Jump to content

Controls Demystified(?)


48 replies to this topic

#21 Koniving

    Welcoming Committee

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Guide
  • The Guide
  • 23,384 posts

Posted 11 November 2013 - 10:48 AM

Note.

I discovered that the cheap 20 dollar PS3 controllers in Walmart are zero-order controllers. I have amazing accuracy with the sticks. It seems more accurate with the left stick than the right, however.

#22 Loc Nar

    Member

  • PipPipPipPipPipPipPipPip
  • 1,132 posts

Posted 11 November 2013 - 01:54 PM

Quote

I discovered that the cheap 20 dollar PS3 controllers in Walmart are zero-order controllers


I think you may be missing some pieces of the puzzle here, as I'm pretty sure those are not zero-order controllers even if you have found a way to use them as such, unless you have discovered a previously unknown breed of PS3 controllers that are mechanically very different from a regular game controller. (link? more info?)...

Again, any joy returning absolute inputs is operating in zero-order (currently only possible via mouse emulation in MWO), where displacements are directly converted into position. However being able to operate in zero-order does not make it a zero-order controller any more than operating if first-order makes my stick or a mouse for that matter a first-order controller while operating in that mode. The distinction comes by which task they are primarily mechanically engineered around. For a joystick (ps3 controller is just 2 little joysticks, some buttons, and a split pressure axis) that would mean completely different gimbals. If they are isotonic (spring centering) like any other game controller's dualsticks, then it is not engineered for zero-order.

This does not mean you can't make it work well or you aren't having luck with it, merely a misappropriation of titles as a ps3 controller does not fall under the umbrella of zero-order controllers. That list has devices like mouse, absolute tablet, trackball, my joystick, etc, and the traits they share in common that define them as primary zero-order devices are not found in typical joysticks or game controllers.

BTW, if you like game controllers, Thrustmaster makes one with 16bit dualsticks ($30 on amazon) where most are just 8bit, and for zero-order use this actually does matter while it is a non-issue for first-order which by comparison is a clumsy affair of nudging the target the direction you want it to move rather than precisely moving it there directly like zero-order.

That new Steam controller with it's circular trackpads looks capable of native zero-order as well as native first-order, but also seems like it would do a poor job of both rather than a good job at either one, but I've yet to use one and only base this on extrapolation from what I can see and what I already understand.

#23 Loc Nar

    Member

  • PipPipPipPipPipPipPipPip
  • 1,132 posts

Posted 23 November 2013 - 10:31 PM

For as much time as I spend railing on about zero-order control I sure don't have a lot of visual aid. Here is a neat little short vid showing a miniature 6DoF hexapod Stewart Platform being controlled by a joystick handle mounted to another 6DoF hexapod platform.

This joystick is using absolute inputs and the displacements of its axes directly manipulate the positions of the axes of the other platform. In other words, this joystick is operating in zero-order control.

Although the example is 6 axes, my mechstick works the same way as this. Just like the stick in the video describes the same motions as the platform it is controlling, my stick moves in pitch/yaw (+/-15deg and +/- 45deg) same as the mechs it is controlling, and uses absolute inputs that directly manipulate the position of pitch/yaw of the mech.

edit: after a little more research after chasing a clue from this video's description, I have learned and verified the nomenclature of this type of relation between the input device and the object being controlled in zero-order/direct input is called Master/Slave, and the most common application of this control type is in robotics which was very vindicating to discover.

By definition, my joystick is a Master/Slave controller for a mech, and since I do not use 1:1 scaling of my inputs, it's further sub-categorized as a Replica controller.

slave/master/replica

Edited by Loc Nar, 24 November 2013 - 10:05 PM.


#24 Sorejack

    Rookie

  • 1 posts

Posted 08 December 2013 - 10:27 AM

Isn't an isometric controller a zero order input device? Force is applied and registered when force is released it returns to zero instantly. Seems the rest is emulation . Would be easy to turn an isometric controller like the warthog or x65 to register relative . Ugh never mind . It would only work like stick. Except with better precision when returning to a zero state. Still it is illuminating. Maybe it is not as much about the order of control as the skill required. Seems zero order devices are reflexively intuitive where others require a degree of orientation. So is using head tracking zero order? My ideal stick would now have a track ball next to the hat switch for my thumb.

#25 Loc Nar

    Member

  • PipPipPipPipPipPipPipPip
  • 1,132 posts

Posted 09 December 2013 - 02:18 PM

Sorejack, I recoginize your name from SC threads that I too am lurking pretty hard. Nice to see you here and there, probing into the same matters of interest. ;)

Quote

Isn't an isometric controller a zero order input device? Force is applied and registered when force is released it returns to zero instantly.


