Hyperbolic Functions Calculator

Enter any real number x and instantly compute all 12 hyperbolic and inverse hyperbolic values — complete with domain validation, step-by-step formulas, and a live curve graph.

sinh cosh tanh csch sech coth
arcsinh arccosh arctanh arccsch arcsech arccoth
Share this tool

Enter a Value for x

Any real number • Inverse functions show domain validity

Quick Load:
Computing functions…

Enter a value for x above and click Calculate to see all 12 hyperbolic function values.

What Are Hyperbolic Functions?

Hyperbolic functions are analogues of the trigonometric functions, but defined using the hyperbola x² − y² = 1 rather than the unit circle x² + y² = 1. Just as (cos θ, sin θ) traces the unit circle, the pair (cosh t, sinh t) traces the right branch of the unit hyperbola.

The six hyperbolic functions — sinh, cosh, tanh, csch, sech, and coth — appear across engineering, physics, and applied mathematics. They describe the shapes of hanging cables, the geometry of special relativity, solutions to many differential equations, and signal behaviour in electrical engineering.

Unlike trigonometric functions, which are periodic, hyperbolic functions grow without bound as x increases. The deep link between the two families is Euler's formula extended to imaginary arguments: sinh(x) = −i sin(ix) and cosh(x) = cos(ix), where i = √(−1).

Shape of the Three Primary Hyperbolic Functions

The diagram below shows sinh (teal S-curve), cosh (blue U-curve), and tanh (purple bounded S-curve) over x ∈ [−3, 3]. Dashed lines mark the asymptotes y = ±1 that tanh never crosses.

−3 3 4 −4 0y=1 y=−1 min = 1sinhcoshtanh
sinh(x) — odd, unbounded cosh(x) — even, minimum 1 tanh(x) — bounded (−1, 1)

Exponential decomposition

eˣ = cosh(x) + sinh(x)

cosh is the even part of eˣ, sinh is the odd part. Every eˣ splits into these two halves.

The catenary

y = a · cosh(x / a)

The shape a hanging chain takes under gravity — the cosh U-curve scaled by a constant a.

The Six Hyperbolic Functions — Definitions and Formulas

Every hyperbolic function is expressed in terms of the exponential function eˣ. The two primary functions are sinh (hyperbolic sine) and cosh (hyperbolic cosine); the remaining four are derived from these two.

FunctionDefinition (exponential form)DomainRange
sinh(x)(eˣ − e⁻ˣ) / 2(−∞, ∞)(−∞, ∞)
cosh(x)(eˣ + e⁻ˣ) / 2(−∞, ∞)[1, ∞)
tanh(x)sinh(x) / cosh(x)(−∞, ∞)(−1, 1)
csch(x)1 / sinh(x)x ≠ 0(−∞,0) ∪ (0,∞)
sech(x)1 / cosh(x)(−∞, ∞)(0, 1]
coth(x)cosh(x) / sinh(x)x ≠ 0(−∞,−1) ∪ (1,∞)

Key observations

sinh is an odd function: sinh(−x) = −sinh(x). Its graph passes through the origin and is anti-symmetric.

cosh is an even function: cosh(−x) = cosh(x). It has a minimum of 1 at x = 0 — it can never be zero or negative.

tanh is bounded: tanh always lies strictly between −1 and 1, approaching ±1 asymptotically. This makes it a popular activation function in neural networks.

Inverse Hyperbolic Functions — Definitions and Domains

The inverse hyperbolic functions undo their forward counterparts. Because hyperbolic functions are not all one-to-one over all of ℝ, some inverses have restricted domains. Each inverse has a closed-form expression involving the natural logarithm.

FunctionLogarithmic formDomainRange
arcsinh(x)ln(x + √(x²+1))(−∞, ∞)(−∞, ∞)
arccosh(x)ln(x + √(x²−1))x ≥ 1[0, ∞)
arctanh(x)½ ln((1+x)/(1−x))|x| < 1(−∞, ∞)
arccsch(x)arcsinh(1/x)x ≠ 0x ≠ 0
arcsech(x)arccosh(1/x)0 < x ≤ 1[0, ∞)
arccoth(x)arctanh(1/x)|x| > 1x ≠ 0

Why domains are restricted

The inverse of a function only exists where the original function is one-to-one. arccosh is only defined for x ≥ 1 because cosh is only one-to-one on that half (its minimum is at x = 0, so the function is symmetric and not invertible across its full domain). Similarly, arctanh requires |x| < 1 because tanh maps (−∞, ∞) onto (−1, 1), so only values in that open interval have a well-defined pre-image.

Hyperbolic Identities You Should Know

Hyperbolic identities mirror trigonometric identities in form, but with crucial sign differences. The most important is the Pythagorean identity for hyperbolics — a minus sign where circular trigonometry has a plus sign.

CategoryIdentity
Pythagoreancosh²(x) − sinh²(x) = 1
Pythagorean1 − tanh²(x) = sech²(x)
Double anglesinh(2x) = 2 sinh(x) cosh(x)
Double anglecosh(2x) = cosh²(x) + sinh²(x)
Reciprocalcsch(x) = 1 / sinh(x)
Reciprocalsech(x) = 1 / cosh(x) • coth(x) = 1 / tanh(x)
Derivatived/dx sinh(x) = cosh(x)
Derivatived/dx cosh(x) = sinh(x)
Derivatived/dx tanh(x) = sech²(x)

Osborne's Rule

To convert any trigonometric identity to its hyperbolic equivalent, replace sin → sinh and cos → cosh, then negate every product or implied product of two sinh functions. This is why cos²θ + sin²θ = 1 becomes cosh²x − sinh²x = 1 (the product sinh·sinh picks up a minus sign).

