SOLVED Math, tweening between 2 numbers - Chamaeleon is cool

[Foldup]

Hello my friends!!! I am not the best with math, I am sorry to say, and I have been struggling.

I want to use a function like LERP to move smoothly between 2 given values: currentwidth and goalwidth for smooth camera zooming.

I thought lerp(currentwidth,goalwidth,.5) was the way to go, but that's to literally give me the 50% value of those two numbers so I'm confused.

Actual code:

currentwidth = camera_get_view_width(view_camera[0])
camwidth=lerp(currentwidth,bracket[width],.5);
camheight=camwidth/ratio;

Any thoughts?

bracket[width] is an array that contains all the potential camera scaling options. I can just jump between them, but thought it would be nice to do it smoothly.

 

[Foldup]

Still working on this but my get camera view width is plainly wrong and returning wrong data....hm....

 

[NightFrost]

Whoops, ignore

 

[chamaeleon]

1. The second parameter, 0.5 in your example, is supposed to be a value between 0 and 1, with 0 resulting lerp's first parameter, 0.5 resulting in the midway point between the two values, and 1 resulting in the second value
2. Keep track of the starting value and the end value before you start the transition, and keep them the same throughout the entire transition, and over a suitable number of steps vary a variable for the third argument to lerp() between 0 and 1

Using the current value as the first value for lerp() and a constant 0.5 will give you the behavior of Zeno's Paradox, which is typically not what you want out of lerp() (although that is similar to ease-in/ease-out behavior).

 

[Foldup]

1. The second parameter, 0.5 in your example, is supposed to be a value between 0 and 1, with 0 resulting lerp's first parameter, 0.5 resulting in the midway point between the two values, and 1 resulting in the second value
2. Keep track of the starting value and the end value before you start the transition, and keep them the same throughout the entire transition, and over a suitable number of steps vary a variable for the third argument to lerp() between 0 and 1

Using the current value as the first value for lerp() and a constant 0.5 will give you the behavior of Zeno's Paradox, which is typically not what you want out of lerp() (although that is similar to ease-in/ease-out behavior).

Oh......OK. So you're suggesting that LERP should be used as a slider and I am meant to be popping in a variable there. Hm.....OK.....let's have a think.

Thanks for the input.

 

[chamaeleon]

var msg = "";
var i = 0;
var value = 0;
repeat 11 {
    value = lerp(0, 100, i/10);
    msg += msg == "" ? string(value) : ", " + string(value);
    i++;
}
show_debug_message(msg);

msg = "";
i = 0;
value = 0;
repeat 11 {
    value = lerp(value, 100, 0.5);
    msg += msg == "" ? string(value) : ", " + string(value);
    i++;
}
show_debug_message(msg);

0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100
50, 75, 87.50, 93.75, 96.88, 98.44, 99.22, 99.61, 99.80, 99.90, 99.95

Now imagine that instead of using a repeat loop, you perform the code in the code block once per step, incrementing i every step.

 

[Foldup]

Ouch...brain...no....

 

[chamaeleon]

Some initialization code

currentwidth = camera_get_view_width(view_camera[0]);
lerping  = false;
start_value = undefined;
end_value = undefined;
lerp_position = undefined;
number_of_steps = 2*game_get_speed(gamespeed_fps);

Some step event code

if (!lerping && should_start_lerping()) {
    lerping = true;
    start_value = currentwidth;
    end_value = bracket[width];
    lerp_position = 0;
    currentwidth = lerp(start_value, end_value, lerp_position / number_of_steps); // or currentwidth = start_value
} else if (lerping) {
    lerp_position++;
    currentwidth = lerp(start_value, end_value, lerp_position / number_of_steps);
    if (lerp_position >= number_of_steps)
        lerping = false;
}

Perhaps it works, perhaps it doesn't. Totally untested, and I might have overlooked something.

Edited to fix bug that compared lerping with number_steps

 

[Foldup]

I'm gonna go out on a limb and say yours would be better. BUT I have something that starts to improve on what I had without hurting my brain too much.

It's not a nice exponential movement, but it takes this arbitrary "currentwidth" value and tries to compare it to the goal of "camwidth"

camwidth=bracket[width];
if currentwidth != camwidth
currentwidth = (currentwidth + camwidth) / 2
camheight=round(currentwidth/ratio);

Ohhhhhh.....I think I can actually understand what you're doing in that code. I may be getting smarter just by being around you, haha.

 

[Foldup]

By the way, here's the actual game where all this camera funzies is happening. Starting to look like something here.

 

[Foldup]

OH! Holy crap, I DO understand it!!!!

The only error in the code was
if (lerping >= number_of_steps)
should have been
if (lerp_position >= number_of_steps)

That IS smoother. You're too cool, Chamaeleon. For reals.

 

[chamaeleon]

Oops, sorry for the bug. I'm glad it was of some help!