Not as I understand things. isotonic controllers can be zero-order but isometric is almost mechanically incapable of what zero-order is by definition. Zero-order directly converts displacement into position. Isometrics however are proportional controllers like a regular joy, but proportional to force rather than displacement and the arrangement is good for little other than velocity commands such as first/second-order owing to its extreme mechanical preference of its state of rest... -It really (really) wants to stay in the middle and the last thing you want a positional controller doing is telling YOU what position it wants to be in.

Optimized positional controllers are engineered around this notion, which for joysticks generally means no spring centering or center detentes. Isometric arrangement is basically the ultimate spring center/detente.

Foust brought up a good question before that led to realizing it possibly *could be used for zero-order (asked about using absolute inputs with it), where force was would actually be directly converted into position. It could be done, but this would be an imprecise and fatiguing way to try to control position, although it may technically satisfy the req criteria for nomenclature of zero-order, but either way would fall far short of being usable or something you want to do.

Quote

Seems the rest is emulation . Would be easy to turn an isometric controller like the warthog or x65 to register relative .


Warty not isometric unless mounted on aftermarket FSSB base, which is in a whole different league than x65 BTW. Real FS sticks in modern aircraft are actually hybrids and not pure FS as they allow a small amount of displacement for tactile feedback and that makes all the difference. X65 has none and is static, and suffers as a result, leading to its discontinuation by Saitek with its controversial reputation.

While they might be classified absolute devices at the hardware level (not sure about isometrics ATM), they are highly mechanically optimized around relative inputs in terms of signal so they already 'register' relative. Not quite sure what you mean beyond this, but if I missed the point, please have another go at it.

Quote

Maybe it is not as much about the order of control as the skill required..


The main thing really does boil down to control-order being gatekeeper, because so long as the name of the game is direct positioning, and that task can be handled 100% effectively with positioning controllers, velocity based controls of any kind will never be as good at performing the same task. Being better/more ergonomic/intuitive from one device to another only affects the small range between better-worse within the ballpark of like-controllers, but will generally not be the make/break decider outside of corner cases.

Quote

Seems zero order devices are reflexively intuitive where others require a degree of orientation. So is using head tracking zero order? My ideal stick would now have a track ball next to the hat switch for my thumb


Reflexive intuitiveness overall is a huge factor underlying why adhering to control order is so important. I need to find more eloquent ways of expressing this, because underneath it all this is what it means to a human operator. Not just requiring higher orientation; and the less reflexive/intuitive a task becomes, the higher the straight up finesse/judgement/compensation skill bar raises to achieve like levels of performance, but orientation is part too.

Head tracking is usually used zero-order, but not limited to it. The trackball for the thumb idea has merit, and would be preferable if infinite travel was needed on the axes it controls. If limited range could be used, another option would be a little trackpad sunning absolute inputs with a stylus trapped on a miniature x/y plotter mechanism such as Mr 144 was/is working on in SC. (another thread worth lurking). For SC I'm definitely on the head tracking tip, and here's my album of mods I've done with my TrackIR5 setup to address what I felt were it's only 2 shortcomings (camera position and passive marker quality)

Even within similar controller types intuition/reflexiveness is a factor too, such as the difference between master/slave/replica vs master/slave/nongeometric analogic . Replica master controllers move identically to their slaves so reflex and intuition is considered optimal. If similar but not identical movements are used like in nongeometric analogic master/slave control, this reduces reflexive intuitive control factor, and the bigger the disconnect the less ideal it becomes, which applies to most things operated by a human. There is some on this at the bottom of the page here:

#26 shotokan5

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 550 posts
  • Google+: Link
  • Locationvirginia

Posted 12 December 2013 - 08:03 AM

With all the great minds in post could someone please tell me how to get the target radical to work on a X52 Pro everything else works great. l have used user cfg. and their dead zones and everything works great except that. Please your help would be very much appreciated.

Oh. Roberts in his new game made the joysticks work from the get go, with the cry engine. Guess he is not very, smart. When from early beta they said that the joystick is low priority. Guess after over a year of programming Christmas lights and naked women has taken up to much of their time. Guess they made a mistake in putting a joystick in the mech and not a keyboard. Don't get me wrong I have great respect for those who can use keyboards even when few did.

#27 KulaGGin

    Member

  • Pip
  • 12 posts
  • Google+: Link
  • LocationUkraine

Posted 29 December 2013 - 06:35 AM

Loc nar this is very good topic. Deep... Unique... Useful... And hard to understand...
I was reading this topic for last 2 hours or so. I just finished reading this topic and I feel my brain is tired of understanding what are you trying to say.
If I understood correctly, you using controller which windows recognize as a mouse. And this controller responds to changes in amperage so you can use resistors/hall sensors instead of optocoupler/optical encoder. But where did you get motherboard like that?
Could you please post video how your joy work in windows and in MWO? I'd like to see how zero-order joystick works in an FPS game.

