Discrete Convolution

Discrete-time convolution is a method of finding the zero-state response of relaxed linear time-invariant (LTI) systems. It is based on the concepts of linearity and time invariance and assumes that the system information is known in terms of its impulse response \(h[n]\).

The superposition principle allows expressing the response to input signal \(x[n]\) as the sum of scaled and shifted versions of the impulse response \(h[n]\). This relation defines the convolution operation, specifically referred to as linear convolution. The mathematical expression used to compute the response \(y[n]\) is known as the convolution sum.

\(\boxed{y[n] = x[n] \ast h[n] = \sum_{k = -\infty}^{\infty} x[k]\,h[n - k]}\)

\(\boxed{y[n] = h[n] \ast x[n] = \sum_{k = -\infty}^{\infty} h[k]\,x[n - k]}\)

As with continuous-time convolution, the order in which the operation is performed does not matter, the arguments \(x[n]\) and \(h[n]\) can be interchanged without altering the result.

When evaluating the convolution sum, keep in mind that \(x[k]\) and \(h[n - k]\) are functions of the summation variable \(k\). The summations frequently involve step functions of the form \(x[n] = u[n - \alpha] \to x[k] = u[k - \alpha]\) and \(h[n] = u[n - \beta] \to h[n - k] = u[n - k - \beta]\). Since \(u[k - \alpha] = 0\) for \(k < \alpha\) and \(u[n - k - \beta] = 0\) for \(k > n - \beta\), these can be used to simplify the lower and upper summation limits to \(k = \alpha\) and \(k = n - \beta\), respectively.

Discrete Convolution Properties

Commutative

In linear time-invariant (LTI) systems, the input \(x[n]\) and impulse response \(h[n]\) can be interchanged mathematically due to the commutative property of discrete convolution.

\(\boxed{y[n] = x[n] \ast h[n] = h[n] \ast x[n]}\)

Associative

This property states that the order of discrete convolution operations can be rearranged without affecting the result.

\(\boxed{y[n] = x_{1}[n] \ast \left(x_{2}[n] \ast x_{3}[n]\right) = \left(x_{1}[n] \ast x_{2}[n]\right) \ast x_{3}[n]}\)

Distributive

Discrete convolution is a linear operation and obeys superposition.

\(\boxed{y[n] = \left(x_{1}[n] + x_{2}[n]\right) \ast h[n] = x_{1}[n] \ast h[n] + x_{2}[n] \ast h[n]}\)

For constant \(c\),

\(\boxed{c \cdot y[n] = c\left(x[n] \ast h[n]\right) = \left(c \cdot x[n]\right) \ast h[n] = x[n] \ast \left(c \cdot h[n]\right)}\)

Multiplicative Identity

The discrete convolution of any signal \(x[n]\) with an impulse \(\delta[n]\) reproduces the signal \(x(t)\).

\(\boxed{y[n] = x[n] \ast \delta[n] = \delta[n] \ast x[n] = x[n]}\)

Time Shifting

Discrete convolution is a time-invariant operation and implies that shifting the input \(x[n]\) or the impulse response \(h[n]\) by \(\alpha\) shifts the output response \(y[n]\) by \(\alpha\).

If \(\displaystyle y[n] = x[n] \ast h[n]\), then

\(\boxed{x[n - \alpha] \ast h[n] = x[n] \ast h[n - \alpha] = y[n - \alpha]}\)

If both \(x[n]\) and \(h[n]\) are shifted,

\(\boxed{x[n - \alpha] \ast h[n - \beta] = y[n - \alpha - \beta]}\)

Summation

The sum of the samples in \(x[n]\), \(h[n]\), and \(y[n]\) are related by

\(\boxed{\sum_{n = -\infty}^{\infty}y[n] = \sum_{n = -\infty}^{\infty}x[n] \ast h[n] = \left(\sum_{n = -\infty}^{\infty}x[n]\right) \left(\sum_{n = -\infty}^{\infty}h[n]\right)}\)

The discrete convolution of a signal \(x[n]\) with an unit step \(u[n]\) is the running sum of the signal \(x[n]\).

\(\boxed{y[n] = x[n] \ast u[n] = \sum_{k = -\infty}^{n} x[k]}\)

Conjugation

The complex conjugate of a discrete convolution involves taking the complex conjugate of each function before performing the convolution.

\(\boxed{\overline{y[n]} = \overline{x[n] \ast h[n]} = \overline{x[n]} \ast \overline{h[n]}}\)

Some Useful Discrete Convolution Results

\(\displaystyle u[n] \ast u[n] = \sum_{k = -\infty}^{\infty} u[k]\,u[n - k] = \sum_{k = 0}^{n} 1\)

\(\boxed{u[n] \ast u[n] = (n + 1)\,u[n] = r[n + 1]}\)

\(\displaystyle \alpha^{n} u[n] \ast \alpha^{n} u[n] = \sum_{k = -\infty}^{\infty} \alpha^{k} \alpha^{n - k} u[k]\,u[n - k] = \alpha^{n}\sum_{k = 0}^{n} 1\)

\(\boxed{\alpha^{n} u[n] \ast \alpha^{n} u[n] = (n + 1)\,\alpha^{n} u[n] = \alpha^{n} r[n + 1]}\)

\(\displaystyle \alpha^{n} u[n] \ast u[n] = \sum_{k = -\infty}^{\infty} \alpha^{k} u[k]\,u[n - k] = \sum_{k = 0}^{n} \alpha^{k}\)

\(\boxed{\alpha^{n} u[n] \ast u[n] = u[n] \ast \alpha^{n} u[n] = \frac{1 - \alpha^{n + 1}}{1 - \alpha} u[n]}\)

Discrete Convolution of Finite-Length Sequences

The convolution of two finite-length sequences \(x[n]\) and \(h[n]\) is also of finite length and is subject to the following rules:

  • The starting index of \(y[n]\) equals the sum of the starting indices of \(x[n]\) and \(h[n]\).

  • The ending index of \(y[n]\) equals the sum of the ending indices of \(x[n]\) and \(h[n]\).

  • The length \(L_y\) of \(y[n]\) is related to the lengths \(L_x\) and \(L_h\) of \(x[n]\) and \(h[n]\) by \(L_y = L_x + L_h - 1\).

    \(\boxed{\mathrm{length}(y[n]) = \mathrm{length}(x[n] \ast h[n]) = \mathrm{length}(x[n]) + \mathrm{length}(h[n]) - 1}\)

Sum-by-Column Method

This method is based on the idea that the convolution \(y[n]\) equals the sum of the (shifted) impulse responses due to each of the impulses that make up the input \(x[n]\). To find the discrete convolution \(y[n]\) of \(x[n]\) and \(h[n]\):

  • Arrange the sequences \(x[n]\) and \(h[n]\) vertically below each other, aligning the starting indices.
  • Create a grid or table with rows representing the sequence elements and columns representing the corresponding indices.
  • For each element in \(x[n]\), multiply it with the corresponding element in \(h[n]\). Place each product in the column corresponding to the sum of the indices of the multiplied elements.
  • Sum the products in each column. Each column corresponds to a specific shift, representing the discrete convolution \(y[n]\) result at that particular shift.
  • The starting index (and the marker location corresponding to index \(n = 0\)) for the discrete convolution \(y[n]\) is found by adding the starting indices of \(x[n]\) and \(h[n]\).
Links to this page
#signal #system