page last updated 5/26/02
This page is intented as a reference for negative cutters using DigiConform, as well as producers and film editors confronted with the choice between cutting a particular show in 24 fps or in 30 fps. This choice has to be made before the creatve editing process begins.
Systematically, there is quite a difference between film projects cut in 24 fps and in 30 fps on an Avid - or any other NTSC video editing system. This is not that obvious at first glance, since video outputs from both kinds of shows come in 30 fps.
This page explains how 24 fps projects are treated differently from 30 fps film projects in computer editing systems, and why there is really no ultimately satisfactory way to conform 24 fps film to a video reference tape that was cut in 30 fps.
Here is an overview of this page:
Telecine 24 fps Film to 30 fps Video in NTSC
Telecine 24 fps Film to 30 fps Video in NTSC
Regardless of whether a show is cut at 24 fps or at 30 fps, the film has to be transferred to video tape first. There are 24 film frames and 30 video frames for each second of running time. (The fact that NTSC video is actually not 30 fps, but 29.97 fps is of no importance in this context.)
The task at hand is to fill 30 different frames of video with the content of only 24 film frames. Some film frames will obviously have to be repeated in order to create the 6 additional video frames needed per second.
The smallest common denominator of 24 and 30 is 6: As you know, 6 x 4 = 24 and 6 x 5 = 30. As a matter of fact, the film's frame lines line up perfectly with the video's 'frame lines' 6 times per second, every 4 film frames, or every 5 video frames. Therefore it is sufficient to look at just a 6th of a second of running time when it comes to transferring film to NTSC video. Once a solution works for a 6th of a second, it can simply be repeated over and over.
When it comes to using 4 film frames to fill 5 frames of video, it is helpful that each video frame is sub-divided into 2 fields. These 2 fields appear on video screens and TV sets one after the other.
Note that the time code burn-in window on video dailies tapes frequently shows the field number in the 9th, or last digit. Usually, when you jog very slowly through a video tape each horizontal line of noise going through corresponds to one field.
Eventually, 4 film frames need to be distributed into 5 video frames with 2 fields each. The next picture shows how 4 film frames are used to fill 10 video fields. This process is called 2:3 pulldown:
Cutting at 24 fps in the Avid (or similar computer-based editing systems)
In a film project at 24 fps, usually only the first field of each video frame is captured while loading the dailies tapes into the computer. This results in 5 full video frames for each 4 film frames.
To get to the true 24 film frames in the Avid, every 5th frame has to be dropped entirely:
Subsequently, the film can be cut in the Avid at 24 fps. This ensures that for each potential cut point in the Avid there is exactly one corresponding cut point on film:
24 fps shows in DigiConform
When you work in 24 fps in DigiConform, you deal with video that has one field only. It does not matter whether that is 1st fields only or 2nd fields only. When you start, you are asked to identify one frame that appears twice, in this example the second B-frame (You could just as well choose the first B-frame.):
From now on DigiConform will skip over every 5th frame starting with the one that was identified, thus displaying just the 24 genuine film frames per second.
This makes sure that the footage and the frame lines are exactly in sync between film, the Avid in 24 fps and DigiConform:
This means DigiConform transports exactly one frame of video for each frame of film, and the next film frame appears on the DigiConform screen exactly when the synch-block's frame line is at "12 o'clock".
'Fast' 30 fps tapes from 24 fps shows
One of the available modes in DigiConform is NTSC without Pulldown. In this mode one video frame is assumed to correspond exactly to one video frame. Output tapes for this mode can be made from 24 fps shows in the Avid, although it is rarely ever done in NTSC. The video played in a regular video deck will run about 25 percent faster than normal, and sound will be of noticeably higher pitch than usual. Depending on the software version in the Avid, this mode is either called "Video Rate 100% +" or "29.97 NTSC" (as opposed to "23.976 NTSC").
This mode assumes that video has been cut in 24 fps, or shot and cut at 30 fps. Shooting film at 30 fps is not commonly done.
If you would like to find out more about this, please contact us for more information.
Cutting film at 30 fps in the computer or on video
Any NTSC dailies tape can be loaded into any computer-based editing system at 30 fps, or they might be cut in a tape-to-tape setup. One of the problems with conforming film based on a 30 fps video cut is that there only is a direct correspondence between potential cut points on video and on film in 1 out of 5 cases (green arrows):
This means that in 4 cases out of 5 the film cut point is slightly off from the point that the video is cut at (dotted lines). Another problem is the accumulation of slight differences in length. For example, if the video is cut between the AA and the BB frames the most obvious point to cut the film at is the frame line between the A and B frame lines. However, this point in time occurs slightly later on film than it does on video.
