Hard Light Productions Forums

Modding, Mission Design, and Coding => FS2 Open Coding - The Source Code Project (SCP) => Topic started by: Wanderer on December 22, 2008, 07:34:28 am

Title: Blind trial on joystick
Post by: Wanderer on December 22, 2008, 07:34:28 am

So... a minor test on the joystick behavior. The archive contains four somewhat different FreeSpace Open release builds (using SVN trunk code) marked with id code. If you want to help in figuring out the preferred method for handling the joystick sensitivity/attenuation please try the builds out.

The archive (http://koti.mbnet.fi/vekkup/FS2/Arkisto/joystick_blind_trial.7z) - Only for windows, sorry, but i couldn't have kept it a blind trial otherwise.

What to do
- Test each build with different sensitivity settings - disregarding the maximum sensitivity setting. Note that you can change the value during the mission via options screen by pressing F2
- No need to test every value, just testing with one high, one mid and one low sensitivity setting should be enough
- Using same sensitivity values for different builds may help when trying to rate the builds
- If you want to you can also rank the builds and post the ranking order into the thread
- Maximum two ( 2 ) votes per user - should you be unable to decide between a pair of builds
- Feedback and comments on the different builds is always welcome.
- Please do not test using maximum sensitivity setting as all versions use same 'ramp' with maximum joystick sensitivity setting

Title: Re: Blind trial on joystick
Post by: Reprobator on December 22, 2008, 02:04:02 pm

ah ...i always play at maximum sensitivity, is it bad?

Title: Re: Blind trial on joystick
Post by: Wanderer on December 22, 2008, 02:26:54 pm

Well.. At maximum sensitivity you tend to lost all the fine control over the ships movements (try aiming a single static pixel on screen). Maximum sensitivity setting translates the current stick movement into ship movement as almost directly proportional to it (only deadzones affect it).

But then again its often rather difficult to do fine moves using maximum sensitivity. So you gain some and you lose some.

Actual sensitivity setting of course depends only on you prefer. But the idea behind this test is to see if a blind trial could be done on such subject and also if we could find a better tradeoff for the sensitivity setting (so while still keeping the equal amount in ability to comfortable control of the ship we would lose less in the fine control).

Title: Re: Blind trial on joystick
Post by: Reprobator on December 22, 2008, 03:00:38 pm

thanks for the explanation , so i ll try a bit less sensitive (1 or 2 point under the maximum and try those different builds)

ok after trying it i prefer the 552 one.

Title: Re: Blind trial on joystick
Post by: FUBAR-BDHR on December 22, 2008, 03:21:14 pm

I've always had the setting near maxed as well.  Using a couple of different joysticks and it seems to be more a stick issue then a game issue.  What I wouldn't give for a good old sidewinder pro.  Now there was a stick I could make all 8 Kysers on an Ery hit at max distance. 

Title: Re: Blind trial on joystick
Post by: Wanderer on December 22, 2008, 04:14:26 pm

Quote from: Reprobator on December 22, 2008, 03:00:38 pm

thanks for the explanation , so i ll try a bit less sensitive (1 or 2 point under the maximum and try those different builds)

ok after trying it i prefer the 552 one.

It might have been had you tested on the sensitivity values from the other end of the spectrum as well but them again as maximum (or near maximum) sensitivity is that what you want to use it might be difficult to compare the other values.

Title: Re: Blind trial on joystick
Post by: Reprobator on December 23, 2008, 01:09:22 am

i've try with 2 point under the max value if it's what you wanted to know?

Title: Re: Blind trial on joystick
Post by: Wanderer on December 23, 2008, 09:01:47 am

Well... as i said if you are used to that setting then that is your result


However what i meant was that it would be good (but not required) to test different sensitivity settings. For example like so...

Code: [Select]

High sensitivity test
  X X X X X X X X - -

Mid sensitivity test
  X X X X X - - - - -

Low sensitivity test
  X X - - - - - - - -Where X marks the light blobs in the options table

Title: Re: Blind trial on joystick
Post by: Reprobator on December 24, 2008, 01:29:28 am

ok , well now i won't get back home till monday , so i ll can't give a try now.
but i'll do

Title: Re: Blind trial on joystick
Post by: Wanderer on December 24, 2008, 02:53:03 am

If you want to... But just to make it clear.. There are no wrong answers to this test.

