Most previous video watermarking algorithms cannot be supported by real-time processing. Our algorithm proposed the specific embedding method in the spatial domain directly rather than the frequency domain. Also the algorithm supports the robustness from the video attacking skills. In the paper, for example, watermark is inserted immediately into the output frame of Digital Video (DV) camcorder. We select the Y component from the DV signal, and then the watermark information is inserted in all of the Y frames. The watermarked video frames put in the video MPEG encoder. We consider embedding information to the high quality video streams, such as a DVD, HDTV. Our experimental results show the high quality of the video even if compressed. Therefore, the robustness from compression is tested by MPEG-2 of 6Mbits/sec of 720x480 frame size and the invisibility is proved by measurement of PSNR. The results also show the robustness from several video editing methods, such as a cut-and-splice and cut-insert-splice, and video conversions, letterboxing, pan & span, and wide screen of media.