Reducing the resolution of media data

摘要

Input data at a first resolution are received from a source. The input data are compressed. The input data can be downsampled to generate compressed downsampled data at a reduced resolution. The compressed downsampled data can be used to generate a frame at the reduced resolution. When the frame is needed as a reference for another frame, the compressed downsampled data can be decoded to generate decompressed downsampled data at the reduced resolution. Motion compensation is performed on the decompressed downsampled data at the second resolution to generate motion compensated decompressed downsampled data used as a reference for a second frame. Thus, by performing motion compensation on downsampled data directly, data processing operations such as transcoding can be accomplished quickly and effectively while saving computing resources.