Is a JVC receiver any good?

In the case of I pictures, the actual image data is then passed through the encoding process described below. P and B pictures are first subjected to a process of "motion estimation", in which the encoder searches for similarities with the previous (and in the case of B pictures, also the next) image in time order. For each macroblock, the encoder searches for a good "reference sample" — a same-sized area in the previous or next image that is most similar to it. A motion vector is encoded to describe the relationship between the current macroblock, and the reference sample it is predicted from. Usually, even after motion compensation is performed, there is still a difference between the macroblock and its reference sample. This difference (or "residual") may be described by the encoding process described below. Note that if the residual is very small - in particular, if it's probably too slight of a difference to see, the encoder may choose to not encode any residual at all in order to keep bitrate low. In this case, the entire macroblock's appearance is described by just two numbers: the two components of its motion vector.