A method of decoding video data, the method including via a motion compensation predictor generating a prediction block of a current prediction unit using a reference index and a motion vector of the current prediction unit; via an entropy decoder, entropy decoding a bitstream to generate a quantized coefficient sequence; via an inverse quantizer/inverse transformer, generating a quantized block using the quantized coefficient sequence and inversely quantizing the quantized block to generate a transformed block using a quantization parameter, and inversely transforming the transformed block to generate a residual block, wherein the inverse quantizer/inverse transformer generates the quantization block by selecting two effective quantization parameters that are available and exist among left, upper, and previous quantization parameters according to an order of priority levels set for the left, upper, and previous quantization parameters and using an average of the two effective quantization parameters; and via an adder, generating a reconstructed block using the prediction block and the residual block, wherein the motion vector is derived using a motion vector predictor which is set equal to one of an effective spatial motion vector candidate and an effective temporal motion vector candidate.