Film to Tape Conversion

by Diana Weynand & Marcus Weise

There are many situations when film is transferred to HDTV, or HDTV to film. Film is often used as an archive medium because the film standard has not changed in generations. When film is used as the originating medium, but the delivery requirement is video, film must go through a conversion process. Converting film to an electronic form is often referred to as telecine, which can also refer to the machine used in the conversion process. In PAL systems, the existing video frame rate matches the 25 fps film rate, and the conversion process is simply a matter of converting from one medium to the other. However, in the NTSC system, where the video frame rate of 30 fps does not match the film rate of 24fps, a different conversion process must take place.

Converting Different Frame Rate Sources

Whether the need for conversion is from film to video or video to video the HDTV 24 fps standard makes video and film more compatible with each other, especially in the United States where film is shot at 24 fps. However, to create 30 images per second in video from a source producing 24 images per second, additional video fields are added. This is done by duplicating images from the 24 fps source. The process of adding additional fields to create additional frames is known as a 2:3 pulldown ystem or sometimes as a 3:2 pulldown system.

The 2:3 or 3:2 pulldown process is used when 24 frames per second progressive scan video, used in image capture, must be converted to 30 fps for either editing or transmission. It is also used when transferring 24 fps film to 30 fps video. When transferring 24 frame per second film or video into a 30 fram per second video system, it is essential to decide which sequence will be used 2:3 or 3:2 and then maintain that sequence.

Using 2:3 and 3:3 Pulldown Sequences

When transferring 24 frames per second in a 2:3 sequence, the system will map four frames from the original source to every five frames, or ten fields, to the transfer source. The first frame that is transferred is referred to as the A frame. It is transferred, or pulled down to two video fields. The second frame is referred as the B frame and is transferred to the next three consecutive video fields. The third film frame, the C frame, is transferred to the next two consecutive video frames. And the four frame, the D frame is transferred to the next three consecutive video frames. (Figure 13.4)

The resulting transfer process yields five video frames for every four original 24 fps source frames. Video frame 1 and video frame 2 are each derived from two separate film frames, A and B. Video frame3, however, is a composite of one field from film frame B and one field from film frame C. Video frame 4 is also a composite composed of one field from film frame C and one field from film frame D. Video frame 5 is composed of two fields, both from film frame D.

When scanning through video images that have been transferred using this process, the video frames that contain two different frames will appear as a double image. When these frames are viewed in motion, however, the double image is not discernable. When editing video that has been transferred from film using this pulldown process, the sequence of frames must be maintained. If the sequence of video frames is broken, for example by editing two combination frames consecutively, the resulting conflict of images will be discernable when displayed in motion.

When the pulldown sequence is changed to 3:2, the resulting frames yield a different result. The five video frames would be AA AB BC CC DD (Figure 13.5). Here the second and third frames are composite, rather than the third and fourth. In the 3:2 sequence there is one clean frame followed by two composite frames followed by two clean frames. Therefore when converting 24 fps to 30fps, it is essential that one of these processes is applied consistently so the order of clen and composite frames will be maintained.

Maintaing Consistency in Pulldown Sequences

Just as the pulldown sequence must remain constant, so should the transfer process. Once the transfer of a scene or section has begun, it is essential the transfer be completed without stopping. If the transfer is stopped and then restarted ina single scene or sequence, the system will assume the first frame again is an A frame and begin the sequence from there. Later, when attempting to edit this sequence together in video, it is very possible that the frame sequence will have been broken and the edited version will contain overlapped images from incompatible frames. The result could be stutter or jump in the images.

During the film or video transfer process, as the footage is digitized into the video system, a computer file can be created that will indicate where the associated composite frames are for each section. The added fields can also be seen and identified by scrolling through the video a field at a time. With a correctly compiled computer file, it is possible for a video system to extract and recombine the overlapping fields to create a clean 24 fps sequence from the 30 fps 2:3 or 3:2 pulldown sequence created in the video transfer. The process will discard the redundant fields used to create the 30 fps video from the original 24 fps material. This process is referred to as inverse telecine. Recreating the original material is complex and can only be dine in an uncompressed video stream.

Excerpted from "How Video Works" by Diana Weynand and Marcus Weise (Published by Focal Press). An easy to understand explanation of the entire world of video. A complete guide from analog video to all the new digital technologies, including HD, compression and encoding. Purchase Now

HOME