Throughout a reel these small discrepancies are bound to accumulate, as on average 4 out of 5 cut points are slightly off. To compensate for that, cut lists often have one frame added or subtracted at the tail end of a cut.
In 30 fps the pulldown is unpredictable for each cut
Also, there is no way to predict within a 6th of a second how the film frames have been pulled down to fill the video frames, how the 'pulldown sequence' has fallen for a particular piece of film. Any of the bottom 4 situations could apply:
The resulting problem, of course, is that inherently the film seems to be drifting against the video, since a film frame may start half a video frame early or late. This uncertainty gets even more aggravated when the video has been cut on a computer using single fields of video.
30 fps single field or double field
When working in 30 fps dailies tapes can be loaded into a computer with both fields, or with just one field (usually, the first field):
Working with tapes that have been cut with both fields can create a problem that appears to be a discrepancy between the video cut and the resulting conformed reel of film. For example, if the video is cut between the BC and CD frames, then the last film frame displayed on the computer screen is the C film frame, because that is what fills the second field of the last video frame before the cut.
The conformed reel of film, however, may very well have the splice right after the B film frame, in order to maintain the proper length for the reel. Thus, the last frame of film before the cut will be B. This may make a crucial difference when doors open or close, eyes blink, shots are fired, etc. And this may even be aggravated by an adjustment in the cut list that may call for another frame to be dropped at the end of the cut.
This set of problems is not specific to any equipment at all. It is inherent to the process of cutting 24 fps film in 30 fps, no matter what editing system is used. And it makes conforming film to a video tape reference that was generated in 30 fps all the more difficult.
30 fps shows in DigiConform
One way to treat 30 fps shows in DigiConform is to pretend they were 24 fps shows. This means that inherently you should expect to see the wrong frame to be skipped from display in about 4 out of 5 cuts.
This is illustrated in the following example:
Even if the skipped frame is the correct one for one cut, it can be off in the next cut. Here, 2 frames are cut out:
The resulting join, indicated by the thick, vertical line, throws the correct display off:
These are the frames that will be displayed on the DigiConform screen. After the cut point the B-frame gets displayed twice, and the D-frame is missing.
To make things worse, cut points may be off by one frame due to length interpolation inherent to cutting film in 30 fps. However, the cut list may indicate these discrepancies.
The advantage of using the NTSC 24 Pulldown mode is that you will still see the video frame change exactly when the sync-block's frame line is at "12 o'clock". Of course, the overall synch between the video in DigiConform and the film is not affected.
Another solution in DigiConform
Recently, a new mode, called NTSC 24:30 was implemented in DigiConform that transports 24 film frames for each 30 video frames. The drawback is the fact that only every 5th video frame will be transported in exact correspondence with the film frame line. In this example, this is indicated by the green arrows. The other 4 out of 5 video frames will be off by either 1/4 of a video frame (BB and DD) or by 1/2 a video frame (BC). Practically, this means that these video frames will not be transported when the frame line in the sync block hits 12 o'clock.
The shadows indicate that DigiConform displays video frames combining 2 fields.
The advantage of this mode is that no video frames get skipped over on the screen, and that the number of times you should expect to see different frames in the synch block and on the DigiConform screen should be greatly reduced.
How to determine if a show has been cut in 24 fps or 30 fps
This is not an easy task, since both kinds of output tapes come in 30 frames per second.
Just asking if a show has been cut on an Avid is not enough. Avids come in 2 different flavors, as Media Composers and as Film Composers. But both these can cut in both 24 fps and 30 fps. The question that should be asked is: "Is this a video project or a film project?". If the Avid is still available, you can ask someone in the cutting room to click 'Info' in the Project Window to display the frame rate of the show.
Something else to try is to look at the cut list: If you see the occasional cut having frames added or subtracted at the tail, it is a 30 fps show for sure. However, the reverse is not necessarily true, because display of those frames in the list is optional.
Blended video frames on an output tape sometimes hint at a 24 fps project. On some 24 fps shows film frames in the Avid are digitized by combining 2 video fields from 2 different video frames. Typically, this happens with a C film frame:
The result is that you see 2 different Time Code numbers on top of each other on one frame out of five. However, you always want to see the key code window to absolutely sharp and in focus. Any blurring in the key code numbers, as well as in footage counters, hints a problem with loading the dailies footage into the editing system. This is because key codes and footage counts always refer to exactly one, and no more than one film frame.
The Bottom Line
Considering that there is no un-ambiguous way whatsoever to convert a 30 fps cut into a 24 fps list, and given the choice between cutting at 24 fps or at 30 fps, it is highly recommended to cut at 24 fps. This does not preclude a video finish in any respect, but it facilitates conforming the film in a fashion that is 100 percent reflective of the cut made in a computer-based editing system.
© DigiConform, Rainer Standke, 2002