This project is a different approach to speeding up or slowing down video. Instead of spline-based approaches (blending in-between frames using linear interpolation) or more complicated optical flow methods, I tried to approximate each in-between frame as a convex linear combination of every frame of the video, sinc weighted as is customary in signal processing. Treating a video as discrete-time samples of a bandlimited 3-dimensional signal is very mathematically simple and programmatically concise (NumPy's tensordot works wonders) but creates time-domain fringing artifacts as shadows and highlights that flash in and out at regions of the video with high temporal frequency and low spatial frequency.