#28 Loc Nar

    Member

  • PipPipPipPipPipPipPipPip
  • 1,132 posts

Posted 29 December 2013 - 01:19 PM

Quote

Loc nar this is very good topic. Deep... Unique... Useful... And hard to understand...
I was reading this topic for last 2 hours or so. I just finished reading this topic and I feel my brain is tired of understanding what are you trying to say.


Thanks KulaGGin. The takeaway is that:

1) controls are defined by their axis behavior.

2) Axis behavior is defined using a system of classification called control-order.

3) If unclear, repeat steps 1 and 2

To understand controls, you must understand axis behavior, and using proper control-order naming convention avoids much confusion these discussions lead to. Luckily games only deal with zero-order (aka mouse aim), and first or second-order control inputs like vehicular movement, so no need to worry about what is happening beyond that. Within this though, it's really vital to understand at least first-order and zero-order control, becasue this is where most of people's lack of understanding gets them in trouble and prevents meaningful discussions on the sugject.


Different control-order tasks tend to breed mutually exclusive hardware/mechanical arrangements to augment their primary tasks, and such is the case between zero and first-order so mixing controllers predictably yields poor results in many circumstances, like if a task is zero-order positioning (such as reticule aim in MWO) and but you are using a regular joystick which is made for first-order inputs. So many misconceptions as to what is happening here and why this is. So many misconceptions about the implications of an application being in zero-order, and what can be done as workarounds.

Spoiler


The key to understanding controls though, is first understanding axis behavior because it's all built around that. Like anything technical in nature, you shouldn't make up your own terms or definitions if you want to be able to communicate these things so again it's back to control-order. I used an analogy of a marble on a sheet of glass as an example in my OP, but here's a further simplified example of the relationship of order of controls.

...imagine playing chess, or more precisely the act of moving pieces on a chessboard:

Zero-order control -using your hand to directly move the chess piece to it's new square.

First-order control -strap the piece to a tiny rc car and drive it to its new square using the r/c controller.


Now imagine you are racing an opponent to the new square. He's using zero-order and you have your piece strapped to the little car. As you can imagine, it's much more difficult to translate r/c controller inputs (proportional x/y displacements) into velocity/direction commands in just the right amounts to make the car arrive at it's new destination in a timely fashion, while the other guy was able to immediately move his piece very accurately. MWO basically *is such a race... and so long as anyone is able to simply use their hand to directly move the piece the new square, it doesn't matter if your little rc car is a Ferarri or not whether it's blessed by PGI or not, because even a guy with a broken hand will still win and you're still bringing a knife to a gunfight.

Games where joysticks are an actual option do not allow this type of direct positional control, and instead have what's called a 'virtual joystick', where the mouse is reduced to a vector controller that responds to displacement from a fixed 'home' and is not positional in nature. Zero-order positional is a mutually exclusive scheme, and so long as direct positional exists in MWO, indirect/vector commands will always take backseat. There are no magic tricks here to make this good, only work less bad and only up to a certain point.

Control orders do continue to climb in number, and complexity/response time with each progression, although again games only require attention focused on zero and first-order since those are so very common in games yet so different from each other. That said, some axes are technically second order inputs (like steering a mech for instance), but their behavior in game using the same controllers is nearly identical so the differences largely academic, while zero and first-order are night and day...

Quote

If I understood correctly, you using controller which windows recognize as a mouse. And this controller responds to changes in amperage so you can use resistors/hall sensors instead of optocoupler/optical encoder. But where did you get motherboard like that?


Nope. My joystick is recognized as a regular joystick by windows, and it's only MWO that thinks it's a mouse; and only because I happen to be using 3rd party software to emulate one. The only reason I use mouse emulation is because it is currently the only way to use absolute inputs in MWO, and absolute inputs is the key to using a joystick for positional control. Literally any joystick running absolute inputs in a zero-order application is a positioning device of varying quality, so whether it's any good at this task or not is entirely dependent on the mechanical bits it moves on in addition to the usual ergonomic considerations.

Any joystick can do this, and can do this right now with PGI's help or not and there is nothing special about mine's board beyond having useful features to me but this scheme can be accomplished by any stick and mouse emulation and has been this way since the beginning. It's easiest for TARGET based sticks (Thrustmaster's kickass software).

Using mouse emulation for absolute inputs is not 100% ideal, but despite misguided hopes around here even full-on support by PGI would not be much different. The key to success would still be predicated on having a gimbal that is mechanically suited for positioning instead of vectoring. Those don't grow on trees and can't be bought in stores (well, the Steel Batallion controller has one and it *can be used in MWO...)

