Although much research have investigated transparent scrambling video techniques, an issue of which transcoding relates to downsizing scrambled video without unscrambling has hardly been solved in single-layer MPEG-2 coded video content. In the paper, a novel algorithm based on approximation for compressed domain linear operation and using a mix of linear transformation of coefficient values in DCT for scrambling video is proposed for downsizing the scrambled video by a factor of two. Experimental results show that the proposed algorithm can reduce computational complexity and efficiently degrade video quality. Compared to the existing scrambling video methods, this algorithm not only can scramble video that is compliant to block-coded format but also can realize downsizing the scrambled video in the compressed domain by a factor of two without unscrambling and impact on the performance of the compression.