Quadratic Bézier Curve

In the digital world, smooth, elegant curves form the backbone of everything from the fonts on this page to the aerodynamic shell of a modern vehicle. But how does a computer, which thinks in straight lines and discrete numbers, capture the continuous, flowing essence of a curve? One of the most fundamental and elegant answers is the Bézier curve, and in its simplest, most intuitive form, the quadratic Bézier curve. While many may encounter its formula as a given, a deeper understanding reveals a beautiful interplay of geometry, algebra, and physics. This article peels back the layers of this essential tool, moving beyond rote memorization to build a powerful, intuitive grasp of its nature.

This exploration will unfold in two main parts. First, the chapter on ​​Principles and Mechanisms​​ will demystify the curve's creation through a simple geometric game, translating that motion into its famous algebraic formula. We will uncover the secrets of its properties, like the "safe enclosure" of the convex hull and the surprising revelation that every quadratic Bézier curve is a familiar parabolic arc. Following this, the chapter on ​​Applications and Interdisciplinary Connections​​ will journey through the practical world, showing how these principles empower artists, designers, and engineers. We will see how these curves are used in vector graphics, pieced together to form complex shapes, and how their underlying mathematics connects to advanced simulation techniques in fields like computational physics, demonstrating their role as a universal language for describing shape.

To truly understand the quadratic Bézier curve, we must not see it as a static formula to be memorized, but as a dynamic, living thing. It is born from a simple and elegant process, a kind of geometric game that we can play in our minds. Once we understand this game, the mathematics behind it becomes not a chore, but a beautiful and logical description of the game's rules.

Imagine you have three points, let's call them $P_0$, $P_1$, and $P_2$, sitting in a plane. These are our ​​control points​​. $P_0$ is the beginning, $P_2$ is the end, and $P_1$ is the guide, the influencer, the puppet master that dictates the curve's bend.

Now, let's connect these points with two imaginary lines: one from $P_0$ to $P_1$, and another from $P_1$ to $P_2$. Think of these lines as tracks. On the first track, $P_0P_1$, we place a bead, let's call it $Q_0$. On the second track, $P_1P_2$, we place another bead, $Q_1$.

We are going to move these beads along their tracks according to a single, simple rule. We have a master clock, or a parameter, which we'll call t, that runs from 0 to 1. At $t=0$, $Q_0$ is at $P_0$ and $Q_1$ is at $P_1$. As $t$ smoothly increases towards 1, $Q_0$ glides along its track towards $P_1$, and $Q_1$ glides along its track towards $P_2$, both at a constant rate. They complete their respective journeys at the exact same moment, when $t=1$. So at any instant $t$, $Q_0$ has traveled a fraction $t$ of the way from $P_0$ to $P_1$, and $Q_1$ has traveled the same fraction $t$ of the way from $P_1$ to $P_2$.

But where is our curve? We're not done yet. As our two beads, $Q_0$ and $Q_1$, are moving, we connect them with a third imaginary track, the line segment $Q_0Q_1$. And on this moving, shrinking, and rotating track, we place our final point, the star of the show, which we'll call $B$. This point $B$ plays the same game: as $Q_0$ and $Q_1$ move, $B$ glides along the track between them, always having traveled the same fraction $t$ of the distance.

The path that the point $B$ traces as $t$ goes from 0 to 1 is the quadratic Bézier curve. This recursive process of linear interpolation is known as ​​de Casteljau's algorithm​​. When $t=0.5$, for instance, $Q_0$ is the midpoint of $P_0P_1$, $Q_1$ is the midpoint of $P_1P_2$, and our curve point $B(0.5)$ is the midpoint of the segment connecting those two midpoints. This construction gives us a powerful intuition: the curve is a blend of blends, a smooth compromise between the straight lines of the control polygon.

This geometric dance is delightful, but for a computer, we need a more direct language: algebra. Let's translate our game into equations.