Because of this, my joystick is as much a fish out of water in a flight sim as a flightstick is in MWO... Since my joy is digitally/electronically nothing special, what really separates if from the herd is mechanical in nature. http://imgur.com/a/ixi64 Dedicated, purpose-built positioning gimbal, that's what makes it special.

This has nothing to do with the sensors though, which are nice but not unique on my stick. They're sealed Hall pots that I made, to replace the regular pots but they do the same thing the same way -directly convert angular position into a 0-5vdc output, which is the same as is happening in any joystick/thottle/etc.

Initially I almost used a mouse's optical sensor instead of pots to pick up movement, which might have helped people that just 'don't get it' to understand what this stick *actually is and how it works. It doesn't matter if I'm using the mouses optical sensors, or my Halls, or regular pots, or magnetoresitors, etc to pickup/convert it's movements into inputs though, the result would be largely indistinguishable in-game and the real work remains tied up in mechanical aspects of precision kinematics.

What's hard about making this gimbal was to make *feel right, meaning the correct blend of damping/resistance vs being able to freely move it, and having no slop whatsoever on any moving parts. Too loose and my aim is too shakey and it would be hard to hold targets, too stiff and I can't affect fine inputs and will be fatiguing. Not stiff enough it won't hold position, wrong grease on the rubs it will ratchet when I move it, etc. With any controls project I've worked on, proper kinematics trump the rest of the project by such a large margin that the rest of the project winds up trivial by comparison.

Quote

Could you please post video how your joy work in windows and in MWO? I'd like to see how zero-order joystick works in an FPS game.


I don't have any videos yet, but the video I did post above demonstrates exactly what a master/slave/replica controller does. My stick/mech share the same relationship as the 2 Stewart platforms. The input platform (slave) is moved manually by the operator, and the robot/slave platform tracks and mimics it's position at all times.

My stick is the same though -the mech always tracks my stick's position. If my stick is all the way up/left, the mech is all the way up/left and if I take my hand off the stick they both stay like that. 50% left stick twist = 50% left torso turn etc. I can literally look at my joystick and be able to tell what position my mech is in without looking at the screen.

#29 VX Seraphin

    Member

  • PipPip
  • 36 posts
  • Locationcanada

Posted 30 July 2014 - 07:03 PM

thankyou for the very in depth explanations of how various input devices work
i am more enlightened and that makes my day a better day :)
what follows is a discussion on how real world mechanics SHOULD translate into the mechwarrior world,or how i would like to use them in the mechwarrior world
this is by no means a definitive absolute observation,just My Humble Opinion (i hate acronyms,they just seem lazy... lol)
i will be using an excavator as my real world representation of a mech
the whole purpose of the excavator is to postion an impliment where you want it and how fast you want it there
to do this joysticks and pedels are used
for the sake of less confusion i will only use joysticks for my examples,as the foot pedel function is the same
in my real world,joysticks have 2 functions
role #1
positional
in this case,the joystick positions an element at a disired hieght/angle/distance/depth
eg:
i want my bucket to stop at 3 feet above the ground
i want my bucket to start digging 20 feet away from me
i want my bucket to dig a hole 45 degrees off parrallel to the road
i want my bucket to dig down 10 feet
role#2
rate of function
in this case the joystick controls the rate of positioning the element
eg:
i want my bucket to be 3 feet above the ground when it is 20 feet away
i want my bucket to get to 20 feet away from me as quickly as possible
i want my bucket to be at 45 degrees off parrallel at the same time it is 20 feet away from me
i want my bucket curl and go down at the right speed to dig down that 10 feet
while any one of these actions can be done indipendently,the most efficient way is to do them collectivlly, at the same time
so lets apply this to Mechwarrior Online
the whole purpose of the mech is to position the aiming reticule where you want it and how fast you want it there
in the mechwarrior world joysticks have 2 functions

role #1
positional
in this case,the joystick positions the reticule at a disired hieght/angle/distance/depth
eg:
i want my reticule to go up 4 meters above the ground targeting the torso
i want my reticule to be at that position 200 meters away from me
i want my reticule 45 degrees off parrallel to my mech,s legs
i want my reticule to go down 2 meters to target the leg
roll#2
rate of function
in this case the joystick controls the rate of positioning the element
eg:
i want my reticule to be 4 meters above the ground when it is 200 meters away
i want my reticule to be at that position 200 meters away from me as quickly as possible
i want my reticule 45 degrees off parrallel to my mech,s legs at the same time it is 200 meters away
i want my reticule to go down 2 meters to target the leg at the right speed to maintain target
in both cases,real or virtual we are looking for a point in space
how we get there, how well we control the rate of speed getting there,and how well we control that point once we are there is all determined by the MACHINE involved,and the OPERATOR,S skill level
so i have capitalized to words,MACHINE and OPERATOR as i believe these two words are paramount to where i am headed with this discussion
first OPERATOR
in the real world,there are poor operators and skilled operators
most people, no matter what their walk in life can tell the difference,just by watching
jerkiness,quick,abrupt movements are an indication of an inexpierienced or relatively new operator
over time muscle memory,confidence,expeirience will all contribute to becoming a skilled operator
then there are the natural born operators
theses people can literally jump on any piece of equipment and after only a few hours can operate it well and after a few weeks can operate at a highly skilled level
then they take it that one step further when they do things bordering on the impossible,a level that is reached by very few and which is only noticed by the highly skilled, a level which all good operators hope to achieve but very few do

