Convolution Calculator

Discrete convolution is an operation that takes two sequences, denoted as $f$ and $g$, and produces a third sequence representing how the shape of one is modified by the other. Mathematically, the discrete convolution $(f * g)[n]$ is defined as the summation over all integers $m$ of the product $f[m]g[n - m]$. In terms of array indices, if the first sequence has length $N$ and the second has length $M$, the resulting convoluted sequence will have a length of $N + M - 1$. The process essentially involves reversing the second sequence, shifting it across the first sequence point by point, multiplying corresponding overlapping elements, and summing the results.

Let us calculate the discrete convolution of two simple sequences: Sequence A = [1, 2, 3] and Sequence B = [4, 5]. First, we determine the length of the output sequence, which will be $3 + 2 - 1 = 4$ elements. We set up the operation by reversing Sequence B to get [5, 4] and sliding it across Sequence A. For index 0, the only overlap is $1 imes 4 = 4$. For index 1, the overlap products are $(1 imes 5) + (2 imes 4) = 5 + 8 = 13$. For index 2, the products are $(2 imes 5) + (3 imes 4) = 10 + 12 = 22$. For index 3, the final overlap is $3 imes 5 = 15$. Thus, the resulting convoluted sequence is [4, 13, 22, 15].

When working with convolution manually or using digital tools, always verify the indexing and length of your input arrays to avoid truncation errors. Remember to pad your sequences with zeros if you are comparing signals of different lengths or preparing data for Fast Fourier Transform (FFT) algorithms. Additionally, keeping track of the starting time index of your signals ensures that the output peak aligns correctly in time-series or audio processing tasks.