With the notable exception of JPEG2000, video encoding, mainly H.264 and HEVC, involves grouping a series of  frames into sequences called Group of Pictures (GOP)s. By processing and compressing GOPs instead of individual frames, file sizes and stream bitrates can be significantly reduced. GOPs are comprised of I,P, and B frames. The first frame in a GOP is an intra-frame called an I frame which contains most of the vital information for the following sequence of P frames which are forward predictive and B frames which are both forward and backward (bi-directional) predictive. A B frame estimates changes to the frame based on the previous and following frames. What’s important to keep in mind is that I frames contain more data than P and B frames. Although it’s entirely possible to encode in HEVC or h.264 with only I frames, the compression efficiencies won’t be as great as with using GOPs.