created_at2020-08-11 15:53:50.913667
updated_at2024-04-03 02:49:06.750522
descriptionA tweening library for games
Jonathan Spira (sanbox-irl)




tween is an std-optional tweening library, designed for use in games and animations.

Quick Start

To install, add the following to your Cargo.toml:

tween = "2.0.2"

You can make a Tweener, like this:

use tween::Tweener;

let (start, end) = (0, 100);
let duration = 15.0;

let mut tweener = Tweener::sine_in_out(start, end, duration);
let mut position = 0;

const DT: f32 = 1.0 / 60.0;

// and then in your main loop...
loop {
    position = tweener.move_by(DT);
    if tweener.is_finished() {

assert_eq!(position, 100, "we've moved to the end of the tween");


A Tween is a function which returns values from one specified number to another specified number over a specified amount of time. The simplest Tween which everyone is familiar with is lerp, or "linear interpolation". In this library, it is called Linear -- a Linear Tween moves from its start to its end by the formula start * (1.0 - p) + end * p, where p is the percent over time you are into the tween. So at time 0, or 0%, you end up with the start value, and at time 1, or 100%, you end up with the end value.

There are many kinds of Tweens beyond Linear, of course; all of which can produce feel and animations! We often use tweens to move positions of objects in games, but you can use Tweens for animating a sprite, selecting behavior, audio manipulation, or even drawing fonts with Cubic Bezier Tweens.

This library gives you access to all the tweens initially made by Robert Penner -- you can see them in action here.

This library exposes three kinds of structs:

  • Zero-Sized Tweens which implement the Tween trait. They also expose the method tween inherently, so you can tween easily with them, like tween::Linear.tween.
  • Wrapper Tweens which implement the Tween trait. These are Looper, Oscillator and Extrapolator. These all wrap around other Tweens. See their documentation for more information.
  • Tweener and FixedTweener, both of which "drive" a Tween. You should use FixedTweener in a fixed timestep application; otherwise, use Tweener. Although you can use a Tween directly, a Tweener manages all the Tween state for you.

For 99% of users, you'll want to construct Tweeners or FixedTweeners with a Tween for this library, occasionally looping or oscillating them.

Making Tweens Yourself

If you'd like to make your own tween, you absolutely can! For that, you'll need to see the main trait of this library: Tween. You can prototype your own Tween implementations with a simple closure, since FnMut(value_delta: Value, percent: f32) -> Value implements Tween, or just use that yourself. Closures are very nice because you can add, or even composite, tweens in interesting ways.

For example, here's a Linear tween averaged with a SineIn tween:

use tween::{Tweener, SineIn, Linear};

Tweener::new(0.0, 10.0, 10, |value_delta, percent| {
    (Linear.tween(value_delta, percent) + SineIn.tween(value_delta, percent)) / 2.0

To see a documented example of a Cubic Bezier Tween, see examples/

Storing Tweens

Very often in a game or animation engine, you'll want to store Tweens by what they act on, without caring about what kind of Tween example it is. To do that, you'll want to box the Tween within the Tweener. Since all the Tweens in this library are ZSTs, the Box won't actually allocate, though you will have to use dynamic access (which will be more than fast enough).

use tween::{Tweener, Looper, Linear, SineIn, Tween};

// very often game engines need sync/sync
type SendSyncTween<Value, Time> = Tweener<Value, Time, Box<dyn Tween<Value> + Send + Sync>>;

let mut my_tweener: SendSyncTween<i32, i32> = Tweener::new(0, 100, 100, Box::new(Linear));
let mut going_up = true;

// we lerp from 0 to 100 over 100 frames, and then we flip our tween back
// into a SineIn tween over 10 frames, so this looks like a slowwwwwww buildup
// and then a SHARP drop down.
// we put this in a thread here to demonstrate `Send + Sync`
std::thread::spawn(move || {
    loop {
        let _output_assigned_somewhere = my_tweener.move_by(1);
        if my_tweener.is_finished() {
            my_tweener = if going_up {
                Tweener::new(100, 0, 10, Box::new(SineIn))
            } else {
                Tweener::new(0, 100, 100, Box::new(Linear))
            going_up = !going_up;

To see a documented example of erased Tweeners, see examples/

Implementing TweenValue

This library uses two traits: TweenTime and TweenValue. You can implement these yourself, but implementing TweenTime would only have fairly obscure uses.

On the other hand, TweenValue needs to be implemented for any tweenable value. By default, all numerical types are already implemented in this library. Additionally, several math libs have a feature flag (see below) which gates an implementation for their structs as appropriate.

Going Fast ⚡️

This library is, ultimately, a math library, and benefits enormously from being in release mode.


tween has the following features:

  • std: enabled by default, gives access to faster floating point math and helper methods with Box
  • libm: enable this, without default features, for no-std tweening
  • glam: enable this for glam types to implement TweenValue
  • nalgebra: enable this for nalgebra types to implement TweenValue
  • vek: enable this for vek types to implement TweenValue
  • ultraviolet: enable this for ultraviolet types to implement TweenValue
  • ultraviolet-f64: enable this for ultraviolet/f64 types to implement TweenValue
  • cgmath: enable this for cgmath types to implement TweenValue

This library supports glam as its first-class math lib.

Std Optional

This library uses std with the default feature std. Disable default features, and enable libm, for a no-std experience. (We need to use libm for the floating point math), like so:

tween = { verison = "2.0.1", default_features = false, features = ["libm"] }

MSRV and Safety

This crate has no MSRV yet. If it sees good adoption, an MSRV policy will be decided.

Additionally, this crate is #![deny(unsafe_code)], since no unsafe code was needed. Changing this policy would constitute a minor breaking change.

Breaking Changes

This crate follows normal rules for breaking changes except for math libraries besides glam. We make no promises that we'll update perfectly with math libraries -- you are encouraged to make forks of this repo instead if you have version handling requirements.


Next up for this library is handling Splines of Tweeners.


Dual-licensed under MIT or APACHE 2.0.

Commit count: 166

cargo fmt