In general terms, convolution is defined for two functions which both can be continuous and then it's essentially an "weighted integral". In 1D case (sound), wh

In general terms, convolution is defined for two functions which both can be continuous and then it's essentially an "weighted integral". In 1D case (sound), when both functions have discrete domain and domain of filter is finite, then it is the same thing as FIR filter. Convolution with continuous domain kernel (eg. sinc()) is useful for example for resampling/scaling.