in Mechwarrior Online you find the same levels of skill
you see the ones learning, you see the ones with a well developed skill set,and then there are the natural born
the natural born though in many cases get malighned as hacking or cheating or doing something other than what they truly are
they are just that good at what they are doing,thats it,thats all
so this brings us to the MACHINE
staying with the excavator,there are hundreds of variations,from a tiny Komatsu PC01, aprox 300kg to all out behemoths such as the Cat 6120b at an astounding 1,400 tons!
all have the basic joystick/footpedal method of control
some have quicker track (leg) speed than others
some have quicker turning rates than others
some have faster acceleleration/deceleration than others
some swing the upper structure (torso twist) quicker than others
some accelerate/decelerate the upper structure(torso twist)quicker than others
some the boom cycle time (torso pitch) is quicker than others
in many ways,the excavator is like the mech we wish operate
in Mechwarrior online,we have light mechs and heavy mechs
we have the standard mouse/keyboard control,with the OPTION for joystick control....
this is, unfortunatly where the excavator/mech analogy falls apart
we have all the nice travel speeds/acceleration/deceleration rates, pich/yaw rates, turn speeds all the things that are needed to move the various parts of the mech in harmony with one another to get the targetig reticule where we want it,,
but only in theory
to illustrate what i mean,lets go to the mechlab and look at the stat sheet on, lets see..
the direwolf
big sonamagun,100 tons
first the acceleration/deceleration
mat the thing and it takes some 40^2 milli seconds toget to a top speed of a lousy 48 kilometers an hour
back off the throttle and it slows down to 0 after about 30^2 milli seconds
so you play with the mouse /keyboard and thats what happens push the button to go,let go of the button it slows down
same with a joystick/throttle combination
you want to maintain a certian speed ,hit the appropriate button,or hold the throttle at a certain distance and you will go that speed
all good
so now we want to turn the thing
turn rate is controled by travel speed
the faster you are moving,the slower your turn rate
so standig still, you will turn this thing at about 37 degrees per second down to some 25 degrees per second at full throttle
using the joystick you can get a fairly nice turn rate that is fairly representative of the numbers they have here,with it changing as the travel speed changes, OR you can turn at a SLOWER rate at a given speed by using LESS deflection/twist on your joystick
when using the buttons on the keyboard there is only one speed,and that is the speed of the turn at the given speed
in my opinion,this gives a jerky,unrealistic,and hard to use turning method,and really limits what the mech is actually capable of
if you like aiming with you legs and torso,you know what i mean
so now we get to the core of everything that i have put forth here
the yaw and pitch of the upper torso *as the arm movement can be locked or unlocked i will not include it,s effect*
the pich is 34 degrees per second
the yaw is 63 degrees per second
using a joystick,s x/y axis will allow you to move the torso at a MAXIMUM rate of 34/63 degrees per second
NO faster but with the ability to go slower if you want
with the mouse,you can twitch that torso as fast as you can,totaly making those numbers meaningless
and that is where the whole control scheme is broken, in my opinion
so how would i fix it?
first thing i would do , make the mechs work like the specs say they should
secondly i would quit trying to make a joystick work like a mouse
and thirdly have the 4 axis (x/y/z/slider) implimented to the simple 255 step axis protocal that microsoft recognizes.
in my most humble opinion these three things , would go a long way in bringing the right feel to the game,what ever your control of choice is
VX_Seraphin

Edited by VX Seraphin, 30 July 2014 - 07:14 PM.


#30 Odins Fist

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 3,111 posts
  • LocationThe North

Posted 31 July 2014 - 07:15 AM

Please do not take this post the wrong way..

I can always tell when i'm fighting against a pilot that is using a joystick for aiming..
Because he "DIES".

He will drift his fire around too much, and if it's applicable during the match, I will be able to find him due to CONSTANT weapons fire drift, then manuever engage and destroy.

Joysticks are great for immersion, and a good pilot can manage good level of skill especially if he is using longer range weapons where he can take his time to aim.
With that said, the flightstick/joystick has a severe disadvantage for aiming when put against the Mouse for aiming, this is cut and dry, no agrument can be made.

