Hard Light Productions Forums

Modding, Mission Design, and Coding => FS2 Open Coding - The Source Code Project (SCP) => Topic started by: Wanderer on May 23, 2009, 06:29:51 pm

Title: New attempt on the joystick tests
Post by: Wanderer on May 23, 2009, 06:29:51 pm
So.. here goes a second attempt to do some sort of joystick evals (again windozers only - sorry).

To test download the package, try the builds out flying a mission (of your own choice) which should feature precision aiming at distant spots as well as chasing of agile targets. DO NOT use only single sensitivity setting per builds, also do note that the setting can be changed even during the mission via F2 key. Also avoid extremes of the sensitivity settings (especially full sensitivity). Best would be if it is possible to test the build on wide variety of sensitivity settings.

After you have tested the builds post the identifiers of the builds in preference order (best -> second -> third -> fourth). Please do so even when you are unsure which build was better (but you can note that elsewhere in the post). Any comments or opinions of the sensitivity settings would be useful as well.

I'll later provide the info on what to change should Mac or Linux users want to take part in to the testing as well.

Link to the archive (http://tinyurl.com/oewyja).
Title: Re: New attempt on the joystick tests
Post by: Herra Tohtori on May 23, 2009, 08:14:50 pm
Results in spoiler tags to prevent expectations and bias from other testers. :p

Spoiler:
4760:

   -pretty good
   -low sensitivity is really low, but middle ground gives rather good control.
   
5336:

   -not bad, but didn't feel as comfortable as 4760.
   
5645:
   -awesome.
   
9915:

   -retail, if I'm not mistaken.
   -pretty horrible as such.


Also, the builds seem to consistently crash the mission "The Sixth Wonder" which I was using for testing. Happens always when the Colossus is about to open fire at the Hawkwood.
Title: Re: New attempt on the joystick tests
Post by: Wanderer on May 24, 2009, 02:21:10 am
Ok... archive reuploaded with builds marked with 'b'. Could you see if those work any better?
Title: Re: New attempt on the joystick tests
Post by: Herra Tohtori on May 24, 2009, 02:53:57 am
Confirmatory, no crash any more on that mission.

If I am to assume that the handling is the same, then the commentary stays same. 5645 feels very intuitive and natural on all sensitivity settings. Others are too non-responsive on lowest settings.
Title: Re: New attempt on the joystick tests
Post by: portej05 on May 25, 2009, 08:31:48 am
Spoiler:
I like the 4760 build on mid-high sensitivity.
Is it an exponential curve? - slow in the middle, and gets faster towards the outside

I'd rate 5645 second, 5336 third, and 9915 was not so good.
Title: Re: New attempt on the joystick tests
Post by: colecampbell666 on May 26, 2009, 03:12:20 pm
So what exactly is this?
Title: Re: New attempt on the joystick tests
Post by: Zacam on May 26, 2009, 10:35:06 pm
Various builds that you fiddle with the in-game joy sensitivity with to determine if you find one in favor of the other.

Note: Does not work so well with the absolute extreme ends of the settings for sensitivity and deadzone.
Title: Re: New attempt on the joystick tests
Post by: colecampbell666 on May 26, 2009, 11:29:01 pm
OK.
Title: Re: New attempt on the joystick tests
Post by: Wanderer on May 28, 2009, 05:04:57 pm
As it seems this thing wont progress... well here are the bits of code used to make the four different sensitivity settings..

Spoiler:
joy.cpp line 1143

4760:   percent = pow(percent, (4.0f - (Joy_sensitivity / 3.0f)));
5336:   percent = (pow(percent,2) * sensitivity_percent + pow(percent,4) * non_sensitivity_percent);
5645:   percent = pow(percent, (3.0f - (Joy_sensitivity / 4.5f)));
9915:    percent = (percent * sensitivity_percent + pow(percent,5) * non_sensitivity_percent);

Joy_sensitivity: 0 .. 9
sensitivity_percent: Joy_sensitivity / 9
non_sensitivity_percent: 1 - sensitivity_percent

9915 is the same as currently used setting
Title: Re: New attempt on the joystick tests
Post by: karajorma on May 30, 2009, 01:28:55 pm
Spoiler:
5645 without a doubt. I won't type out my comments cause they pretty much match HT's (Which I only checked after making my list.
Title: Re: New attempt on the joystick tests
Post by: kodiak7251 on November 04, 2009, 10:16:36 am
I get very little difference in sensitivity no matter what i set it at. even when i go into my logitec program and turn the settings way down the stick is still way tooo sensitive to my movements. is ther something else i should be setting besides the dead zone and sensitivity.
Title: Re: New attempt on the joystick tests
Post by: Herra Tohtori on November 04, 2009, 10:25:36 am
I get very little difference in sensitivity no matter what i set it at. even when i go into my logitec program and turn the settings way down the stick is still way tooo sensitive to my movements. is ther something else i should be setting besides the dead zone and sensitivity.


Are you trying to use these test builds, or is this more of a general issue? If it's a general issue, thread split to support might be a good idea.

It's recommendable to keep the dead zone as small as possible. That increases the effective range of motion of the stick and thus increases the amount of control you have over the control inputs.

Also, with the retail I personally recommend keeping the sensitivity maxed as well. The control curve that the sensitivity setting adjusts is rather strange and unintuitive and makes the controls unresponsive in the middle with huge spikes at the ends of the motion range.

Then there's the matter of logitech's sticks generally being of rather low quality in general. What stick do you have, and what kind of axis response curve does it have, if you can observe it somewhere?