Video coding and decoding interprediction using differing spatial resolutions
A video encoder is provided which generates a bitstream for decoding by a compliant decoder having an interprediction mode for generating a predictor from a reference picture for a subsequent picture having a spatial resolution different from the reference picture. The encoder provides a bitstream in which the decoder can scale the spatial resolution of a block of pixels of the reference picture using a scale factor signaled in the encoded bitstream and using a filter chosen for the particular block of the reference picture and not the reference picture as a whole to form a predictor.
1. A video encoder configured to:
encode a bitstream for decoding by a decoder, the decoder having an inter-prediction a mode for generating a predictor from a reference picture for a subsequent picture having a spatial resolution different from the reference picture, the decoder configured to:
receive an encoded bitstream including a first picture and a subsequent picture and a header associated with the first picture but not the subsequent picture, the subsequent picture having a different spatial resolution than the first picture, the header associated with the first picture not including parameters for utilizing the mode;
decode the first picture in response to the header and not utilizing the mode;
retain the decoded first picture as a reference picture for the subsequent picture;
scale the spatial resolution of a block of pixels of the reference picture using a scale factor signaled in the encoded bitstream and using a filter chosen for the particular block of the reference picture and not the reference picture as a whole to form a predictor; and
decode a block of pixels of the subsequent picture by adding residual pixel values to the predictor.
2. A decoder having an inter-prediction a mode for generating a predictor from a reference picture for a subsequent picture having a spatial resolution different from the reference picture, the decoder configured to:
receive an encoded bitstream including a first picture and a subsequent picture and a header associated with the first picture but not the subsequent picture, the subsequent picture having a different spatial resolution than the first picture, the header associated with the first picture not including parameters for utilizing the mode;
decode the first picture in response to the header and not utilizing the mode and determine from information in the bitstream that the first picture is available for use as a reference picture for the subsequent picture,
scale the spatial resolution of a block of pixels of the first picture using a scale factor signaled in the encoded bitstream and using a filter chosen for the particular block of the reference picture and not the reference picture as a whole to form a predictor; and
decode a block of pixels of the subsequent picture using the predictor.
3. A non-transitory computer readable media storing an encoded bitstream thereon and being decodable by a decoder having an inter-prediction a mode for generating a predictor from a reference picture for a subsequent picture having a spatial resolution different from the reference picture, the encoded bitstream including a first picture and a subsequent picture and a header associated with the first picture but not the subsequent picture, the subsequent picture having a different spatial resolution than the first picture, the header associated with the first picture not including parameters for utilizing the mode, the decoder being configured to:
receive the encoded bitstream;
decode the first picture in response to the header and not utilizing the mode and determine from information in the bitstream that the first picture is available for use as a reference picture for the subsequent picture,
scale the spatial resolution of a block of pixels of the first picture using a scale factor signaled in the encoded bitstream and using a filter chosen for the particular block of the reference picture and not the reference picture as a whole to form a predictor; and
decode a block of pixels of the subsequent picture using the predictor.