Frame interpolation plays a crucial role in very low bit rate video coding, video format conversion, and slow motion replay. Motion-compensated interpolation is known to be the best solution for frame interpolation of video. In this work, three widely used motion-compensated algorithms for frame interpolation are compared and evaluated in terms of interpolated image quality, computation cost, and flexibility. The first algorithm considers motion occlusion, the second is the one used in the H.26L/JVT video codec, and the third uses bidirectional motion estimation. The results of the comparison and evaluation provide insight and serve as a guide to select an appropriate algorithm for a specific application.