Tanks - tracked vehicles driving and handling

[Beagle 684]

49 minutes ago, darkChozo said:

The Rhino is actually the fastest of the DLC vehicles. It just has crap acceleration, which I wouldn't mind seeing increased a little.

Its the fastest only in theory, but it is also difficult to control since it accelerates without input by its own. Sincew it wont slow down when you relese the forward input you have to "brake" which slows it down too much. So it's a constant stop and go instead of smooth driving. Very bad for a vehicle that needs to use agility instead of armor to be effective.

 

Oukeij mentioned thre Rhino is beeing worked on in the regard.

 

But again I cant get over the impression that besides actual control and input handling by the game, another problem is that acceleration seems to be linear instead of reversed exponential. A lot of the issues can't be seen on "paper" sidne the values of the vehicles seem o.k., but only on flat ground in the virtual garage.

 

All vehicles in general need  a better acceleration in the lower speed segment >50km/h while at speeds close to top speed the torquey assets are too good currently.

The issue with slowing down should not be underestimated...currently it is like "negative thrust" which makes a controlled speed reduction difficult and results in reverse instead of a stop. This control mechaqnics make brakign and steering at the sme tiem tricky. But braking, accelarating and slowing down and steering at the same time is what makes a vehicle stay on path. The result in arma III is much to often twitchy behaviour, especially with keyboard, and not enough repsonse with mouse.

 

unfortunately an analogue steering option and a brake function is not given for tracked vehicles. It makes drivint he tanks very uncomfortable and unprecise over long time.

 

Currently the only solution is to use different player profiles for different uses with HOTAS like plane, helo, tracked, wheeled, boat.

This is not a perfect solution at all and really bad for MP purposes.

[x3kj 1247]

1 hour ago, Beagle said:

But again I cant get over the impression that besides actual control and input handling by the game, another problem is that acceleration seems to be linear instead of reversed exponential. A lot of the issues can't be seen on "paper" sidne the values of the vehicles seem o.k., but only on flat ground in the virtual garage.

1 hour ago, Beagle said:

I think a big problem with mobility (not top speed) is about the way how shifting  is simulated. The logic is shy of shifting a gear down resulöting in the feel of "starving in high gear".

Another problem exclusively to tracked vehicle simulation is the absence of a real functional brake, instead you have onyl reverse which makes the already unprecise steering (no analogue sterring for tracked) even more twitchy.

No there is no shfiting simulation problem. The engine is using optimal points in torque-rpm curve of the engine for shifting.

However there is no torque converter simulation in Arma, which is a big issue for heavy vehicles like tanks. They now have fake "start assist" forces to help them start moving, but for stampin on accelerator when already travelling 20kph there is nothing.

 

Energy conservation is not respected either, so yes this can absolutely be a problem. If you change Moment of Inertia value of the wheels in the configuration a bit (factor of 1.5), the vehicle can turn from "barely accelerates" to "it accelerates without pressing gas pedal and starts flying" and that is no exaggeration.

 

Another issue is the modelling of the clutch. The clutch is assumed to two disks that are pressed against another but are permamently slipping (i'd like to see an Nvidia Physx programmer try simulating a permanently slipping clutch with a real truck *gg* magical conversion to permanently stuck clutch). And since the gear change is discrete (clutch is always closed, gear simply changes from X to Y) the clutch needs to be slipping otherwise it would be extremely jerky looking and weird on gear change. That does mean however, that with increasing "resistance", less and less of the power delivered by the engine is lost (because clutch slip increases with load).

It's going full circle to 4 years ago when i first posted this: https://feedback.bistudio.com/T77541
 

Quote

The result in arma III is much to often twitchy behaviour, especially with keyboard, and not enough repsonse with mouse.

unfortunately an analogue steering option and a brake function is not given for tracked vehicles. It makes drivint he tanks very uncomfortable and unprecise over long time.

They have improved handling significantly and made it less twitchy with the update... It makes a different when steering with and without thrust now - have you even noticed?. Keyboard, Mouse, My XBox Controller and also Joystick all work perfectly fine for me... 

 

 

[Beagle 684]

4 hours ago, x3kj said:

No there is no shfiting simulation problem. The engine is using optimal points in torque-rpm curve of the engine for shifting.

