Why doesn’t the curve always pass through the middle control points?

Only the endpoints are interpolated (for a standard Bézier). Interior controls attract and shape the curve but generally do not lie on it; they define the Bernstein blend, not a polyline through every vertex.

What is special about de Casteljau versus the Bernstein formula?

Both define the same curve. de Casteljau is numerically stable and geometric: each step is a linear interpolation you can draw, which makes the construction intuitive and robust.

What happens if I place all control points on a straight line?

The Bézier curve collapses to that line segment (still parameterized by t). The scaffolding interpolations stay on the line, illustrating that Bézier evaluation preserves affine combinations.

Bézier & de Casteljau

Drag control points; live recursive linear-interpolation scaffolding evaluates B(t).

About this model

A Bézier curve of degree n is determined by control points P₀…Pₙ and evaluated with the Bernstein basis, or equivalently by de Casteljau’s algorithm: recursive linear interpolation B(t) built from nested segments (1−t)A + tB. This simulator lets you drag control points while the scaffolding of intermediate interpolated points is drawn live, showing how the curve is carved from the control polygon. The parameter t ∈ [0,1] traces from P₀ to Pₙ; the curve stays in the convex hull of the controls. The page focuses on evaluation geometry — not CAGD continuity between spans, rational NURBS weights, or curve fitting. Move points and scrub t to see de Casteljau’s pyramid construct B(t).

Who it's for: Computer graphics, CAGD, computational geometry, and intermediate programming / math visualization courses.

Key terms

Bézier curve
de Casteljau algorithm
Control points
Bernstein polynomials
Linear interpolation
Convex hull

How it works

Bézier curves are everywhere in graphics, fonts and CAD. The de Casteljau algorithm evaluates B(t) by repeated linear interpolation: take each pair of adjacent control points, slide a fraction *t* along the segment, and you get a new — shorter — control polygon. Iterate until a single point remains. It is numerically stable, geometric, and reveals beautiful scaffolding — drag any Pᵢ to reshape the curve in real time.

Key equations

B(t) = Σᵢ C(n,i) (1−t)ⁿ⁻ⁱ tⁱ Pᵢ

de Casteljau: Pᵢ⁾ʳ⁾ = (1−t)Pᵢ⁾ʳ⁻¹⁾ + t Pᵢ₊₁⁾ʳ⁻¹⁾

B(t) = P₀⁾ⁿ⁾, cost: O(n²)

Frequently asked questions

Why doesn’t the curve always pass through the middle control points?: Only the endpoints are interpolated (for a standard Bézier). Interior controls attract and shape the curve but generally do not lie on it; they define the Bernstein blend, not a polyline through every vertex.
What is special about de Casteljau versus the Bernstein formula?: Both define the same curve. de Casteljau is numerically stable and geometric: each step is a linear interpolation you can draw, which makes the construction intuitive and robust.
What happens if I place all control points on a straight line?: The Bézier curve collapses to that line segment (still parameterized by t). The scaffolding interpolations stay on the line, illustrating that Bézier evaluation preserves affine combinations.

Continue with similar topics in this category — or all 85 in Math Visualization.

View category →

NewSchool

Convex Hull (Graham & QuickHull)

Launch Simulator

Click to add points, drag to move; Graham scan with step playback or QuickHull divide-by-farthest; compare vertex sets.

NewSchool

Linear Regression: OLS, Ridge, Lasso & R²

Launch Simulator

Click/drag scatter points; fit y = β₀ + β₁x with OLS, Ridge (L2 on slope), or Lasso (L1 on slope). Spike Δy on the largest |x| point to see outlier sensitivity; compare SSE and R².

NewSchool

K-Means Clustering (Lloyd)

Launch Simulator

Click to add points, choose k, randomize centroids, then step Lloyd iterations (assign to nearest centroid, update means). Optional Gaussian-mixture demo; watch within-cluster SSE decrease.

FeaturedSchool

Trigonometry Circle

Launch Simulator

Unit circle with live sin, cos, tan values as you drag.

NewUniversity / research

Mandelbrot Deep Zoom

Launch Simulator

Drag/wheel deep zoom into the Mandelbrot set with smooth continuous coloring and named landmarks.

NewUniversity / research

Lagrange vs Cubic Spline

Launch Simulator

Click knots: Lagrange polynomial vs natural cubic spline; Runge preset shows edge oscillations.

Bézier & de Casteljau

About this model

How it works

Key equations

Frequently asked questions

Related simulators

Convex Hull (Graham & QuickHull)

Linear Regression: OLS, Ridge, Lasso & R²

K-Means Clustering (Lloyd)

Trigonometry Circle

Mandelbrot Deep Zoom

Lagrange vs Cubic Spline

Bézier & de Casteljau

About this model

How it works

Key equations

Frequently asked questions