How do I know.?? I have (2) Saitek Flight sticks, I tried them, I liked the immersion factor, BUT it is simply impossible to get on target as fast or to achieve accuracy as FAST a mouse.

ALSO, I do not use the keyboard W.A.S.D., I use something similar to the Split Fish, I use the left side of a progammable Gamepad (Xbox style).. The left thumbwheel, left shoulder trigger(s), and the D-pad for functions, and have been doing this in First Person Shooters since COD-4.

This approach is the best of both worlds, you only have to use your left thumb for movement programmed (W.A.S.D) to thumbwheel , your middile and fore fingers can be used at the exact same time for different functions/commands, much better then dedicating (3) fingers to W.A.S.D., while freeing your right hand for using the mouse for Aiming, Firing, and switiching weapons groups.

EDIT: With my gamepad I can pop the Left Thumbwheel and d-pad. flip it to put the D-pad forward of the thumbpad ... Saitek P-3600 gamepad.

Getting on target Fast and Accurately is a MUST for any Shooter, and the mouse cannot be beat. The learning curve and situational dependencies of a Flightstick/Joystick (only) setup limit it's performance for "aiming" against the Mouse.

When caught in a brawl against a pilot with a Flightstick/Joystick setup it's pretty obvious, the lack of pinpoint aiming against certain spots on a mech like the side torso will be witnessed first hand. I have been in fights where I can hear my teammate on TeamSpeak say, "That guy can't shoot, you got this easy", to which I reply, "He can shoot, but he's using a flight stick, it's over for him", and unless he has backup in that exchange, he's done for. I don't even have to Torso Twist to avoid losing a component, and that says a lot.

I like immersion it's a big part of this game for me, but so is NOT getting my Mech destroyed.

Want to debate which is better for MWO..??

Ok then, do this... Run a Spider 5-D with only one ERPPC and DO NOT just try and sit back and snipe, as a matter of fact run any light mech and try to achieve any sort of pinpoint aiming, while NOT sitting back and sniping.

Most people cannot do it effectively, exceptions to the rule do not disprove it.

Question: Can some pilots achieve good results with a Flightstick/Joystick (only) setup..??
ANSWER: Yes

Question: Does it matter when comparing Mouse versus Flightstick aiming..??
ANSWER: No, and I would bet a case of beer on that, without hesitation.

Edited by Odins Fist, 31 July 2014 - 07:34 AM.


#31 Loc Nar

    Member

  • PipPipPipPipPipPipPipPip
  • 1,132 posts

Posted 31 July 2014 - 10:07 AM

View PostVX Seraphin, on 30 July 2014 - 07:03 PM, said:

thankyou for the very in depth explanations of how various input devices work
i am more enlightened and that makes my day a better day :)


Heh, you are a real life mech driver. At first I was ready to point out that 'role 1' (positioning) and 'role 2' (rate) are already handled by the controller, but after reading further see you were not assigning them to separate devices, but as the two functions of the same device.

My summary: yep, position of a stick should directly equate to position of torso, and the rate which we move the stick should determine the rate of travel, all the way up the the speed of the mech's servos which can not be surpassed.

View PostOdins Fist, on 31 July 2014 - 07:15 AM, said:

Please do not take this post the wrong way..

-snip-

How do I know.?? I have (2) Saitek Flight sticks, I tried them, I liked the immersion factor, BUT it is simply impossible to get on target as fast or to achieve accuracy as FAST a mouse.

ALSO, I do not use the keyboard W.A.S.D., I use something similar to the Split Fish, I use the left side of a progammable Gamepad (Xbox style).. The left thumbwheel, left shoulder trigger(s), and the D-pad for functions, and have been doing this in First Person Shooters since COD-4.

This approach is the best of both worlds, you only have to use your left thumb for movement programmed (W.A.S.D) to thumbwheel , your middile and fore fingers can be used at the exact same time for different functions/commands, much better then dedicating (3) fingers to W.A.S.D., while freeing your right hand for using the mouse for Aiming, Firing, and switiching weapons groups.

EDIT: With my gamepad I can pop the Left Thumbwheel and d-pad. flip it to put the D-pad forward of the thumbpad ... Saitek P-3600 gamepad.

Getting on target Fast and Accurately is a MUST for any Shooter, and the mouse cannot be beat. The learning curve and situational dependencies of a Flightstick/Joystick (only) setup limit it's performance for "aiming" against the Mouse.

When caught in a brawl against a pilot with a Flightstick/Joystick setup it's pretty obvious, the lack of pinpoint aiming against certain spots on a mech like the side torso will be witnessed first hand. I have been in fights where I can hear my teammate on TeamSpeak say, "That guy can't shoot, you got this easy", to which I reply, "He can shoot, but he's using a flight stick, it's over for him", and unless he has backup in that exchange, he's done for. I don't even have to Torso Twist to avoid losing a component, and that says a lot.

