Frame Rate Basics

When transferring film to video, you need to take into account the differences in film and video frame rates. Film is commonly shot at 24 frames per second (fps), although 25 fps is sometimes used when the final project is to be delivered as PAL video (as opposed to the more common technique of just speeding up 24 fps film to 25 fps). Video can have a 29.97 fps rate (NTSC), a 25 fps rate (PAL), or either a 24 fps or 23.98 fps rate (24p), depending on your video standard.

The frame rate of your video (whether you sync the audio during the telecine transfer or not) and the frame rate you want to edit at can determine what you need to do to prepare your clips for editing. You may find it useful to read Determining How to Prepare Source Clips for Editing before you make any decisions about frame rates.

Working with NTSC Video

The original frame rate of NTSC video was exactly 30 fps. When color was added, the rate had to be changed slightly, to the rate of 29.97 fps. The field rate of NTSC video is 59.94 fields per second. NTSC video is often referred to as having a frame rate of 30 fps, and although the difference is not large, it cannot be ignored when transferring film to video (because of its impact on audio synchronization, explained in Synchronizing the Audio with the Video).

Another issue is how to distribute film’s 24 fps among NTSC video’s 29.97 fps.

The most common approach to distributing film’s 24 fps among NTSC video’s 29.97 fps is to perform a 3:2 pull-down (also known as a 2:3:2:3 pull-down). If you alternate recording two fields of one film frame and then three fields of the next, the 24 frames in 1 second of film end up filling the 30 frames in 1 second of video.

Note: The actual NTSC video frame rate is 29.97 fps. The film frame rate is modified to 23.98 fps in order to create the 3:2 pattern.

Figure. Diagram showing the pattern of fields added when a 3:2 pull-down is used to convert 23.98 fps film to 29.97 fps NTSC video.

As shown above, the 3:2 pattern (actually a 2:3:2:3 pattern because frame A is recorded to two fields followed by frame B recorded to three fields) repeats after four film frames. Virtually all high-end commercials, movies, and non-live television shows use this process prior to being broadcast.

Note that there is not a one-to-one correspondence between film frames and video frames after this pull-down occurs. The duration of a video frame is four-fifths the duration of a film frame. Because of this discrepancy, if you tried to match a specific number of whole video frames to some number of whole film frames, the durations would seldom match perfectly. In order to maintain overall synchronization, there is usually some fraction of a film frame that must be either added to or subtracted from the duration of the next edit. This means that in the cut list, Cinema Tools occasionally has to add or subtract a film frame from the end of a cut in order to maintain synchronization. For this reason, if you edit 3:2 pull-down video, the Cinema Tools cut list is only accurate to within +/– 1 frame on each edit.

This accuracy issue is easily resolved by using the Reverse Telecine feature (or third-party hardware or software) to remove the extra fields and restore the film’s original 24 fps rate before you begin editing digitally, providing a one-to-one relationship between the video and film frames. Setting the Final Cut Pro editing timebase in the Sequence Preset Editor to 24 fps (or 23.98 fps—see Synchronizing the Audio with the Video) allows you to edit the video and generate a very accurate cut list. See Determining How to Prepare Source Clips for Editing for more information about issues related to these options.

What’s an A Frame?

You will see and hear references to “A” frames whenever you are involved with 3:2 pull-down video. As the previous illustration shows, the A frame is the only one that has all its fields contained within one video frame. The others (B, C, and D frames) all appear in two video frames. Because the A frame is the start of the video five-frame pattern, it is highly desirable to have one as the first frame in all video clips. It’s common practice to have A frames at non-drop frame timecode numbers ending in “5” and “0.”

See About A Frames for more information.

Working with PAL Video

The PAL video frame rate is exactly 25 fps. There are two methods used when transferring film to PAL: running the film at 25 fps (referred to as the 24 @ 25 method), and adding two extra fields per second (similar to NTSC’s 3:2 pull-down, referred to as the 24 & 1 method, or the 24 @ 25 pull-down method).

24 @ 25 Method

Running the film at 25 fps sets up a one-to-one relationship between the film and video frames. The drawback is that the action in the film is sped up by 4 percent, and the audio will need an identical speed increase to maintain synchronization. To take advantage of the wide variety of 25 fps video equipment available, you can choose to edit with the action 4 percent faster. Another option is to use the Cinema Tools Conform feature to change the clip’s timebase to 24 fps, correcting the speed. The video can then be edited with Final Cut Pro as long as the sequences using it have a 24 fps timebase.

Figure. Diagram showing 1 second of 24 fps film frames converted to video and sped up to 25 fps.

Note: Final Cut Pro includes an Easy Setup and sequence preset with “24 @ 25” in their names, as well as a timecode format named “24 @ 25.” These are all intended to be used with clips that originated as PAL 25 fps video but have been conformed to 24 fps video. See Working with 25 fps Video Conformed to 24 fps for more information.

24 & 1 Method

Adding two extra video fields per second (also known as the 24 @ 25 pull-down method in Final Cut Pro) has the advantage of maintaining the original film speed, at the expense of losing the one-to-one film-to-video frame relationship. This method records an extra video field every twelfth film frame.

Figure. Diagram showing how the 24 film frames in 1 second are stretched to 25 frames by adding an extra field after the twelfth and twenty-fourth frames.

Working with 24p Video

With its frame rate and progressive scanning, 24p video is well suited for use with telecine transfers. It uses the same frame rate as film, providing a one-to-one relationship between the film and video frames without requiring a frame rate conversion.

