View Full Version : using game data always the best method?

09-12-2007, 11:19 AM
Hi all,

I actually have never asked this question before, but after being invited to the Open Source Motion Forum it kind of got me thinking about the software side of things and the question as to which method is the best and why for motion control.

(BTW thanks for the invite -- I plan to share any of my hardware designs for sure)

so....let me draw attention to the software side which has provoked the big question.

When it comes to game control, there are generally two methods for interaction.

1. Game Data Method
2. Game Controller Method

I have chosen the later for now as it works with all PC games as well as provides all the information needed for representing most of a vehicle's motion.

Game data is a future method that I would be looking into because when all the hardware is working and runs to satisfaction, I would welcome the extra motions that only game data can provide.

Game data extraction method seems to be a proverbial favorite amoung many (if not all) motion cockpit builders, yet it does requires so many more things to be in place....... maybe I am wrong?

From my humble opinion, Game data appears to require that the target game have a method of built-in (or at least capable of) communication through third-party data extraction. This method requires an active involvement with the actual game developers or with individuals who are skilled in programming this functionality. Optionally, some of the popular sim titles seem to have public game extraction software created for them already which can make things much easier for those specific titles and many of you are using these currently.

The Game controller method on the other hand allows you to control your motion with literally thousands of more titles which opens some real posibility for great legacy games including space sims or even newer games in which the developers are reluctant to provide such game telemathy.

Now, this is just a personal opinion, but it seems that it would be easier or at least advantages to develop game controller apps for motion cockpits because at the very least you can test things out while playing a game before having to focus on the game data hooks and everything associated with those.

Game controller control also does not need encoders which can save on time and extra hardware if your just starting out. Now, with no encoders it is basically a "follow the stick or pedel" method with limiters for your range end, but I have not seen that the accuracy of encoders is needed unless you want to go half way through a move like a half roll and then hold. Certainly this would add more fine tuning, but it's seems to be a good enhancement more then a requirement. For game data, encoders and the code/electronics for them are certainly a requirement.

Again, I am not condeming game data as it is certainly beneficial because it can yield some very unique effects and if its offered, by all means tap into it.....but is it the only one that makes sense?

Below is an overview that compares the two (again my observation) Please offer some comments because I am just wondering why motion simmers have "game data" as a preference

Maybe it is that all the games that most motion simmers would want to play does have game data apps available?

maybe I am just thinking along a different line?

here is the list:

Game Controller Actions
Landing Gear/Boost

Game Data Only
Ground Textures
Taking Fire
Physical Bumps

For any of you that have worked with both, Does game data pervide a better experience for the Game Control Actions listed above?

So there's my question/s I certainly have game data control on the radar, but for know I am focusing on game controller control cause it seems to be a good fit. But maybe it is all personal preference or maybe there is some other a reason why many have stayed away from this method?

I also just went to X-sim site and looked at how they handle the game controller method and it seems they do but are saying:

"d) if your a game does not support ForceFeedback you can use either a game specific plug-in or your joystick movements (not recommended)"

Hmmm .......again not reccomended.

Has anybody played around with game controller control in X-sim or even your own home grown app?

Regards Kyle

09-13-2007, 12:59 PM
Hi kyle check out craig warners website he is currently working with X-sim and has posted some data regarding your query :-) you can find his webpage listed in the links on Jims website

Bob Reed
09-13-2007, 02:07 PM
One of the problems you will encounter with controller data only is if the visuals get out of sink with the motion.. Have a barf bag ready!!

09-13-2007, 03:17 PM

Got to say I have not heard of that one yet. :)

But I can see it. Probably depends on the user and motion speed and travel of the actions.

I would think once setup delay should not be any different from game data, but I can see more processing happening with the stick to motor path. Although probably something that might just happen time to time.

I would view stick control as a starting point though with game data being the light at the end of the tunnel.


do you have a link to Craigs thread on stick control with X-sim I have not found anything on his site or looking around..... still looking.

Regards Kyle

09-14-2007, 05:13 PM

I think to an extent the choice depends on what a builder wants from the sim. If the builder's interest is to make the motion cues as realistic as possible (given the usual limitations of a stationary simulator) then it makes a lot of sense to derive those cues from data that explicitly defines the motion of the aircraft using variables and units that are part of the physics of the system. For example to obtain an understanding of how close to realistic your force cues are you need to be able to quantify the force effects and to do this you need to know actual magnitudes and directions of accelerations in recognisable units eg m/s^2, ft/s^2 or even g's. If you want to determine the difference between the heave effects of turbulence and those of manoeuvres so that you can scale them differently you need to know the magnitude of the different effects etc etc. The best place to get this realistic quantified data is directly from the physics engine of the simulator - because it has to work it out to realistically simulate the aircraft motion.

Generally I would think that it would be easier to develop more realistic (and effective?) simulations where you have an accurately quantified description of the physics of the motion being simulated. True, the downside is not all "games" make their internal physics accessible - a balance between quality and quantity?

A further point to note is that the motion of the aircraft is described fairly completely by the internal physics - the force effects of all manoeuvres, orientations etc can be determined from a fairly short list of velocities, accelerations, orientations etc so the effects of climbs, rolls, braking, accel, side slip, turbulence effects etc etc are all in there in a small set of game data.

Just a thought. I've kept away from controller data so far because it keeps me at arms reach from the real physics of the motion.


09-17-2007, 10:28 AM

yea, I think the concensus has been like you said "quality vs quantity" as far as games.

heave effects for example have very little value being controller (button) driven. Retractable landing gear, activating a pitstop (jack up car) maybe, but other then that, there are far to many of these effects that are game data only. For the others I think that realism can be added with the extra information like m/s^2, ft/s^2 or even g's. It would be interesting to compare both methods to log the actual differences.

I am looking at X-sim currently as they seem to have a good thing going with that tool and the next version is suppose to support pwm output to one of the I/O cards I am using.

As mentioned I am using controller method for testing and I do want to be able to experience many games that are not data, but with the sims that have this internal ability I definately see the value.

It has been interesting getting others opinion. Definately comes down to the users needs.

thanks for your response

Regards Kyle