-snip-

Want to debate which is better for MWO..??

-snip-

Question: Can some pilots achieve good results with a Flightstick/Joystick (only) setup..??
ANSWER: Yes

Question: Does it matter when comparing Mouse versus Flightstick aiming..??
ANSWER: No, and I would bet a case of beer on that, without hesitation.


Sigh... don't take this the wrong way, but I can't help but thinking you didn't bother reading this controls writeup you are responding to since you have mistakenly assumed it is some kind of defense of joysticks or attempting to initiate a mouse vs joystick debate. Quite the opposite in fact, as I wrote this as an attempt to kill those debates because they are typically noisy and lack fundamental understanding, to the point of causing pain. I'm a mechanical engineering student with leanings towards human factors (the people that design/make controls). It was an attempt at raising the quality of those discussions, by arming people with correct terminology and basic understandings of controls fundamentals, which gaming communities sorely lack.

"Want to debate which is better for MWO..??"

No. I want people to understand *why* a regular joystick sucks for reticule aim in MWO or any shooter that uses zero-order control, although I enjoy meaningful controls discussions. Keep in mind, I too cringe when I see some joystick'er on my team, which is immediately obvious, which tank goodness happens almost never anything north of Elo hell...

For the same technical reasons, a Warthog flightstick is as much a paperweight in MWO as my dedicated positional controller is a fish out of water in a flight sim, which has just as little business in an aircraft cockpit as a flightstick has in a mech. The controller I fabricated to play with is not your father's joystick... http://imgur.com/a/ixi64

Your observations of behavior based on regular old flight sticks is not incorrect, but only applies to normal sticks which are made for flight sims or other tasks where you are controlling either velocity (first-order control) or acceleration (second-order control). As such, you're statements are only 99.999% correct :D ...but that last 0.0001% may require a bit of digging into why controls behave the way they do to understand, so I recommend starting with my op.

What a mouse does is direct positional control, called zero-order control, and I spent a lot of words already pointing out why it's a terrible idea to use anything but controllers that are 100% dedicated to this task, because positional control breeds mutually exclusive hardware from the tasks of velocity/acceleration that regular joys are engineered for. Sadly no companies make a joystick suitable for positional control, so there is no reason you would have any first hand frame of reference to judge this by so I understand skepticism. The right stick of a Steel Battalion controller is the only example I can think of, and it's a poor one at that.

A positional joystick fundamentally requires absolute inputs (not that horrible thing they just patched in) to work, which to this day can only be achieved using absolute mouse emulation, which adds another barrier to viable alternate controls. Achieving absolute inputs is easy enough though, but a positional controller needs to mechanically support this type of input/movement so ideally should have:

No spring centering -the last thing you want is a spring influencing your position... could you imagine if your mouse had rubber bands pulling it to the center?

No center detentes -just as harmful to a positional joystick as a big ass crater would be in the middle of a mousepad

No deadzones -a dealbreaker of positional control, make the above crater also not track position

Kinematic related movements -the closer an input device moves compared to what it is controlling, the more intuitive inputs are. A direct match is called 'replica control', and is considered the best, which is how high end robotics like teleosurgery are handled.

Proper damping -needs consistent smooth resistance, strong enough to hold it's position when not being moved but not so strong it fights inputs. This is hard to achieve, and has the ability to make/break this scheme.

Moving on...

If your D-pad thingie is analog, then you are operating at max efficiency (well, doesn't sound like it would lend itself well to maintaining less than full speeds) however if it's generating 'w' and 's' that would be less than stellar due to the extremely slow ramp up/down this imposes on throttle. By comparison, analog throttle is instantaneous. If you only pilot fatties I suppose you might not notice, but it's night/day comparison when piloting lights, which is my default preference. When spectating, I can definitely tell who uses throttle vs kb...

Question: Can some pilots achieve good results with a Flightstick/Joystick (only) setup..??

ANSWER: Not really... not with a 'flightstick/joystick. They widdle themselves down to Elo hell and "adapt" by piloting mechs that are so big/slow that the mismatch of joystick velocity control are felt less. However a positional joystick a different beast altogether and actually has the potential to compete on even footing with other positional controllers such as mice

Question: Does it matter when comparing Mouse versus Flightstick aiming..??

ANSWER: Absolutely not, and if you read my op you would even be able to explain why this is the case! :wacko:

#32 Odins Fist

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 3,111 posts
  • LocationThe North

Posted 31 July 2014 - 10:32 AM

View PostLoc Nar, on 31 July 2014 - 10:07 AM, said:

Sigh... don't take this the wrong way, but I can't help but thinking you didn't bother reading this controls writeup you are responding to since you have mistakenly assumed it is some kind of defense of joysticks or attempting to initiate a mouse vs joystick debate.

