning....you are wrong and will miss out on a great flight experience.

 

[FlyBaby 659]

9 minutes ago, EmaRacing said:

I'd like also to be able to edit the curve, by pushing the reactivity towards the end of the joystick movement.
As it is now, even if you set sensitivity to -90% you have violent reactions past 30% of the total movement, not good.

True....MSFS needs the same custom options as XP and DCS, where you can clip the Y axis / response if need be.

[EmaRacing 302]

1 minute ago, FlyBaby said:

True....MSFS needs the same custom options as XP and DCS, where you can clip the Y axis / response if need be.

Yes, I could do with half the control surfaces travel, I don't do aerobatics and the small planes don't really simulate the loss of authority at close to stall speeds.

[norman_99 209]

So if I have to reduce sensitivity by as much as 80%, just so I can use a joystick range of more than a millimetre or two, that’s ok?

As a test, I literally flew a complete circuit in the C280 entirely within the rage of play from the spring on my joystick. About 2mm side to side. If I went full left or right, I could roll the 208 quicker than a Pitts! That is not physically possible in the real world, and it shouldn’t be in the sim either.

As Robert said, the issue is excessive control authority, not my joystick. 

Edited August 19, 2020 by norman_99

[hangar 661]

Yes...it's true that the flight dynamics of some (if not all) of the default aircraft are flawed in certain ways...there is NO doubt of this. It's not JUST the sensitivities that are wrong. There are other flight dynamic issues with these planes as well, like thrust/drag and even some engine issues.

But...the only way forward is for 3rd party intervention, period.

Don't expect Asobo to change the flight modeling of their planes as, according to them, it's not likely to happen.

So...I assume that most (if not all) of these things can be rectified by either modifying the flight files or creating new aircraft. This has been the way of things since the beginning. Is it disappionting that we don't get realistic flight models out of the box? You Betchya! But at least, thanks to 3rd party devs, and other FDE tinker'ers there's a pretty clear path forward. 1 year from now all of this will be a non-issue...right now it's all we have and thats a let down I can't deny....but at least now we have a future within this franchise which is something we didn't have before.