For decades, cubic Béziers have been the backbone of 2D vector graphics. They're everywhere — in fonts, illustrations, UI design, animation. Despite their ubiquity, they have well-known limitations: controlling curvature smoothly is difficult, and representing elegant curves requires careful manipulation of control handles.
Raph Levien, a researcher who has spent years exploring alternatives, has proposed a new curve family he calls the hyperbezier — and it might be the first serious candidate to supplant cubic Béziers in interactive design.
The Problem with Béziers
Cubic Béziers are versatile. That's their strength and their weakness. The same mathematical form can represent gentle curves, sharp corners, and everything in between. But this versatility comes at a cost: it's hard to control curvature.
When a designer manipulates Bézier control handles, they're indirectly controlling curvature. Small changes in handle position can produce dramatic changes in the curve's shape. The relationship between the controls and the resulting curve is not intuitive — it requires training and practice.
Alternative curve families like Euler spirals (where curvature changes linearly with arc length) produce beautifully smooth curves, but they fail at representing regions of high curvature variation — exactly where cubic Béziers excel.
The Hyperbezier
The hyperbezier is defined by a Cesàro equation — curvature as a function of arc length:
κ(s) = (as + b) / (cs² + ds + 1)^1.5
This deceptively simple formula produces a curve family that:
- Closely approximates cubic Béziers at low deflection angles, making it a natural replacement
- Has smoother curvature variation — the primary complaint about Béziers
- Contains valuable analytic curves within the family
- Handles both smooth and high-tension regions — the key advantage over Euler spirals
The key insight is that the denominator (cs² + ds + 1)^1.5 creates a "softening" effect on the curvature, allowing it to peak and then decrease smoothly — something that's difficult to achieve with polynomial-based curves like Béziers.
Why This Matters
If you're not a font designer or a vector graphics enthusiast, you might wonder why curve mathematics matters. Here's why:
Every piece of text you read is defined by curves. Font designers spend enormous effort tuning Bézier control points to produce letterforms that look right at any size. A curve family that's easier to control means better fonts, faster.
UI design tools (Figma, Sketch, Illustrator) all use cubic Béziers as their primary curve primitive. Designers spend significant time fighting the controls to get the curves they want. A better curve primitive could fundamentally change how design tools work.
Animation paths use curves to define motion. Smoother curvature means smoother motion — and fewer manual keyframes to achieve it.
The Interactive Demo
Levien has created an interactive tester that maps cubic Bézier control points to corresponding hyperbezier parameters. You can drag the Bézier controls and see both curves overlaid, along with curvature combs that visualize the curvature distribution.
The result is striking: where the Bézier has irregular curvature (visible as bumps in the curvature comb), the hyperbezier has smooth, monotonic curvature. The visual difference is subtle, but the perceptual difference is significant — curves with smooth curvature simply look better.
The Road to Adoption
Replacing cubic Béziers is not a small undertaking. They're embedded in:
- Every major font format (OpenType, TrueType)
- SVG, PDF, PostScript
- Every design tool
- Every graphics API
- Every font rendering engine
The hyperbezier is not going to replace Béziers overnight. But the path to adoption could be:
- Design tools add hyperbezier as an alternative curve primitive
- Font designers experiment with it for new typefaces
- Open font formats (like the proposed GlooNF format) include it
- Graphics libraries add support
- Eventually, it becomes a standard option alongside Béziers
The key advantage is that the hyperbezier can approximate Béziers closely enough that conversion between the two is practical. Existing Bézier-based designs can be converted to hyperbeziers with minimal visual change, and designers can then refine using the superior curvature control.
The Deeper Lesson
The hyperbezier is a reminder that the tools we use are not inevitable. Cubic Béziers were chosen for practical reasons in the 1960s — they were computationally tractable, mathematically well-understood, and good enough. But "good enough" in 1962 is not the same as "good enough" in 2026.
Every fundamental computational primitive — from character encodings to instruction set architectures to curve representations — was chosen under constraints that no longer apply. The question is not whether better alternatives exist (they do), but whether the ecosystem can overcome the inertia of existing standards.
Raph Levien's hyperbezier is a serious attempt to answer that question for one of the most fundamental primitives in visual computing. Whether it succeeds or not, the work is a model for how to challenge established standards: not with vague complaints, but with a rigorously designed alternative that's better on every dimension that matters.
Read Raph Levien's full technical writeup at linebender.org. The interactive demos are worth playing with — you'll never look at a font the same way again.
Top comments (0)