Sectioned Convolution /
Block Convolution
Dr. J. Naga Vishnu Vardhan
Professor & HoD, ECE, BVRIT HYDERABAD
Introduction
In spite of computational advantages of performing convolution using
DFT, but there are some difficulties.
For example if x(n)is very long,
It takes significant amount of time to compute DFT
Very long processing delays
For instance, high fidelity digital audio requires a data rate of about
5 Mbytes/min, while digital video requires about 500 Mbytes/min.
With data rates this high, it is common for computers to have
insufficient memory to simultaneously hold the entire signal to be
processed. There are also systems that process segment-by-segment
because they operate in real time
Introduction (contd..)
Henceforth, solution to these problems is to use SECTIONED /
BLOCK CONVOLUTION that involves segmenting the signal to be
filtered x(n) into sections.
Each section is then filtered with the filter having impulse response
h(n).
The filtered sections are pieced together to form output sequence
y(n). To perform these, there are two block convolution techniques
namely
Overlap-add method
Overlap- Save method
Overlap Add Method
Consider an input signal x(n), impulse response of the filter/
system h(n) and output response y(n), then the convolution sum
is defined as
Overlap Add Method
Overlap Save Method