Question: Does it matter when comparing Mouse versus Flightstick aiming..??

ANSWER: Absolutely not, and if you read my op you would even be able to explain why this is the case! :)


Nope, just pointing out that a Hybrid of the (2), Flightstick/Gamepad & Mouse is superior then just going with a only a Flightstick setup.

I would have liked it if MWO was Flightstick rudders only, but I realize that isn't possible.
Also that would have severely limited it's player base.

I do apologise if I came across as adversarial in showing my Frankenstein GamePad/Mouse setup, and advertising what I think are the advantages in gameplay.

Props to anyone that puts time and effort into projects like Mech Pits and Flightstick control mods and such, it really shows a dedication beyond just gameplay.

#33 Warblood

    Member

  • PipPipPipPipPipPipPip
  • The Hunter
  • The Hunter
  • 503 posts
  • LocationMontreal, Qc, Cnd

Posted 01 August 2014 - 12:38 PM



#34 AztecD

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 656 posts
  • LocationTijuana. MX

Posted 14 August 2014 - 12:55 PM

Just give me a step-step walktrough in seting up a joystick proerly within MWO :D

Please :wacko:

#35 Smokeyjedi

    Member

  • PipPipPipPipPipPipPipPip
  • Liquid Metal
  • Liquid Metal
  • 1,040 posts
  • LocationCanada

Posted 15 August 2014 - 04:11 AM

I think the lack of real world feed back (actual feel of mech)like twisting momentum and the bi pod mechs natural time to come to rest from torso twist will always limit the accuracy........If you were strapped in a chair with hydraulic risers and could feel the torso movements from the legs your aim would increase greatly.

#36 GreenHell

    Member

  • PipPipPipPipPipPipPip
  • 543 posts
  • LocationGrandmas House

Posted 24 December 2015 - 01:20 AM

You......... YOU!!! . . . . . . .

Now I have to buy a joystick just so I can tear it apart...
GRAAAAHHHH!!!!

As a question from somebody who just spent cash on a new T16000M (aka ME) do you think I could get away with only removing the built in spring, adding a "resistance disc" (mylar sheet) and using the TARGET software? I really don't want to moddify this thing to death like you guys. I mean... I play at a normal desk so......... yeah. I still have my left stick (logi Attack3), my "center console" (the separate stick shift part from a G27 racing wheel), and pedals (from the same racing wheel) so I'm covered there. I'm also doing a more Steel Battalion style as well with the pedals controlling throttle and the left stick doing the turning. Could I get away with a basic mod like that though? Are the sensors good enough?

#37 Foust

    Member

  • PipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 394 posts
  • LocationKentucky

Posted 24 December 2015 - 06:26 AM

View PostGreenHell, on 24 December 2015 - 01:20 AM, said:

You......... YOU!!! . . . . . . .

Now I have to buy a joystick just so I can tear it apart...
GRAAAAHHHH!!!!

As a question from somebody who just spent cash on a new T16000M (aka ME) do you think I could get away with only removing the built in spring, adding a "resistance disc" (mylar sheet) and using the TARGET software? <snip> Are the sensors good enough?


Yes you can and yes they are. In fact, I actually left the spring in mine after my first friction disc needed replaced.

#38 GreenHell

    Member

  • PipPipPipPipPipPipPip
  • 543 posts
  • LocationGrandmas House

Posted 27 December 2015 - 01:05 PM

I've noticed something while using FreePIE and my (admitedly bad) Logi-Stick. When I move slowly it moves things at one speed, but if I move quickly things move faster and further almost like it has some acceleration built in. Any ideas on how to kill this, or is it just my bad (and temporary) stick giving me issues?

#39 GreenHell

    Member

  • PipPipPipPipPipPipPip
  • 543 posts
  • LocationGrandmas House

Posted 03 January 2016 - 02:50 PM

Maybe you could help me with something Foust. I've gotten my new T16000M stick, and removed the spring (easy stuff). I've been tinkering with the TARGET software, and I must say that it works far better than I expected. However, my aim is still far too fast, especially when zoomed in on a target. When zoomed out I can aim "decently" but when I zoom in, each tiny movement still moves the aim more than the width of the target. Am I doing something wrong, or have I reached the limit of what can be done with my current setup? I really don't want to get into lengthening the throw on the stick, or custom building one, or re-writing the game...

#40 Foust

    Member

  • PipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 394 posts
  • LocationKentucky

Posted 03 January 2016 - 03:15 PM

MWO applies some scaling to the input when you zoom in. I started working on a way that when you toggle each zoom level that it would counter the scaling so that your movements don't get super sensitive while zoomed in, and return to normal when your zoomed out.

I haven't put much effort into it. I'll see what I can come up with.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users