Video and TV Broadcasting Standard Conversion

While most of the technical differences between the various video and TV broadcasting standards (PAL, NTSC, SECAM) are easily overcome by today's software or hardware chips, there are three things that remain troublesome and where the ultimately perfect solution has yet to be invented: frame rate conversion, frame size conversion, and sound (re)synchronization.

On this page we will have a look at the basic principles in the techniques used to convert a PAL video to a NTSC video and vice versa. We will see that there often is a trade off between conversion speed and conversion quality.

Video Standard Conversion the Primitive Way

Converting NTSC videos to PAL involves reducing the number of frames per second (fps) from 30 to 25. A cheap way to do this (time- and software wise) is to remove every sixth frame. This makes it possible to watch the video on a PAL system, but the video will have a visible skip every sixth of a second. If we are talking DV tapes, the height of each frame has to be expanded from 480 pixels to 576 pixels, which often is done either by adding 96 pixels of blank space or stretching the image vertically.

Converting from PAL to NTSC is the same thing in reverse. The height of each frame is either downscaled or cropped 96 pixels, and five frames added to every second. This will make the video seem to pause 1/30th of a second every sixth of a second.

Another cheap method is to simply adjust the frame rate. Converting from NTSC to PAL will speed up the video by 4%, which will make everyone talk a little faster and with a higher pitch. Converting from PAL to NTSC slows down everything.

Video Standard Conversion the Advanced Way

The better, but more intensive, way of converting between the broadcasting standards is to sort of blend the frames to match the new frame rate. The higher end, professional conversion software averages the information from one frame to the next by storing frames digitally in computer memory, comparing the stored frames, and adjusting each frame to produce natural motion. The better software can detect and differentiate between moving and stationary objects to further improve motion quality. The frame size still has to be cropped or stretched respectively, but the result will be much more pleasing to the eye.

These techniques are computationally very expensive and demanding processes. On slow systems, it may take 15-20 hours to convert just one hour of movie (if it works at all). At Video we only do broadcasting standard conversion the professional way. On our systems, it usually takes 6-8 hours to convert one hour of movie.