Title: Re: Blind trial on joystick
Post by: Nuke on December 28, 2008, 01:15:17 am

i always perfered games to support joisticks as perfectly linear with no deadzone. most decent joysticks come with software to set deadzone and gain. but since fs already came with a deadzone feature. i guess it can stay. the only time i really like a deadzone is when using newtonian physics (like in orbiter). either way i usually keep the advanced settings in the joystick software.

does fs handle joystcks as 8 bit or as 16 bit?

Title: Re: Blind trial on joystick
Post by: karajorma on January 04, 2009, 05:29:01 am

:bump:

It is very much in the interest of those of you with joysticks to test this. Speaking as an SCP coder myself I do know that this sort of thing follows a distinct pattern.

1) Write the code
2) Ask for help testing it
3) Get little or no help
4) Commit the code the way you and the few hardcore people who did test like it
5) Get complaints from those who don't like it that way weeks or months after you last touched the code.
6) Stick fingers in ears and go "LA LA LA I CAN'T HEAR YOU!"

:p

Title: Re: Blind trial on joystick
Post by: Nuke on January 04, 2009, 09:08:04 am

sorry been kinda slow getting around to it, finding a new job and all. :D

Title: Re: Blind trial on joystick
Post by: Nuke on January 04, 2009, 10:09:28 am

i used my ch fighterstick pedals and throttle using my usual ch profile, which has no ramping or deadzones set. i flew the first training mission in the game flying a myrmadon. traditionally i play freespace with the setting maxed. i did 2 test passes on the instructor, one was a close persuit test, where i follow the ship keeping my crosjhair on the lead. and the other a full afterburn joust from 1000 meters. the first test was to see how easily it was to track presicely, as the second was a test of how quickly you could line up before you had to veer away.

413 is really good at the lower settings but as you pass mid it gets sensitive a tad too fast. its the easyest to hold a target at low sensitivity. very easily handled the pursuit test at all settings. and the joust was best on mid to max-1 (which is to be expected).

552 sucks, its just horrible, feels like theres a massive deadzone at the lowest setting. higher settings are better but it dont feel right. feels almost laggy. failed the pursuit test, which is kinda sad since the instructor doesnt fly all that fast. couldnt lign up at all on the joust.

717 is about the same as 552. i was having difficulty keeping the reticle on the instructor in the first training mission on the lowest setting. the higher settings also seem laggy.

979 wins at a distant second, has the same low sensitivity problem as the last 2, but handles higher settings better. persuit was easyer at higher sensitivity while the joust worked on mid to upper.

conclusion: 413 felt the best all across the board. none of the other builds were even remotely playable at the lowest or sometimes mid levels.

Title: Re: Blind trial on joystick
Post by: Reprobator on January 05, 2009, 04:07:31 am

after testing it with different sensitivity setting i still prefer the 552 one with the full -1 sensitivity set.

I got a little question anyway, i ve asked it already a long time ago , i wonder if it would be possible ( (i mean hard/long) to make freespace2 to handle multiple input (my x52 and my g25 wheel natively) as i d like to use it for it's pedals...

Karajorma reply me to use ppjoy , but even with the differents tutorials i've never managed to get it worked :p

anyway this is not a really big issue for me i can live with that :D 

Title: Re: Blind trial on joystick
Post by: karajorma on January 05, 2009, 05:06:11 am

It's possible and it is coming AFAIK but due to the changes coming in FS2_Open 3.7 it has to wait until then.

Title: Re: Blind trial on joystick
Post by: Reprobator on January 05, 2009, 06:44:00 am

That's a great news!

Title: Re: Blind trial on joystick
Post by: Nuke on January 05, 2009, 11:10:52 am

should be possible with sdl, but im more intrested in seeing more axes used by the game, lateral and vertical thrust axes, reverse trhust axes or a split thottle mode with forward and reverse. and having additional axes script accessible.

