De-interlace filters remove interlacing from video to facilitate video filtering tasks. Operations like repositioning, rotating and blurring cause unwanted artifacts on interlaced footage, so removing the interlacing before performing these steps is essential.
There are four main methods of de-interlacing. The first is called field doubling or field stripping where one field is removed and the other doubled to take its place. This obviously results in only half the original vertical resolution.
The second method is field blending, where each field is averaged against its neighboring field. This retains vertical detail at the expense of some softening of the image.
A third method uses motion estimation to warp the two fields into a single moment in time. This can produce accurate results but will produce errors in complex or grainy footage.
A fourth and most desirable method is field separation or frame doubling. In this process, each field is extracted and doubled–just like field doubling–but the first fields are placed in the editing timeline at the frame marks and the second set of fields are placed halfway between each frame in time. When all the editing operations are complete, the two sets of fields are then merged back together to create a final interlaced output.