Video and Film Pull-down Issues
Monaco's Chief Colorist John Carlson Clarifies Film to Video Transfer Rate Questions
(Reprinted from www.filmarts.org)
Filmmakers transferring their footage to video often need clarification when the final project is going to output to film. The first area of confusion about pull-down can occur when discussing the transfer rate which refers to the film speed as shot in camera and transferred on the telecine. For a film finish, you would use 24 fps since that is the speed of film projectors. Some transfers are done at 30 frames per second but are always a video finish, i.e. not ending up on film. Your videotape would always conform to the NTSC video time-code standard of 29.97 Non Drop Frame per second or Drop Frame for broadcast finish since the time code would then equal real time.
Another confusing area is the request for time-code and/or keycode windows as being made visible in black strips. This could be mistaken as a request for letterboxing a transfer to show the wide-screen geometry (aspect ratio) of a projected film. Other information from a transfer, including time code, keycode, audio time code, footage and frame, Acmade, audio roll, lab reel, camera ID, video reel and camera date, can be made visible in windows on the video.
Be sure to order your window dubs from a reputable transfer house to ensure that the window dub information matches at the head of each camera roll. Using an edit-controller, the film can easily be rewound back to the head of each camera roll after the keycode reader has caught the code, giving uninterrupted time code with the correct keycode numbers.
Another help would be to ask the transfer house to supply a diskette with a flex file to assist you in digitizing and logging your film clips. Most editing software (FCP) will support this useful tool which contains transfer information available in windows above. The database files are used in non-linear editing systems or in matchback programs. The correlation of time code to keycode is contained within these files, which are used to go from a video edit to film cut. A film frame can be an A, B, C, or D frame when transferred (see information below). Files are based on A frames because only they use one time code increment. A frames are also the start of the 2:3 sequence, which becomes important when digitizing on a film editor. (Note: If your editor is set to field two dominance, and you were only seeing one field of video per frame, you might think that there was 1 frame of A, B, and C and 2 frames of D.)
Pull-Down: The First Essential Thing
Four film frames are equal to five frames of video, but video is further broken down into 2 fields per frame. If we discount High Definition or computer monitors, the fields are interlaced as odd scan lines then even scan lines of the video frame. So now we have 10 fields of video per 4 frames of film. And the reason it’s called 2:3 pull down is because each frame of film is alternately transferred to either 2 or 3 fields of video. If we call the 4 frames of film A, B, C, and D, then A gets 2 fields of video, B gets 3 fields, etc. (See the diagram below for a graphic illustration comparing film frames to video fields).
There are many ways of “skinning this cat”—doing an edit on a non-linear system and coming up with a correct film cut list to keep the edited film and sound track in sync. Here are a few of them:
- Edit with an Avid Film Composer
- Edit initially with any non-linear system, then re-edit on an Avid Film Composer
- Make work prints of the selected takes, cut the work print
- Use the “matchback” software, constantly chasing sync so that you are never off by more than 1 frame
- Use the new Mac software (To us unproven, but theoretically sound)
- Always cut before an A frame—what a pain!
We realize that all of this can be overwhelming, or at the least frustrating. We are happy, however, to answer questions—feel free to contact us.
John Carlson
Monaco Labs, San Francisco
