A video recording/reproducing apparatus having an encoder for encoding a video signal, and a recording unit for recording encoded information from the encoder on a digital recording medium, the encoder comprising a memory for storing at least first and second reference pictures, a prediction circuit for forming a prediction signal for an input signal, using a first reference signal from the first reference picture designated by a first motion vector, and a second reference signal from the second reference picture designated by a second motion vector, a first motion vector detector for determining a third motion vector from a point of intersection of the first motion vector and the second motion vector to a point in the second reference picture where the first motion vector intersects the second reference picture, a second motion vector detector for determining a differential motion vector by subtracting the third motion vector from the second motion vector, and an encoding circuit for encoding the first motion vector and the differential motion vector to obtain encoded first motion vector data and encoded differential motion vector data, and for encoding a difference signal between the prediction signal and the input signal to obtain encoded difference data.