The position of our first bead, $Q_0$, which moves from $P_0$ to $P_1$, is a simple linear interpolation: $Q_0(t) = (1-t)P_0 + tP_1$

Similarly, the position of our second bead, $Q_1$, is: $Q_1(t) = (1-t)P_1 + tP_2$

Our final point, $B(t)$, interpolates between $Q_0(t)$ and $Q_1(t)$: $B(t) = (1-t)Q_0(t) + tQ_1(t)$

Now, let's do something a physicist loves to do: substitute and see what shakes out. We replace $Q_0(t)$ and $Q_1(t)$ in the final equation with their definitions: $B(t) = (1-t)((1-t)P_0 + tP_1) + t((1-t)P_1 + tP_2)$

If we expand and collect the terms for each control point, we find something quite neat: $B(t) = (1-t)^2 P_0 + (1-t)tP_1 + t(1-t)P_1 + t^2 P_2$ $B(t) = (1-t)^2 P_0 + 2t(1-t) P_1 + t^2 P_2$

And there it is. The famous formula for a quadratic Bézier curve. It's not some arbitrary incantation pulled from a hat; it is the direct algebraic consequence of our simple, intuitive game of sliding beads.

Let's look closely at the "coefficients" in front of our control points: $(1-t)^2$, $2t(1-t)$, and $t^2$. These are the ​​Bernstein basis polynomials​​ of degree 2. Think of them as "dials of influence" or "weighting functions". For any value of $t$, they tell us how much "pull" each control point has on the final position of the point $B(t)$.

• At $t=0$, the weights are $(1, 0, 0)$. All the influence belongs to $P_0$, so $B(0)=P_0$. The curve starts at $P_0$.
• At $t=1$, the weights are $(0, 0, 1)$. All the influence belongs to $P_2$, so $B(1)=P_2$. The curve ends at $P_2$.
• At any $t$ between 0 and 1, all three control points have some influence. For example, the influence of $P_1$, given by the term $\beta(t) = 2t(1-t)$, is zero at the ends and reaches its maximum value at $t=1/2$. This confirms our intuition that the middle control point has the strongest pull on the middle of the curve.

These weights have two magical properties. First, for any $t$ in $[0, 1]$, they are all non-negative. Second, they always add up to exactly 1: $(1-t)^2 + 2t(1-t) + t^2 = (1-2t+t^2) + (2t-2t^2) + t^2 = 1$ This "partition of unity" property is profoundly important. An average where the weights are non-negative and sum to one is called a ​​convex combination​​. This means that for any $t$, the point $B(t)$ is located inside the triangle formed by the three control points, $\triangle P_0P_1P_2$. This is the ​​convex hull property​​. It gives designers a wonderful guarantee: no matter where they place $P_1$, the curve will never fly outside the triangle defined by the control points. It is tamed, predictable, and safely contained within its "hull".

So far, we have a geometric definition and an algebraic one. Now, let's put on our physicist hats and imagine a particle moving along this path, $B(t)$, where $t$ represents time. What is its velocity? Its acceleration? The answers reveal the deepest secret of the quadratic Bézier curve.

The velocity vector, $B'(t)$, is found by taking the derivative of our master formula with respect to $t$. The result is wonderfully symmetric and relates back to our control polygon: $B'(t) = 2(P_1 - P_0)(1-t) + 2(P_2 - P_1)t$ Look at this! The velocity itself is a simple linear Bézier curve. It's an interpolation between two vectors: the vector from $P_0$ to $P_1$ (scaled by 2) and the vector from $P_1$ to $P_2$ (scaled by 2). This tells us something incredibly practical. At the start of the curve ($t=0$), the velocity is $2(P_1 - P_0)$. This means the curve starts out moving in the direction from $P_0$ to $P_1$. At the end ($t=1$), the velocity is $2(P_2 - P_1)$, so the curve arrives pointing in the direction from $P_1$ to $P_2$. The first and last legs of the control polygon are tangent to the curve at its endpoints! This gives designers immediate and intuitive control over the curve's start and end directions.