However there is no torque converter simulation in Arma, which is a big issue for heavy vehicles like tanks. They now have fake "start assist" forces to help them start moving, but for stampin on accelerator when already travelling 20kph there is nothing.

 

Energy conservation is not respected either, so yes this can absolutely be a problem. If you change Moment of Inertia value of the wheels in the configuration a bit (factor of 1.5), the vehicle can turn from "barely accelerates" to "it accelerates without pressing gas pedal and starts flying" and that is no exaggeration.

 

Another issue is the modelling of the clutch. The clutch is assumed to two disks that are pressed against another but are permamently slipping (i'd like to see an Nvidia Physx programmer try simulating a permanently slipping clutch with a real truck *gg* magical conversion to permanently stuck clutch). And since the gear change is discrete (clutch is always closed, gear simply changes from X to Y) the clutch needs to be slipping otherwise it would be extremely jerky looking and weird on gear change. That does mean however, that with increasing "resistance", less and less of the power delivered by the engine is lost (because clutch slip increases with load).

It's going full circle to 4 years ago when i first posted this: https://feedback.bistudio.com/T77541
 

They have improved handling significantly and made it less twitchy with the update... It makes a different when steering with and without thrust now - have you even noticed?. Keyboard, Mouse, My XBox Controller and also Joystick all work perfectly fine for me... 

 

 

Im not posting in this section to call out slightly reduced problem as perfectly fine. ArmA III problems are often control option issues, not engine or physics issues. The control options have not changed at all, there is still no control option section for  tracked class (or Boats). A lot of issues could be solved that way.

 

CURRENTLY:Car Analog left/right and Car left/right and Car more left/right, all in the same control tab for "vehicles" are simply conflicting each other.

 

Analog left/right works best for cars and wheeled IFV, Car more left/right works best for tracked vehicles and boats and gives a good feel and result but conflicts with analogue left/right for cars when using a Joystick or Wheel.

Left right works for both but gives not the full range of control on a Joystick etc.

 

"Car back" when used as a brake even on a axis works to abruptly now, like digital input: on or off.

[FirstPanic 39]

Hi Developers,

 

Short question: As we all are aware about the very big issues with the AI driving... will this be changed with this DLC?
To be honest: Today it's nearly impossible to have AI driving (doesn't matter if tanks or wheeled vehicles) in a mission. Please have a look on the examples in the "AI driving feedback" thread.

 

Would be great to have an improvement within this DLC and to get back on the AI driving level which we had 2 years before, before the AI driving model has been changed.

 

Thanks in advance for feedback.

FirstPanic

[scotg 204]

I just want to put in that this all about the technical aspect of the tanks' performances; it's not so much about how players or AI utilize the vehicles' abilities. That's the way I take it, anyway. That being said, how tanks are driven by whoever controls them is an important topic (whether here or elsewhere), but it's a secondary consideration of this thread. If controlling them in some way affects the manner they negotiate their environment, then it's relevant. Stay on topic to make sure you are staying afloat. This is the cusp at which the ocean floor starts to suddenly drop off from the shore, so watch your footing and bring your floaties and shark repellent.

[Yopoman 16]

I've been experimenting with the Tanks DLC on Dev-Branch for a couple of weeks now. I'd like to kindly add my opinion on how ground vehicles, especially armored vehicles and tanks, are controlled in relation to third person camera positioning. Right now, if I am controlling an armored vehicle in third-person, the camera is constantly changing its zoom automatically depending on nearby objects around the vehicle. This automatic zoom-in/zoom-out of the 3PP camera is highly disorienting and results in collisions with buildings and trees as well as my tank being shot by enemy vehicles in scenarios and editor missions while testing, events that would have been fully preventable if I had a steady camera view and camera angle being maintained while driving my vehicle near objects around the vehicle (houses/trees/hils etc...).

 

Also, the 3PP camera angle is far too low in the Z-axis (height relative to ground), and a bit too far to the rear of the vehicle in the X/Y-axes.

 

It would be much easier to command, drive, and be a gunner in armored vehicles/tanks, if the 3PP camera was higher in the air, angled more towards the ground in-front of the vehicle, and a bit closer to the turret of the vehicle being used.

