Matrix Norm Calculator (1-norm, ∞-norm, Frobenius and spectral 2-norm)

Understanding the matrix norm calculator — concept and principle

This calculator computes the common norms of a matrix: the 1-norm (largest absolute column sum), the ∞-norm (largest absolute row sum), the Frobenius norm, the spectral 2-norm (largest singular value) and the max norm. It also checks the inequality ‖A‖₂² ≤ ‖A‖₁‖A‖∞. Use it in numerical analysis, error estimates and condition numbers.

The 1-norm takes the largest column sum, the ∞-norm the largest row sum.

How it is calculated

‖A‖₁ = maxⱼ Σᵢ |a_ij| (columns)

‖A‖∞ = maxᵢ Σⱼ |a_ij| (rows)

‖A‖_F = √(Σ a_ij²)

‖A‖₂ = largest singular value σ₁

Max norm = largest |a_ij|

Check: ‖A‖₂² ≤ ‖A‖₁ × ‖A‖∞

The principle

A norm measures the size of a matrix. The 1-, 2- and ∞-norms measure the largest stretch the matrix can apply to a vector, each using a different way of measuring vector length; this is why they are called operator norms. The Frobenius norm treats the matrix as one long vector of entries. All norms agree on whether a matrix is small or large up to factors depending on its size, so the cheap 1- and ∞-norms are useful quick estimates of the more expensive 2-norm.

Worked example

Default input: A = [1 −2; −3 4].

Column sums: |1| + |−3| = 4 and |−2| + |4| = 6, so ‖A‖₁ = 6.
Row sums: |1| + |−2| = 3 and |−3| + |4| = 7, so ‖A‖∞ = 7.
Frobenius: √(1 + 4 + 9 + 16) = √30 = 5.47723.
Spectral norm (largest singular value) = 5.46499; max norm = 4.
Check: ‖A‖₁ × ‖A‖∞ = 6 × 7 = 42 ≥ 5.46499² ≈ 29.87.

How to read the result

The spectral norm is the largest factor by which A can lengthen any vector. Here it is close to the Frobenius norm because A is nearly rank one: one singular value dominates. Norms are combined with the inverse's norm to give the condition number, which measures how sensitive Ax = b is to errors.

Common mistakes

Mixing up rows and columns for the 1- and ∞-norms.
Forgetting absolute values.
Treating the Frobenius norm as the 2-norm; they differ in general.
Using the max norm as an operator norm; it does not bound Ax in the same way.

Keep in mind: Singular values are computed numerically; very large or ill-conditioned matrices can show small rounding differences.

Frequently asked questions

What is the 1-norm of a matrix?

The largest sum of absolute values in any column.

What is the spectral norm?

The largest singular value, the largest factor by which A stretches a vector.

How is the Frobenius norm computed?

The square root of the sum of all squared entries.

Sources: E. W. Weisstein, "Matrix Norm", Wolfram MathWorld · J. Demmel, UC Berkeley Math 221 lecture notes, Lecture 3 (2011)

Checked 9 Nov 2026. Wolfram MathWorld, “Matrix Norm” (maximum absolute column sum norm ‖A‖₁, spectral norm ‖A‖₂ = (maximum eigenvalue of AᴴA)^½, maximum absolute row sum norm ‖A‖∞, inequality ‖A‖₂² ≤ ‖A‖₁‖A‖∞); J. Demmel, UC Berkeley Math 221 lecture notes, Lecture 3 (Fall 2011): Frobenius norm (Σ|A(ij)|²)^½, max norm, operator norm, ‖A‖₂ = σ(1).

Your numbers