Your Final Cut Pro system needs to be equipped with specialized hardware to capture 24p video, either as compressed or uncompressed clips. Alternatively, some DV cameras, such as the Panasonic AG-DVX100 camcorder, can shoot 24p video and use the 2:3:3:2 pull-down method to record it to tape at 29.97 fps (the NTSC standard). Using Final Cut Pro and Cinema Tools, you can capture this video and remove the 2:3:3:2 pull-down so that you can edit it at 24 fps. See Adding and Removing Pull-Down in 24p Clips for more information.

Note: When used as part of an NTSC system, the 24p videotape recorder’s (VTR’s) frame rate is actually 23.976 fps (referred to as 23.98 fps) to be compatible with the NTSC 29.97 fps rate.

Timecode Considerations

There are several general issues related to timecode that you should be aware of. If you’re using NTSC video, you can also choose between two timecode formats.

General Timecode Tips

When using video or audio equipment that allows you to define the timecode setting, it is recommended that you set the “hours” part of the timecode to match the tape’s reel number. This makes it much easier to recognize which reel a clip originated from. It is also best to avoid “crossing midnight” on a tape. This happens when the timecode turns over from 23:59:59:29 to 00:00:00:00 while the tape is playing.

You have the option to use record run or free run timecode during the production:

  • Record run timecode: The timecode generator pauses each time you stop recording. Your tape ends up with continuous timecode, because each time you start recording it picks up from where it left off.
  • Free run timecode: The timecode generator runs continuously. Your tape ends up with a timecode break each time you start recording.

To avoid potential issues while capturing clips, it is strongly suggested that you use the record run method, which avoids noncontinuous timecode within a tape.

Whenever a tape has noncontinuous timecode (with jumps in the numbers between takes), make sure to allow enough time (handles) for the pre-roll and post-roll required during the capture process when logging your clips. See the Final Cut Pro documentation for additional information about timecode usage.

About NTSC Timecode

Normal NTSC timecode (referred to as non-drop frame timecode) works as you would expect—each frame uses the next available number. There are 30 frames per second, 60 seconds per minute, and 60 minutes per hour. Because NTSC’s actual frame rate of 29.97 fps is a little less than 30 fps, non-drop frame timecode ends up being slow (by 3 seconds and 18 frames per hour) when compared to actual elapsed time.

To compensate for this, drop frame timecode skips ahead by two frames each minute, except those minutes ending in “0.” (Note that it is only the numbers that are skipped—not the actual video frames.) This correction makes the timecode accurate with respect to real time but adds confusion to the process of digital film editing.

With non-drop frame timecode, once you find an A frame, you know that the frame at that frame number and the one five away from it will always be A frames. For example, if you find an A frame at 1:23:14:15, you know that all frames ending in “5” and “0” will be A frames. With drop frame timecode, you are not able to easily establish this sort of relationship.

Note: It is standard practice to have A frames at non-drop frame timecode numbers ending in “5” and “0.”

It is highly recommended that you use non-drop frame timecode for both the video and audio in all film editing projects, even though both Cinema Tools and Final Cut Pro are able to use either type. Whichever you use, make sure to use the same for both the video and audio tapes.

Note: PAL timecode does not have this issue—it runs at a true 25 fps.

What Happens to the Timecode After Using Reverse Telecine?

The Reverse Telecine feature (used to change 29.97 fps video to 23.98 fps video) directly affects the timecode of the video frames. Because Cinema Tools must generate new 23.98 fps timecode for the frames (based on the original timecode), you may see a difference between the burned-in timecode numbers and the numbers shown in Final Cut Pro. Though the timecode discrepancies between the window burn and Final Cut Pro timecode may be confusing, Cinema Tools tracks the new timecode of the 23.98 fps video and is able to match it back to its original NTSC or PAL values, and thus back to the film’s key numbers.

Note: The Reverse Telecine feature is most often used to convert the NTSC video to 23.98 fps to match the audio timecode, but it can also convert the video to 24 fps.

This is what happens to the timecode: reverse telecine removes six frames per second, so the timecode numbers continue to match at the beginning of each second. This means that a clip that lasts for 38 seconds when played at its NTSC rate of 29.97 fps will still last for 38 seconds when played at the reverse-telecined rate of 23.98 fps.

Figure. Diagram showing the fields discarded when using reverse telecine on NTSC video.

In the above illustration, the blue NTSC fields represent fields that are removed during the reverse telecine process on a clip using traditional 3:2 pull-down. (See Adding and Removing Pull-Down in 24p Clips for information about 2:3:3:2 pull-down.) The window burn NTSC timecode will be different from what Final Cut Pro shows for all frames except the first one of each second, regardless of the clip’s length.

What Happens to the Timecode After Using Conform?

There are three common situations you would use the Conform feature for:

  • Converting PAL 25 fps video to 24 fps: The timecode is not changed, which ensures that an EDL exported after the clips are edited will accurately refer to the original PAL timecode. The drawback is that the timecode, at 25 fps, no longer accurately represents the true passage of time when played at 24 fps because each frame is displayed for a slightly longer time. See Working with 25 fps Video Conformed to 24 fps for more information.
  • Conforming 29.97 fps video to 29.97 fps: The timecode is not changed. This process is used to correct issues in a QuickTime file prior to using the Reverse Telecine feature. See Solutions to Common Problems for more information.
  • Converting NTSC 29.97 fps video to 23.98 fps: The timecode is altered, with a number skipped every five frames. This conform situation is rarely used.

See Using the Conform Feature for more information.