 

The best third-person camera (3PP) system in a realistic game with armored vehicles being controlled by the player can be seen in the game 'War Thunder'. I highly encourage the developers of this Tanks DLC to install War Thunder on PC and try operating an armored ground vehicle from the third-person camera in that game, it is far superior to Arma 3's 3PP camera perspective and doesn't automatically zoom-in or zoom-out due to nearby objects around the vehicle. Also, when you zoom in to the maximum level in third-person mode in War Thunder, the camera position is just above the turret of the tank, which is great for driving/fighting in dense urban areas or forests. I would be thrilled, and greatly relieved, if an improved 3PP camera perspective and viewing options was implemented in Arma 3 Tanks DLC as a free platform update. For people who prefer the Arma 3 vehicle 3PP camera as it is now, perhaps the updated "War Thunder-like" 3PP camera can be toggled on/off as an option in Arma 3's Options --> Game settings menu. Thanks for your time, and keep up the great work, BI devs! :) <3

[scavenjer 112]

BTW, found a video of the Nyx Wiesel braking 

Apparently this wasn't even at full speed...

So yeah, it should be able to brake very agressively.

 

[twistking 204]

in joyful anticipation for the upcoming release i played around with the (old) tanks on stable branch and again noticed how big of a negative effect the "always-on turret stabilization" has on the vehicle handling even though not related to the vehicles physics itself.

 

for testing i assigned both movment and turret (vehicle: aim) controls to the same keys, so when "driving" (commanding) from gunner or comamnder seat, the respective turret would turn in sync with the hull: that immediately felt better, because you did not have to manually slave the turret to your driving direction.

this crude workaround would not work with a human crew obviously.

 

so again, i ask for a simple fix for this issue that impacts realism, immersion and handling.

a solution would be a mechanic to lock the active turret facing forward, or alternatively deactivating stabilization for the active turret.

this could be done in a way that the lock is only on while holding a button, or is toggled by a button, yet would reset when either moving the turret or left clicking (this way there would be no chance of beginner players getting stuck in this "lock". compare that to the infantry "lower weapon", "raise weapon" functionality)

 

again: the main reasons for having "turret lock" or disabling of stabilization:

- realistic looking tank manouvering in non-combat situation

- no need for the gunner/commander to constantly pan to keep the gun at 12 o clock when cruising

- locked main turret means better awareness of heading/vehicle movement for the commander when navigating when turned in and when turned out

 

[Tankbuster 1746]

It used to be that if the driver turned out, the turret locked. That was a nice way of doing it.

[twistking 204]

16 hours ago, Tankbuster said:

It used to be that if the driver turned out, the turret locked. That was a nice way of doing it.

 

well, i feel that with the improved tank mechanics and new interiors, this implementation would still be lacking.

 

first of all, there might be situation, where you want the turret in a "locked" cruise-mode, while the driver is inside the tank. especially with the new interiors, where a driver might actually prefer driving from under armor.

secondly it is a bit "uncommon" at least, to give the driver the possibility to lock the gunners turret (if you had to add some restriction because of safety (barrel clipping) issues, it would be more sensible to allow driver to turn out, only when turret is locked by gunner).

 

also if the lock mechanic for the gunner was introduced in a way that the first fire command (mouse click) or turret move command (move mouse) would cancel it, it would be an highly "optional" feature, that would add great amounts of immersion, authenticity and ease-of-use for experienced players, while it would not be in any way a hurdle for beginner players, since they could not get stuck in a mechanic they have yet to learn.

 

[martin_lee 33]

On 3/30/2018 at 12:02 AM, twistking said:

for testing i assigned both movment and turret (vehicle: aim) controls to the same keys, so when "driving" (commanding) from gunner or comamnder seat, the respective turret would turn in sync with the hull: that immediately felt better, because you did not have to manually slave the turret to your driving direction.

this crude workaround would not work with a human crew obviously.

The trouble I am having now is actually the hull trying to align with the turret. As long as you are not turning it too fast or using turn left/right, the hull would try to align with the hull. This current mechanism works okay if you are just cruising around as commander, since it is just like driving with mouse. You move the view, the vehicle turns that way. 

But when you are in combat, if you want to scan to your left? You have to press turn right so that the vehicle doesn't turn with you. In such case, the old system is actually better, since the tank just continues advancing whilst you scan and engage targets.

