Convolution (convolve) in SciPy

Introduction

Convolution helps combine two sets of data to see how one affects the other over time or space.

To smooth noisy data by blending it with a small averaging window.

To detect patterns like edges in images by combining with a filter.

To analyze how a signal changes when passed through a system.

To mix two audio signals to create effects like echo.

To calculate weighted sums in time series data.

Syntax

SciPy

scipy.signal.convolve(in1, in2, mode='full', method='auto')

in1 and in2 are the input arrays to combine.

mode controls the size of the output: 'full' (default), 'valid', or 'same'.

Examples

Basic convolution of two small lists with default 'full' mode.

SciPy

from scipy.signal import convolve
result = convolve([1, 2, 3], [0, 1, 0.5])

Convolution with output size same as first input.

SciPy

result_same = convolve([1, 2, 3], [0, 1, 0.5], mode='same')

Convolution with output only where inputs fully overlap.

SciPy

result_valid = convolve([1, 2, 3], [0, 1, 0.5], mode='valid')

Sample Program

This example smooths the signal by combining it with a small filter that averages neighbors.

SciPy

from scipy.signal import convolve

# Two simple signals
signal = [1, 2, 3, 4]
filter_ = [0.25, 0.5, 0.25]

# Perform convolution
result = convolve(signal, filter_, mode='same')

print('Signal:', signal)
print('Filter:', filter_)
print('Convolved result:', result)

Important Notes

Convolution flips one input before sliding it over the other.

Use mode='same' to keep output size equal to the first input.

Convolution is useful in many fields like image processing, audio, and time series.

Summary

Convolution combines two data arrays to see how one modifies the other.

Use scipy.signal.convolve with inputs and choose output size by mode.

It helps smooth, detect patterns, or analyze signals in data science.