Architecture For An MPEG-PIP-processor

Abstract

To reduce the size of a picture it needs to be lowpass- filtered in horizontal and vertical direction before being sampled down to avoid aliasing. This lowpass-filtering can be combined with the MPEG-decoding-process due to the character of the discrete cosine transform. Equipped with some flexibility an MPEGZ-video- decoder can be used for decoding more than one video stream with reduced image size. It is shown how memory and calculation power can flexibly be used for decoding MPEG-sequences with reduced picture size.