From these I think it is clear that neither the old and the new system were ideal for all situations. A toggle to enable/disable turret stabilisation is perhaps the way to go - it can (kind of?) allow us to toggle between the above two behaviors.

 

if a toggle were not possible, at least what I'd like/prefer is the old turret behaviour (always stablised and AI does not try to align the hull), and the current AI response (instant response to WASD, basically like driving yourself). Together with the PIP driver view, I think that could make a placeholder/compromise between the two - you could see where you are going, plus the ease of scanning during combat.

 

Oh, and one extra annoyance about aligning the hull with turret (and not the other way around). Whilst I am using the Nyx in the campaign, I found that if I turn the turret, the hull turns as well (as expected). But since that turret is not stabilised, what you will get is the combination of the turret rotation plus the turning of the Nyx, causing a very sensitive and imprecies transverse.

 

On 3/30/2018 at 12:02 AM, twistking said:

this could be done in a way that the lock is only on while holding a button, or is toggled by a button, yet would reset when either moving the turret or left clicking (this way there would be no chance of beginner players getting stuck in this "lock". compare that to the infantry "lower weapon", "raise weapon" functionality)

 

I am wondering if "lock ground" could be modified to used here, although tank gun stabilisation is about freezing its orientation, rather than tracking.

[twistking 204]

5 hours ago, martin_lee said:

The trouble I am having now is actually the hull trying to align with the turret. [...]

 

oh, wow. is that a new mechanic that came with the latest dev branch updates? wasn't in the changelogs...

i hope that they are only experimenting with it right now, because this implementation seems even worse than before (did not test myself though, bc i am on stable br.).

 

i think some kind of stab. toggle, or lock toggle would be the only real solution, while having move/hull direction independent from commander turret.

even more important for cooperative play, where a human gunner would never perfectly align the gun with the hull, resulting in some disorientating semi lateral movement of the commander optics etc. that is totally unnecessary and unrealistic in non-combat manoeuvring.

[lex__1 422]

6 hours ago, martin_lee said:

The trouble I am having now is actually the hull trying to align with the turret. As long as you are not turning it too fast or using turn left/right, the hull would try to align with the hull.

I did not like driving, turning the case from the mouse, since OFP. For a long time I use keys to control the rotation, and I do not have these problems. Change and try to drive the vehicle from the buttons.

[Imperator[TFD] 444]

1 hour ago, twistking said:

 

oh, wow. is that a new mechanic that came with the latest dev branch updates? wasn't in the changelogs...

i hope that they are only experimenting with it right now, because this implementation seems even worse than before (did not test myself though, bc i am on stable br.).

 

i think some kind of stab. toggle, or lock toggle would be the only real solution, while having move/hull direction independent from commander turret.

even more important for cooperative play, where a human gunner would never perfectly align the gun with the hull, resulting in some disorientating semi lateral movement of the commander optics etc. that is totally unnecessary and unrealistic in non-combat manoeuvring.

 

This is a player option in the menu which from my rusty memory is called Vehicle Freelook and it allows the user to steer the vehicle with the mouse when turned off.  By doing this you lose turret control if you're driving the vehicle from the commander/gunner slot.

 

My recommendation is that you do not turn off Vehicle Freelook.

[martin_lee 33]

hm... No one experienced this? May be it is my control settings then...??

 

Just to clarify, I am in the commander/gunner seat with an AI driver. If I turn the turret the AI turns the hull as well.

 

EDIT: Okay, I think I know the problem now. If you have mouse left/right mapped to car left/right and an AI driver, you are controlling BOTH the turret and the hull.

 

Now I am kind of embarrassed ... let me hide inside a tank for a while

 

Edited April 7, 2018 by martin_lee
oops

[oukej 2910]

Another piece of docu' - https://community.bistudio.com/wiki/Arma_3_Vehicle_Handling_Configuration - thanks to @Lou Montana! Big thanks! Document also includes properties introduced during the Tanks development and updated, corrected of clarified definitions or example values of few others.

