Linear convolution

Linear convolution is an operation used in signal processing to combine two signals and obtain a
resulting signal that re ects how one signal affects the other
y[n]=x[n]∗h[n]=k=−∞∑∞ x[k]⋅h[n−k]

Circular convolution
Circular convolution takes two periodic sequences and computes their convolution in such a way
that the result also has the same period as the input sequences

Key Differences from Linear Convolution:

1. In circular convolution, the sequences are assumed to be periodic, so values outside the
sequence bounds wrap around to the start.
2. Length of Result: For two sequences of length N circular convolution produces a result of
the same length N, unlike linear convolution, which yields a length of 2N−1 for the same
input lengths.

Using the DFT and its inverse, circular convolution can be computed faster than linear convolution,
especially for large data sets

y[n]=x[n]⊛h[n] is given by:

y[n]=IDFT (X[k]⋅H[k])

gilbert transform
This transform essentially shifts the phase of each frequency component in the signal by 90 degrees,
which has various applications in demodulation, communication systems, and analytic signal
construction.

Autocorrelation
De nition: Autocorrelation is a measure of the similarity between a signal and a delayed version of
itself over various interval
Rxx (k)=∑ x(n)⋅x(n+k)
n
fi
fl
 Cross-Correlation
• De nition: Cross-correlation is a measure of similarity between two different signals as a
function of the time lag between them. It evaluates how one signal aligns with another over
various time shifts.
Rxy (k)=∑ x(n)⋅y(n+k)
n

Z transform
X(z)=−∞∑∞ x[n](z)^−n

Inside the Unit Circle: If ∣pi∣<1 all poles pi , the system is stable.
On the Unit Circle: If ∣pi∣=1 all poles pi, the system is marginally stable but not necessarily stable.
Outside the Unit Circle: If any pole lies outside the unit circle∣pi ∣>1, the system is unstable.

Causal Systems: The Z-transformX(z) of a causal system has its ROC outside the outermost pole
(or includes the unit circle if stable).

Relation
Fourier Transform: X(ejω)=X(z)| z=ejω
DFT: X[k]=X(ejω) | ω=2πk/2
Z and dft : X[k]=X(z) |z=ejN2πk

Decimation :
Decimation reduces the sampling rate of a signal by an integer factor, M. This process is commonly
referred to as "down-sampling." Decimation involves two main steps:

Filtering: A low-pass lter is applied to the original signal to remove any high-frequency
components that could cause aliasing. Aliasing occurs when higher frequency components of the
signal "fold" back into the lower frequencies, distorting the down-sampled signal.

Down-sampling: After ltering, every M-th sample is retained, while the others are discarded. This
reduces the sampling rate by a factor of M.

The overall effect is that the signal's bandwidth is reduced and its resolution is lower since fewer
samples represent the original signal. If the original signal was sampled at , the new sampling rate
becomes / M.

Interpolation
Interpolation increases the sampling rate of a signal by an integer factor, L. This process is called
"up-sampling." Interpolation involves:

Up-sampling: The rst step is to insert L−1 zero-valued samples between each pair of original
samples. This increases the sampling rate but introduces spectral images, or copies of the original
spectrum shifted to higher frequencies.
fi
𝑓
𝑠
fi
fi
fi
𝑓
𝑠
 Filtering: A low-pass lter, known as an interpolation lter, is then applied to the up-sampled signal
to remove these spectral images and smooth out the transition between samples.

After interpolation, the signal's new sampling rate is × , where is the original sampling rate.

Discrete Fourier Transform (DFT)

DFT is a mathematical transformation used to analyze the frequency content of a discrete signal. It
converts a sequence of N complex numbers from the time domain to the frequency domain.

[ ] = ∑( =0 to −1) [ ]⋅ ^ (− 2 / ), =0,1,2,…

where:
X[k]: The DFT of the sequence x[n]
N: The total number of samples
k: The frequency index
x[n]: The input sequence in the time domain

Inverse Discrete Fourier

Transform (IDFT)
IDFT is the inverse process of the DFT. It transforms a sequence from the frequency domain back
to the time domain.

[ ] = 1/N ∑(k=0 to −1) [ ]⋅ ^ ( 2 / ), n=0,1,2,…

Fast Fourier Transform (FFT)

FFT is an ef cient algorithm to compute the DFT and IDFT, reducing the computation complexity
from O(N^2) to O(N logN). It is widely used in digital signal processing for its speed and
ef ciency.
𝑋
𝑥
fi
𝑛
𝑘
𝑛
fi
𝑁
𝑁
fi
𝑥
𝑛
𝑋
𝑒
𝑘
𝑒
𝑗
𝜋
𝑗
𝑘
𝜋
𝑛
𝑘
𝑁
𝑛
𝑁
𝑘
𝐿
fi
𝑓
𝑠
𝑓
𝑠
 Plot square
% De ne the coordinates of the vertices of the square
x = [1, 1, 2, 2, 1]; % x-coordinates
y = [1, 2, 2, 1, 1]; % y-coordinates

% Plot the square

plot(x, y, '-o')

% Set the axis limits to make the square appear correctly

axis([0 3 0 3])

% Add labels and title

xlabel('X-axis')
ylabel('Y-axis')
title('Square Plot')

% Display grid
grid on

Plot circle
% De ne the radius and center of the circle
radius = 5;
center_x = 0;
center_y = 0;

% De ne the angle range for the circle

theta = linspace(0, 2*pi, 100);

% Parametric equations for the circle

x = center_x + radius * cos(theta);
y = center_y + radius * sin(theta);

% Plot the circle

gure;
plot(x, y);
axis equal;
title('Circle Plot');
xlabel('X');
ylabel(‘Y');
fi
fi
fi
fi
Twiddle factor

Fourier
IIR

FIR