BOB and Weave
There are two simple ways to turn interlaced video back into deinterlaced video; the “weave” method bluntly combines the two half images back into one image. With still images this is not a problem, but when using this method with moving images, like in our example, the video will get distorted because of the time difference, as shown in the image on the previous page. The method most commonly used for deinterlacing is called BOB. This method basically doubles the lines in an image, so every image contains 50 lines but is half the original resolution. Because of this BOB will produce jagged lines.
A bad deinterlacer will use either BOB or Weave, the better ones use a combination of the two methods. It will check if the image is a moving one or a still, and use the method both suitable for that frame. Even better deinterlacers are able to use a combination of the two methods in one frame, so different parts of the image are rendered through different methods. The best possible result is achieved with the so called “motion adaptive deinterlacing”. This method analyses every image extensively and uses the most suitable method on based on the movement patterns in the image and then produces the best possible image on the screen. This method requires a lot of processing power, which makes the newer 3D chips very suitable for the task. Even with a powerful GPU there are still differences in the algorithms used, as there is no optimal way of analyzing the movement of images.
As mentioned before deinterlacing is extremely important, as TV-signals are still transmitted via the PAL or NTSC standard. A media PC with a graphics card that sports a decent deinterlacer will therefore produce a better quality when watching TV. Besides a lot of DVD discs still contain interlaced material, like TV-series, and even the highest HDTV-standard (1920x1080) can still contain interlaced material.