To reduce the duplication of information we may replace the info elsewhere with references to this document - feel free to do so as well anytime you find yourself casually editing the wiki ;)

[R3vo 2654]

9 hours ago, oukej said:

Another piece of docu' - https://community.bistudio.com/wiki/Arma_3_Vehicle_Handling_Configuration - thanks to @Lou Montana! Big thanks! Document also includes properties introduced during the Tanks development and updated, corrected of clarified definitions or example values of few others.

To reduce the duplication of information we may replace the info elsewhere with references to this document - feel free to do so as well anytime you find yourself casually editing the wiki ;)

Thats a very nice documentations.

[x3kj 1247]

Could you clarify why the enginePower parameter is important?

After all, it is deducable by max/min omega, max torque and the torque-curve.   P = M * omega

 

dampingRateInAir is missing (unless this was made obsolete?). I would appreciate an indepth description how and what that value does exactly. It is possible to break physx with wrong values (MOI and dampingrateInAir not beeing balanced out), where the tank will accelerate on it's own at 0% throttle - so it would be good to know what exactly it does.

 

From my experience clutch strength for tanks, a starting value of 10 is whoefully inadequate. 30-40 is better.  On the other hand, if clutch strength values are too high it will become noticeable because the vehicle will jerk whenever it changes gears.

Switchtimes near and above 1 seconds (for tanks) produced handling problems in the past, because during switching you can't steer. Idk if that changed when changing the steering system, but i would guess not. The faster the tank the more problematic the issue. Also problematic when the gearbox can't decide if its liking one gear or the other more for a certain terrain grade.

 

driveOnComponents and the other driveOn thing (i forgot) is also missing  - it would be good to have clarification for this as well, how it behaves currently / what those things do.

 

Btw:

Real contemporary tanks up until recently didn't have dampers on all roadwheels in the suspension. Mostly it was just the first and last, or the first two and last, or just the first two wheels that have dampers. This is what results in "springy" look when they drive up a step. Idk how the current situation is, but i strongly suspect since the majority still uses old chassis that this was not changed for most.

 

 

 

[oukej 2910]

On 5/2/2018 at 9:31 PM, x3kj said:

Could you clarify why the enginePower parameter is important?

Used to compute default for peakTorque (if undefined). Otherwise only for deprecated functionalities.Afaict. Default is 50.
 

On 5/2/2018 at 9:31 PM, x3kj said:

dampingRateInAir is missing (unless this was made obsolete?). I would appreciate an indepth description how and what that value does exactly. It is possible to break physx with wrong values (MOI and dampingrateInAir not beeing balanced out), where the tank will accelerate on it's own at 0% throttle - so it would be good to know what exactly it does

It was added in the early development because the tank physics (their wheel differential) didn't work well with wheels that are constantly in the air (sprocket, pulley). It  dampens the wheels that are in the air and as you hinted towards - it is a bit of a "magic" value.
 

On 5/2/2018 at 9:31 PM, x3kj said:

From my experience clutch strength for tanks, a starting value of 10 is whoefully inadequate. 30-40 is better.  On the other hand, if clutch strength values are too high it will become noticeable because the vehicle will jerk whenever it changes gears.

Switchtimes near and above 1 seconds (for tanks) produced handling problems in the past, because during switching you can't steer. Idk if that changed when changing the steering system, but i would guess not. The faster the tank the more problematic the issue. Also problematic when the gearbox can't decide if its liking one gear or the other more for a certain terrain grade.

For tanks' clutch strength we use more or less the values you mentioned (30-50, even up to 70). Also we have the switchTime set to 0 on all tracked vehicles. Again - as u mentioned - because you can't steer during the change. We didn't have the chance to alter that.
 

On 5/2/2018 at 9:31 PM, x3kj said:

Real contemporary tanks up until recently didn't have dampers on all roadwheels in the suspension. Mostly it was just the first and last, or the first two and last, or just the first two wheels that have dampers. This is what results in "springy" look when they drive up a step. Idk how the current situation is, but i strongly suspect since the majority still uses old chassis that this was not changed for most.

If you have tried the first iteration of an overhauled Kuma last year there was basically no damping (minimal stable damping). It's suspension looked better in many cases. But there was an issues with rough impacts - e.g. when jumping. The tanks bounced away from the ground in a toy-like fashion. It didn't feel like a 60-tonne vehicle. IRL the tank's weight wouldn't be transferred so well into the recoil. The vehicle would also be able to 'smooth the terrain'. On vanilla tanks we tried to compensate these missing effects by adding dampers (even though that came at a loss of this nice suspension sway, especially after firing a gun). It makes the tanks drive better over a rough terrain.

