Exponential Fourier Series

The exponential Fourier series represents any periodic signal as a sum of complex exponentials, each spinning at an integer multiple of the fundamental frequency. This compact form powers spectrum analyzers, audio codecs, and every digital communication receiver built today.

Core Concept

Every periodic signal repeats with period T. The exponential Fourier series says you can build that signal by adding together sinusoids whose frequencies are 0, f0, 2f0, 3f0, and so on, where f0 = 1/T. Instead of using separate sine and cosine terms, you use complex exponentials of the form e^(jkω0t), which combine both.

The complex coefficient c_k tells you two things at once: the amplitude |c_k| says how much of frequency kf0 is present, and the angle angle(c_k) tells you its phase. The two-sided spectrum plots c_k for both positive and negative k. For a real signal, c_{-k} = c_k*, so the spectrum is conjugate symmetric.

This form is mathematically cleaner than the trigonometric series. Differentiation, integration, and convolution all become algebraic operations on the c_k coefficients. The exponential form is the direct ancestor of the Discrete Fourier Transform used in every FFT algorithm.

Key Formula

The synthesis equation reconstructs x(t) from its coefficients. The analysis equation extracts each coefficient from x(t). The fundamental angular frequency is ω0 = 2π/T radians per second.

Synthesis: x(t) = Σ c_k · e^(jkω0t), summed over k from -∞ to +∞. Analysis: c_k = (1/T) ∫ x(t) · e^(-jkω0t) dt, integrated over one period. Here k is any integer, T is the period in seconds, and c_k is a complex number with units matching x(t).

Given: x(t) = cos(2πt), period T = 1 s, ω0 = 2π rad/s

Formula: c_k = (1/T) ∫₀ᵀ x(t) e^(-jkω0t) dt

Step 1: Write cos(2πt) = (1/2)e^(j2πt) + (1/2)e^(-j2πt)
         using Euler's formula.

Step 2: Substitute into synthesis form.
         x(t) = (1/2)e^(j·1·ω0·t) + (1/2)e^(j·(-1)·ω0·t)

Step 3: Match term by term with Σ c_k e^(jkω0t).
         c_1  = 1/2
         c_-1 = 1/2
         c_k  = 0 for all other k

Final Answer:
  c_1 = c_-1 = 0.5, all other c_k = 0
  |c_k| has two lines of height 0.5 at k = ±1

Exam Tip: The relation between exponential and trigonometric coefficients is a_0 = c_0, a_k = 2·Re(c_k), b_k = -2·Im(c_k) for k ≥ 1. For a real even signal all c_k are real. For a real odd signal all c_k are purely imaginary. Many GATE problems test whether you can jump between the two representations quickly without re-integrating.

Properties Summary

• Linearity: if x(t) has coefficients c_k and y(t) has d_k, then ax(t)+by(t) has coefficients a·c_k + b·d_k.
• Time shift: shifting x(t) by t0 multiplies each coefficient by e^(-jkω0t0), changing phase but not magnitude.
• Frequency shift: multiplying x(t) by e^(jmω0t) shifts the index, giving coefficient c_{k-m}.
• Conjugate symmetry: for real x(t), c_{-k} = c_k*, so |c_{-k}| = |c_k| and phase is antisymmetric.
• Parseval theorem: average power P = Σ |c_k|², summed over all k from -∞ to +∞.
• Differentiation: the coefficients of dx/dt are jkω0·c_k, turning calculus into multiplication.
• Convolution: if y(t) = x(t)*h(t) over period T, the coefficients of y are T·c_k·d_k.

Quick Revision

• Period T → fundamental frequency f0 = 1/T → fundamental angular frequency ω0 = 2πf0.
• c_k is complex; its magnitude gives the amplitude spectrum, its angle gives the phase spectrum.
• The DC component is c_0 = (1/T)∫x(t)dt, the average value of x(t) over one period.
• Real signals have conjugate-symmetric spectra: |c_k| is even in k and angle(c_k) is odd in k.
• Parseval theorem connects time-domain average power to sum of squared spectral magnitudes.
• Trigonometric form uses a_k and b_k; exponential form uses c_k. Both carry the same information.
• The number of nonzero c_k terms equals the number of sinusoidal components in the signal.
• Exam trap: forgetting the 1/T factor in the analysis integral, or integrating over 0 to 2T instead of exactly one period T.