Scale mouse/joystick sensitivity depending on FOV

[ryvnf]

Fixes #11005 for joystick and mouse input.

How to test

Use the zoom-key and compare the mouse responsiveness when zoomed in and not. It should feel just as responsive.

Compare that to without this patch where it will feel a lot more responsive when zoomed in.

[lhofhansl]

I like the idea. Tested and works.

Took me a minute to reason through why tan(fov / 2.0).

[Zughy]

Really needed feature for FPS games. Tested, works, just two things:

1. mouse sensitivity: what was 0.2 before, it's now 0.25 (in terms of speed)
2. I've tested in my FPS and it feels a bit too slow when zooming in. (I tend to play a lot of frantic FPSs, just to say I know what I'm talking about)

[ryvnf]

• mouse sensitivity: what was 0.2 before, it's now 0.25 (in terms of speed)

• I've tested in my FPS and it feels a bit too slow when zooming in. (I tend to play a lot of frantic FPSs, just to say I know what I'm talking about)

What aspect ratio and FOV do you use? I did make it such that the speed should be equal from before the change when playing on 72 degree FOV at 16:9 aspect ratio. I verified that the commited function returns 1.0 when playing using these values. I assume that is what most people play with.

[nerzhul]

this constant should be a DEFINE in order to understand what is this directly

[ryvnf]

I found it hard to give such a constant a meaningful name so I thought it was more clear just to multiply with it and explain its purpose in the comments. If std::tan was constexpr in C++11 I could do the something like following which would be more clear about its purpose and where it comes from:

static constexpr float default_fov_y = 72.0f * M_PI / 180.0f;
static constexpr float constant = 1.0f / tan(default_fov_y / 2.0f);

This works for gcc but as far as I know std::tan isn't is not guaranteed to be constexpr in C++11.

Maybe you have a better idea what to call such a constant?

Also where should I put it? I think it is better to have a constexpr variable in the function instead of a global macro because it is only used in that function.

[lhofhansl]

Personally, I think the comment right above it good enough to explain it.

[lhofhansl]

Off-topic: @Zughy what mods do you use for your Minetest FPS games?

[Zughy]

@ryvnf I didn't change the default FOV and I have a 16:9

@lhofhansl https://gitlab.com/zughy-friends-minetest/block_league (you can find it on AES server)

[ryvnf]

I didn't change the default FOV and I have a 16:9

Mathematically this shouldn't be the case. The sensitivity is multiplied by 1.3763818698. That constant is given by

1 / tan(degree_to_radians(72) / 2) = 1.3763818698

which when multiplied by tan(degree_to_radians(72) / 2) the result becomes

 tan(degree_to_radians(72) / 2) / tan(degree_to_radians(72) / 2) = 1

[lhofhansl]

Where are we with this? I like this feature and stand by my "One Approval".

[SmallJoker]

Works for me. Initially the movement seemed very slow to me, but I suppose that's because I'm used to the normal speed.

[sfan5]

Fine by me but we might have to add a setting for this once someone complainsnames a good usecase.