Title: Re: Blind trial on joystick
Post by: Wanderer on January 05, 2009, 02:58:10 pm

Two weeks have passed... and imo its time to tell here what was being tested - the actual description of it is within spoiler tags so should you still want to make the test.

That is... First test the builds and post the answer and only after that read what exactly you were testing to avoid having your existing opinions affecting the actual results.

Or just keep reading...

Spoiler:

Ok.. so here is actually what was being tested...

First couple of definitions

   pct = (float, 0 .. 1) amount of stick movement from measured from the edge of the deadzone(s)
   pct final = (float, 0 .. 1) actual turning multiplier applied to the 'max rotation' to get the desired rotation
   sensitivity = (integer, 0 .. 9) the number of the sensitivity 'lights' in options screen - 1
   sensitivity pct = (float, 0 .. 1) amount the current sensitivity setting is from the max sensitivity (ie. sensitivity / 9 )

Then the actual descriptions in mathematical formula.

413   -  pct final = pct ^ ( 3.0 - ( sensitivity / 4.5 ) )
552   -  pct final = ( pct * sensitivity pct ) + ( 1 - sensitivity pct ) * pct ^ 5
717   -  pct final = pct ^ ( 4.0 - ( sensitivity / 3.0 ) )
979   -  pct final = ( pct * sensitivity pct ) + ( 1 - sensitivity pct ) * pct ^ 3

Of the used joystick sensitivity settings the 552 is the same as the one currently in use in the FS Open. Sorta like control.


On graphs... the x-axis is the 'pct' and y-axis is 'pct final'. And they should reveal what was being tested... It should be rather obvious that with sensitivity setting used on 552 and 979 getting good fine control seems to be much more difficult than on the other types.

Also i locked the voting but comments - and additional opinions on the test - are welcome.

[attachment stolen by Slimey Goober]

Title: Re: Blind trial on joystick
Post by: chief1983 on January 05, 2009, 03:47:24 pm

I just want to throw out that I think this needs to be done in some way that it doesn't have to interfere with the original behavior.  This is an instance where a command line option shouldn't cause any headaches for anyone by not having them enabled, so I think leaving retail default in place unless a "Use Alt JS Mode A, Use Alt Mode B, etc" option is enabled would make sense here.  Once an in game menu for command line options is developed, it could be translated to a radiobutton-style option.

Title: Re: Blind trial on joystick
Post by: Nuke on January 05, 2009, 07:54:29 pm

so long as the range is the same (-1 to 1?) when the computations are done it shouldnt matter. not al joysticks are made equal. they have different gain settings, mechanical workings that differ greatly. then you have things like ch control manager, where you get fine control over the gain. some people have higher precision controllers. so long as the range is respected you wont change the way the ships handle. any changes that do occur would be no different that what you get between various kinds of joysticks.

Title: Re: Blind trial on joystick
Post by: chief1983 on January 05, 2009, 08:01:59 pm

What would really be nice is a full blown curve editor like in IL-2.  That's customization there.  And a tester to go with it to show you what the effective position is based on the real position.  That could even be done with just this.

Title: Re: Blind trial on joystick
Post by: Nuke on January 05, 2009, 10:53:38 pm

my ch software does that.

software alone is a good reason to go with ch. there's not really a damn thing it cant do.

Title: Re: Blind trial on joystick
Post by: Wanderer on January 05, 2009, 11:37:17 pm

Full user done customization would be of course nice... But that would require creating new options screen for joystick setup and all. Though it could be done..

Quote from: chief1983 on January 05, 2009, 03:47:24 pm

I just want to throw out that I think this needs to be done in some way that it doesn't have to interfere with the original behavior.  This is an instance where a command line option shouldn't cause any headaches for anyone by not having them enabled, so I think leaving retail default in place unless a "Use Alt JS Mode A, Use Alt Mode B, etc" option is enabled would make sense here.  Once an in game menu for command line options is developed, it could be translated to a radiobutton-style option.

Well... i would agree if i would think there was anything to be saved from the original sensitivity setting. However as there IMO are no redeeming qualities i cant see why the earlier behavior should be kept.