Kurzweil:Convolution

From Sonikmatter

A strict definition of Convolution is the multiplication of the Fourier Transforms of two signals. Essentially the impulse response of one signal is applied to the other signal. With Convolution it doesn't actually matter which signal is considered to be the impulse response, and which is considered to be the signal.

In effect Convolution takes the impulse response and reverses it. For each sample period the reversed impulse response is applied to the signal. Graphically this can be represented by drawing the impulse response on a piece of paper, reversing it, and moving it over the signal. The resulting waveform is the magnitude of the reversed impulse response at all of it's points multiplied by signal at each of the points corresponding to the impulse response

For a more in depth read have a look at Bracewell R N, The Fourier Transform and it's Applications, McGraww Hill, 1986 ISBN 0070070156

The Puppeteer 01:16, 2 Sep 2005 (EDT)

Main Page :

Documentation :

FAQ :

DSP Blocks :

KDFX :

Algorithms