The main difference between RGB (and component) and S-video is color bandwidth. For S-video the color difference signals are bandwidth limitted to 3.57MHz (~455 pixels per line, ~375 pixels onscreen).
A line doubler, on the other hand, attempts to de-interlace the video. A good line doubler will also perform inverse telecine for video which has been transferred from film.
So RGB will give better color, but a line doubler will give a better image.
Oh, and not all DVDs are non-interlaced. There are some truely horrid examples of bad image encoding out there which look fine on an interlaced TV but will drive progressive output DVD players nuts.