Now for the punchline. What about the acceleration, $B''(t)$? We differentiate again. What we find is almost shockingly simple: $B''(t) = 2(P_2 - P_1) - 2(P_1 - P_0) = 2(P_0 - 2P_1 + P_2)$ The parameter $t$ has vanished! The acceleration vector is ​​constant​​ throughout the entire motion. It doesn't depend on where the particle is on the curve. This is a revelation. What kind of path does an object follow if its acceleration is constant? Think of a ball thrown through the air (ignoring air resistance). It is pulled by the constant acceleration of gravity, and its path is a ​​parabola​​.

This means that every single quadratic Bézier curve, for any non-collinear set of control points, is an arc of a parabola. This beautiful, hidden unity connects a modern tool of digital design with the classical conic sections studied by the ancient Greeks and the laws of motion discovered by Galileo and Newton. This is not just a mathematical curiosity. Knowing the curve is a parabola means we can easily find its properties, like its vertex—the highest or lowest point—by simply finding the moment in time $t$ when the vertical component of the velocity is zero.

From a simple game of sliding beads to the profound realization of a universal parabolic form, the quadratic Bézier curve reveals how simple rules can generate elegant, predictable, and deeply beautiful structures.

Having unraveled the elegant mathematics behind the quadratic Bézier curve, you might be asking yourself, "What is all this good for?" It is a fair question. The answer, it turns out, is wonderfully broad and deeply satisfying. These curves are not merely an abstract curiosity for mathematicians; they are a fundamental tool in the modern world, a kind of universal language for describing smooth shapes. Their power lies in their simplicity and predictability, allowing us to bridge the gap between human intention and digital creation. From the letters you are reading right now to the complex simulations that design our cars and airplanes, the ghost of the Bézier curve is everywhere.

Let's begin our journey in the world of the artist and the designer. Imagine you are using a vector graphics program to draw a logo or illustrate a character. You want to create a graceful, symmetric arch, perhaps for a bridge in a landscape. How do you do it? You don’t need to type in a complicated polynomial equation. Instead, you simply place a starting point and an ending point. Then, you grab a third point—our control point $P_1$—and pull. As you move $P_1$ upwards, a perfect parabolic arch rises to meet it. The height of your arch is directly and intuitively controlled by the height of $P_1$. This direct, tactile manipulation is the magic of Bézier curves. They behave just as our intuition suggests they should.

This intuitive behavior extends to other transformations. Suppose you have designed a beautiful logo. Now, your client wants a larger version for a billboard and a smaller one for a business card. Do you have to redraw it? Not at all. Thanks to a property called affine invariance, you can simply take the three control points that define your curve, scale their positions relative to the origin, and the new curve will be a perfectly scaled version of the original. The entire curve scales, rotates, and translates precisely as its "scaffolding" of control points does. This robust predictability is the bedrock of all modern vector graphics, from Adobe Illustrator to the fonts that render on your screen. Furthermore, if you need to perform a precise operation, such as finding where your curved shape should meet a straight structural element in a design, the underlying parametric equations allow for exact solutions. You can solve for the parameter t where the curve's coordinate matches the line's, enabling perfect alignment in computer-aided design (CAD).