Where Hyperbolic Functions Appear in Engineering and Physics

The Catenary — Hanging Cables

A flexible chain or rope hanging under gravity takes the shape y = a cosh(x/a). This catenary curve appears in suspension bridges, power lines, and the Gateway Arch in St Louis, which was deliberately designed as an inverted catenary.

Special Relativity — Rapidity

In special relativity, the rapidity φ of a moving object is related to velocity v and speed of light c by tanh(φ) = v/c. Rapidities add linearly for collinear boosts, unlike velocities, which is why arctanh(v/c) appears so naturally in relativistic mechanics.

Differential Equations

The general solution to y'' − y = 0 (which arises in heat transfer, wave propagation, and beam deflection) is y = A cosh(x) + B sinh(x). Hyperbolic functions are as fundamental to this equation as sin/cos are to y'' + y = 0.

Electrical Engineering

Transmission line equations involve hyperbolic functions. The voltage and current along a lossy transmission line are expressed using cosh and sinh of the propagation constant times distance. The Langevin function L(x) = coth(x) − 1/x models paramagnetism.

Machine Learning — Activations

tanh is one of the original neural network activation functions. It maps any real input to (−1, 1), centres its output near zero, and has a clean derivative: tanh'(x) = 1 − tanh²(x) = sech²(x). It was widely used in early recurrent neural networks.

Hyperbolic Geometry

In the Poincaré disk model of hyperbolic geometry, distances and angles are computed using inverse hyperbolic functions. Hyperbolic space has constant negative curvature, and its metric naturally involves arcsinh and arctanh, making this calculator directly useful for studying non-Euclidean geometry.

Hyperbolic Functions vs Trigonometric Functions

The parallel between hyperbolic and trigonometric functions is deep but imperfect. Both families satisfy Pythagorean-style identities, both have double-angle formulas, and both arise as solutions to simple second-order differential equations. The differences lie in the sign of one term and the nature of the geometry.

PropertyTrigonometricHyperbolic
Pythagorean identitycos²x + sin²x = 1cosh²x − sinh²x = 1
Period2π (periodic)None (non-periodic)
Range of cos/cosh[−1, 1][1, ∞)
GeometryUnit circle x²+y²=1Unit hyperbola x²−y²=1
d/dx of cos/cosh−sin(x)+sinh(x)
Exponential linkcos x = Re(eⁱˣ)cosh x = (eˣ + e⁻ˣ)/2

Frequently Asked Questions

What is the difference between sinh and sin?

sinh is not a periodic function and is defined using exponentials: sinh(x) = (eˣ − e⁻ˣ)/2. In contrast, sin is defined geometrically using the unit circle and repeats with period 2π. Although sinh(x) and sin(x) share the same Taylor series structure (both are odd functions with the same terms), hyperbolic functions grow without bound as x → ±∞, while trigonometric functions are bounded between −1 and 1.

Why is cosh(x) always ≥ 1?

cosh(x) = (eˣ + e⁻ˣ)/2. By the AM-GM inequality, the arithmetic mean of two positive numbers is always at least their geometric mean. So (eˣ + e⁻ˣ)/2 ≥ √(eˣ · e⁻ˣ) = √1 = 1, with equality only when eˣ = e⁻ˣ, i.e. at x = 0. Since cosh is an even function with a minimum of 1 at the origin, it is always ≥ 1 and can never be zero or negative.

What is the domain of arccosh(x)?

arccosh(x) is only defined for x ≥ 1. This is because cosh(t) ≥ 1 for all real t (it reaches its minimum of 1 at t = 0). To invert cosh, we need an input value in the range of cosh, which is [1, ∞). Values below 1 have no pre-image under cosh, so arccosh(0.5) is undefined. This calculator will display 'Undefined: domain requires x ≥ 1' for such inputs.

How are hyperbolic functions related to the exponential function?

Every hyperbolic function is a direct combination of eˣ and e⁻ˣ: sinh(x) = (eˣ − e⁻ˣ)/2, cosh(x) = (eˣ + e⁻ˣ)/2. So sinh and cosh are simply the odd and even parts of eˣ respectively — that is, eˣ = cosh(x) + sinh(x) and e⁻ˣ = cosh(x) − sinh(x). This exponential definition is the basis for all their properties, including the Pythagorean identity cosh² − sinh² = 1.

Where do hyperbolic functions appear in real life?

They appear more often than most people expect: (1) The catenary — a hanging chain or cable follows y = a cosh(x/a), used in bridge engineering and architecture. (2) Special relativity — the rapidity of a moving object is φ = arctanh(v/c). (3) Signal processing — voltage along a transmission line involves sinh and cosh of the propagation constant. (4) Machine learning — tanh is a standard neural network activation function. (5) Heat transfer and beam deflection — solutions often involve linear combinations of sinh and cosh.

Why does csch(0) not exist?

csch(x) = 1/sinh(x). At x = 0, sinh(0) = 0, so csch(0) = 1/0 which is undefined. As x approaches 0 from the right, sinh(x) approaches 0 from the right, so csch(x) → +∞. From the left, csch(x) → −∞. The function has a vertical asymptote at x = 0, which is why this calculator shows 'Undefined at x = 0' for csch and coth when x = 0.

What is the derivative of tanh(x)?

The derivative of tanh(x) is sech²(x) = 1/cosh²(x). This can be derived using the quotient rule on tanh(x) = sinh(x)/cosh(x): d/dx[tanh x] = (cosh²x − sinh²x)/cosh²x = 1/cosh²x = sech²x, using the Pythagorean identity cosh²x − sinh²x = 1. An equivalent form is 1 − tanh²(x), which is particularly useful in neural network backpropagation.