It’s difficult to move in science or image analysis without tripping over a Fourier transform or two, whether it’s getting a frequency spectrum, performing a convolution or just creating a JPEG image. If you’ve done any computational science, you’ve almost certainly used a “fast Fourier transform” (FFT) to actually do a Fourier transform. In this talk Phil peered behind the curtain of these magical algorithms to see what they do, how they work and how the complex mathematics involved is actually implemented – plus, what are they actually useful for?

You can find the slides for this talk here.