Now, let's put on an engineer's hat. Seldom is a complex object—like the body of a car or the path of a camera in an animated film—described by a single, simple curve. Instead, we build them by piecing together smaller, manageable segments. The challenge is to make the joints seamless. We don't want a jarring "corner" where two curves meet, unless we design it intentionally. We want the transition to be as smooth as a well-made railroad switch. This requires what mathematicians call C^1 continuity, which means that the direction and magnitude of the tangent vector are the same for both curves at the point where they join. For quadratic Bézier curves, this leads to a beautifully simple geometric rule: to join a curve from $P_0$ to $P_2$ via $P_1$ smoothly to a new curve starting at $P_2$, the new control point $Q_1$ must lie on the line extending from $P_1$ through $P_2$, and the distance from $P_2$ to $Q_1$ must equal the distance from $P_1$ to $P_2$. In other words, the junction point $P_2$ must be the exact midpoint of the segment connecting its two neighboring control points, $P_1$ and $Q_1$. This simple constraint allows designers and engineers to construct infinitely complex and perfectly smooth shapes from a humble three-point toolkit.

The algorithmic nature of these curves also lends itself to other powerful manipulations. What if a single curve isn't providing enough detail for you to shape it precisely? You can split it! Using a procedure known as de Casteljau's algorithm, you can break any quadratic Bézier curve at any point $t$ into two new, smaller quadratic Bézier curves that join perfectly and trace the exact same path as the original. This is the digital equivalent of a sculptor taking a large block of clay and cutting it in two to work on each half with more finesse. Conversely, sometimes we need to change the fundamental description of a curve to make it compatible with a different software system. Many professional graphics systems, like PostScript, standardized on cubic Bézier curves (which have four control points). Through a process called degree elevation, a quadratic curve can be perfectly represented as a cubic one by calculating a new set of four control points from the original three. The shape doesn't change one bit; only its mathematical dress does. These robust, algorithmic properties—splitting, joining, and converting—make Bézier curves the ideal computational building block. The same principles also allow us to solve for the precise intersection points between two different curves, a critical task in everything from detecting collisions in a video game to designing interlocking mechanical parts.

The beauty of a deep scientific principle is that it often appears in disguise in seemingly unrelated fields. The quadratic Bézier curve is a prime example. If you look at its defining equation, $B(t) = (1-t)^2 P_0 + 2t(1-t) P_1 + t^2 P_2$, you'll see that the point on the curve is always a weighted average of the three control points. The coefficients, or weights, $((1-t)^2, 2t(1-t), t^2)$, are known in geometry as barycentric coordinates. They tell you the "balance" of influence from each vertex $P_0$, $P_1$, and $P_2$ in determining the final position of $B(t)$. This connects Bézier curves to a much broader mathematical landscape, including topology and the study of abstract geometric spaces called simplices. The curve is, in essence, tracing a specific parabolic path through the space of all possible weighted averages of its control points. This perspective reveals a profound geometric truth: the tangent to the curve at its parametric midpoint ($t=0.5$) is always parallel to the line segment connecting the start and end points ($P_0$ and $P_2$).

Perhaps the most surprising and powerful connection is found in the field of computational physics and engineering. When engineers simulate the stress on a metal plate with a curved hole, or the airflow over a wing, they use a technique called the Finite Element Method (FEM). They break the object down into a mesh of small, simple "elements." But how do they tell the computer where the curved edges of the object are? They use interpolation, just like a Bézier curve! For a quadratic element, they define a curved edge using three nodes: two at the corners and one in the middle. The mathematical formula they use to define the shape of this edge is identical to the formula for a quadratic Bézier curve. The positions of the three nodes are the control points. This means that the same simple idea we use to draw a nice-looking curve is also used to model the boundaries of physical objects in complex simulations. It’s important to note this method creates a parabolic approximation of the true boundary. If the hole is perfectly circular, the three-node element will represent it as a parabola that passes through the three nodes. For a fine mesh, this approximation is extraordinarily good, but it highlights a fascinating trade-off between mathematical purity and practical utility.

So, we see the journey of an idea. What starts as an intuitive method for a designer to sketch a curve on a screen—a simple, elegant dance of three points—reveals itself to be a cornerstone of computational geometry, a practical tool for engineering complex systems, and a specific manifestation of deep principles in mathematics and physics. The quadratic Bézier curve is a testament to the unity of science and art, a simple form that contains within it a world of complexity and application.