A technique where a video is scaled much larger before processing, to give filters more information to work on (in theory). Do not do this with interlaced video! You'll smear the field scan lines that deinterlacers (and IVTCers, too) need to work from.