[x3kj 1247]

On 9.5.2018 at 5:58 PM, oukej said:

It was added in the early development because the tank physics (their wheel differential) didn't work well with wheels that are constantly in the air (sprocket, pulley). It  dampens the wheels that are in the air and as you hinted towards - it is a bit of a "magic" value.

But what exactly does it do, when compared to it's non-magic counterpart "damping"? Why do all wheels have dampingInAir instead of just those that are permamently in the air (in the current configurations)? Is the role of regular damping parameter still that of physx default (mDampinRate i assume)? How does dampingrateinair interface with this system - does it modify the dampingrate of a wheel under certain conditions, or does it apply always? And how does it apply it's value?  Why is it possible that a system intended for retarding (damping) wheel speed results in forces that propell it forwards? All calculations and inputs for mass, moi etc are utterly pointless when there is one pink elefant value that just does <whatever> and nobody knows what exactly.

On 9.5.2018 at 5:58 PM, oukej said:

If you have tried the first iteration of an overhauled Kuma last year there was basically no damping (minimal stable damping).

IRL the vehicles are damped regularly, but all the damping forces come from 1-2 wheels per side. The rest do not contribute anything to suspension damping. So in a sense, 1-2wheels are overdamped (for the spring it uses) but normally damped when all spring forces are counted together. That is not the same as no damping ;)

If you look very closely at the suspension movements of this jumping t-90 you can see how the first two wheels and the last one do not wobble when it got into the air -> those ought to have suspension. The other three wheels wobble a great deal in the air. This results in a behaviour difference when the vehicle rocks back and forth, as the first and last wheels can lift of the ground, thus cant contribute to damping momentarily.

[twistking 204]

On 29.3.2018 at 6:02 PM, twistking said:

in joyful anticipation for the upcoming release i played around with the (old) tanks on stable branch and again noticed how big of a negative effect the "always-on turret stabilization" has on the vehicle handling even though not related to the vehicles physics itself.

[...]

so again, i ask for a simple fix for this issue that impacts realism, immersion and handling.

a solution would be a mechanic to lock the active turret facing forward, or alternatively deactivating stabilization for the active turret.

this could be done in a way that the lock is only on while holding a button, or is toggled by a button, yet would reset when either moving the turret or left clicking (this way there would be no chance of beginner players getting stuck in this "lock". compare that to the infantry "lower weapon", "raise weapon" functionality)

 

again: the main reasons for having "turret lock" or disabling of stabilization:

- realistic looking tank manouvering in non-combat situation

- no need for the gunner/commander to constantly pan to keep the gun at 12 o clock when cruising

- locked main turret means better awareness of heading/vehicle movement for the commander when navigating when turned in and when turned out

 

 

any news, if we could see sth. like a turret lock/stabilize toggle with the upcoming patch?

 

i did some wwII mod milsim tanking recently and just cruising around was so much better without the turret constantly moving relatively to the hull, which is unrealistic and confusing for the commander...

[haleks 8212]

From today's SPOTREP :

Quote

Tweaked: Interaction with inventories of dead characters was improved

 Anyone has more info on this? I've been using the RC branch for a few weeks, but I didn't notice anything.

[SuicideKing 233]

1 hour ago, haleks said:

From today's SPOTREP :

 Anyone has more info on this? I've been using the RC branch for a few weeks, but I didn't notice anything.

Probably better suited to the dev branch thread?

[scotg 204]

I'm always speaking from a mod perspective, so...
Was the neutral steer rate ever updated? Specifically, is there a setting to determine how fast a vehicle can turn in place, or disable it entirely? I'm thinking this might be the first step to making better half-tracks.

[oukej 2910]

On 9/10/2018 at 4:02 AM, scotg said:

I'm always speaking from a mod perspective, so...
Was the neutral steer rate ever updated? Specifically, is there a setting to determine how fast a vehicle can turn in place, or disable it entirely? I'm thinking this might be the first step to making better half-tracks.

tankTurnForce and the following related properties