Time-Frequency Duality

April 10, 2026 · Luciano Muratore

Having analyzed the Fourier expansion of specific periodic signals, we now explore a powerful symmetry in signal processing: Time-Frequency Duality. This principle reveals that operations performed in the time domain have a direct, mirrored counterpart in the frequency domain.

1. The Duality Principle

The relationship between the time domain and the frequency domain is governed by a remarkable symmetry. If an operation is a convolution in one domain, it becomes a point-wise multiplication in the other.

2. Multiplication in Time

When two signals are multiplied in the time domain, $y[n] = x[n]h[n]$, their corresponding DTFT undergoes a periodic convolution.

Mathematical Derivation:

Starting with the DTFT of $y[n]$: $Y(e^{j\omega}) = \sum_{n=-\infty}^{\infty} x[n]h[n] e^{-j\omega n}$

By substituting the Inverse DTFT of $x[n]$, defined as $x[n] = \frac{1}{2\pi} \int_{-\pi}^{\pi} X(e^{j\theta}) e^{j\theta n} d\theta$, we get: $Y(e^{j\omega}) = \frac{1}{2\pi} \int_{-\pi}^{\pi} X(e^{j\theta}) \left( \sum_{n=-\infty}^{\infty} h[n] e^{-j(\omega - \theta)n} \right) d\theta$

Recognizing the term in the parentheses as the DTFT of $h[n]$ shifted by $\theta$: $Y(e^{j\omega}) = \frac{1}{2\pi} \int_{-\pi}^{\pi} X(e^{j\theta}) H(e^{j(\omega - \theta)}) d\theta$ $Y(e^{j\omega}) = \frac{1}{2\pi} (X * H)(e^{j\omega})$

3. Convolution in Time

Conversely, convolving two signals in the time domain—the fundamental operation of LTI systems—results in simple multiplication in the frequency domain.

Mathematical Derivation:

Let $y[n] = (x * h)[n] = \sum_{k=-\infty}^{\infty} x[k]h[n - k]$. Taking the DTFT: $Y(e^{j\omega}) = \sum_{n=-\infty}^{\infty} \left( \sum_{k=-\infty}^{\infty} x[k]h[n - k] \right) e^{-j\omega n}$

Exchanging the order of summation: $Y(e^{j\omega}) = \sum_{k=-\infty}^{\infty} x[k] \left( \sum_{n=-\infty}^{\infty} h[n - k] e^{-j\omega n} \right)$

By applying the time-shift property of the DTFT to the inner sum: $Y(e^{j\omega}) = \sum_{k=-\infty}^{\infty} x[k] \left( H(e^{j\omega}) e^{-j\omega k} \right)$ $Y(e^{j\omega}) = H(e^{j\omega}) \sum_{k=-\infty}^{\infty} x[k] e^{-j\omega k}$ $Y(e^{j\omega}) = X(e^{j\omega}) H(e^{j\omega})$

4. Summary of Results

These dualities simplify complex signal analysis:

1. Filtering: Designing a filter in the frequency domain is as simple as multiplying the signal spectrum by the filter’s frequency response.
2. Modulation/Windowing: Multiplying a signal by a window or a carrier wave in time is equivalent to spreading or shifting its spectrum via convolution.

Key takeaway: Multiplication and convolution are two sides of the same coin, depending on which domain you are viewing.