the Bromwich inversion contour
/ BROM-witch /
When you take a Laplace transform you trade a function of time, f(t), for a function of a complex frequency, F(s) — a one-way summary that is often far easier to manipulate. The Bromwich contour is the road back. It is the specific path in the complex s-plane along which you integrate F(s) to recover the original f(t). Without it the Laplace transform would be a trapdoor you could fall through but never climb out of.
Concretely, the inverse Laplace transform is f(t) = (1 / (2 pi i)) times (integral along the Bromwich contour of e^(s t) F(s) ds), where the contour is the vertical line s = c + i y, with y running from minus infinity to plus infinity, and c chosen so the whole line lies to the right of every singularity of F(s). That vertical line is the Bromwich contour. In practice you almost never integrate the straight line directly; instead you close it with a large semicircle to the left and use the residue theorem, so the answer becomes the sum of residues of e^(s t) F(s) at the poles of F — and Jordan's lemma is what guarantees the big arc contributes nothing as its radius grows (for t > 0).
The contour ties signal analysis and differential equations directly to the residue calculus. Each pole of F(s) becomes an exponential mode e^(s_k t) in f(t): poles in the left half-plane give decaying modes, poles on the imaginary axis give steady oscillations, and poles in the right half-plane give blow-up — which is exactly how the geometry of poles encodes stability. The honest subtleties: the constant c must beat the rightmost singularity (its real part is the abscissa of convergence), and if F has branch points rather than poles you cannot just close the contour — you must wrap it around a branch cut instead, which produces a continuous spectrum rather than discrete modes.
To invert F(s) = 1 / (s^2 + omega^2) place the Bromwich line to the right of the poles s = plus or minus i omega. Closing left and summing residues of e^(s t) / (s^2 + omega^2) at those two poles gives f(t) = (1/omega) sin(omega t) for t > 0 — the conjugate pair of imaginary-axis poles producing a pure oscillation, exactly as the pole geometry predicts.
A vertical line right of all poles, closed to the left; residues give the time function.
For t < 0 you must close the contour to the right instead, where (if F is analytic there) you enclose no poles and get f(t) = 0 — which correctly reflects that a Laplace transform reconstructs a causal, one-sided signal.