TECH TIPS


Don't Let 3:2 Pulldown Get You Down
After Effects to the Rescue
by Ben Bardens

One of the more technical aspects of dealing with digital video comes with footage transferred from film. Many of us run into trouble in this area, preferring to concentrate on the artistic aspects of our craft. Yet we all know how important it is to work to the correct technical settings, providing our clients and collaborators with footage in the appropriate format. In this Tech Tip, I’ll offer some background and tips for dealing with one of these more technical areas–removing 3:2 pulldown and processing film-transferred footage to ensure a proper frame rate conversion.

Background
When 24fps film is transferred to 29.97 fps (NTSC) video, the frame rate is converted using the process called 3:2 pulldown, in which the film frames are distributed across video fields in a repeating 3:2 pattern. Technically, 3:2 pulldown consists of two parts: the pulldown, or slowing of the film footage so that it becomes cleanly divisible into the video frame rate, and the cadence and phase order, or pattern in which the film frames are copied across the greater number of video fields (see figure 1).


Fig. 1: A graphical representation of 3:2 pulldown, illustrating pulldown, cadence and phase order.

The 3:2 process spreads four frames of film across five interlaced frames of video (10 video fields). The process results in whole (W) frames, and split-field (S) frames. Whole frames are video frames where both fields match (they are pulled from the same frame of film). Split-field frames contain fields that do not match; each field is pulled from an adjacent frame of film (see figure 2). In order to edit digitally and generate cut lists that match the original film, it is necessary to remove 3:2 pulldown and conform digitized footage back to its 24fps origin.


Fig. 2: A split-field frame from a digitized clip without 3:2 pulldown removed. This footage would either need to be re-digitized without 3:2 pulldown or have 3:2 pulldown removed prior to being used in an editing or compositing system.

By default, when digitizing into a 24p project, the Avid will attempt to remove 3:2 pulldown. Its success will depend on what frame the in-point of the capture starts on. The Avid can only interpret the cadence of the footage and remove pulldown correctly if the in-point of the capture (referred to as the “pullin”) is on field one of either an “A” or “B” frame. The default pullin is the “A” frame, although you can change this in the pullin column of your bin window. Editors should visually inspect their source footage by stepping through it for several frames to identify the cadence and locate an “A” frame prior to initiating capture. If you must capture film-transferred footage using a program other than the Avid, it is best to do so at video frame rate and leave the 3:2 intact–and then remove it after the fact using After Effects (AE).

For captured footage that contains 3:2 pulldown, the easiest and most reliable way to remove the pulldown is to run it through After Effects. After Effects provides the option to specify the cadence and render the footage back out at 24fps. If you captured using a program other than the Avid, your footage will most likely be QuickTime or AVI format, using DV or some kind of Motion JPEG compression. In any case, import your footage into an After Effects project. With the footage selected in the Project window, go to File -> Interpret Footage -> Main. In the main interpret footage dialog locate the Fields and Pulldown area (see figure 3).


Fig. 3: The Fields and Pulldown area of the Interpret Footage Main dialog.

First determine the correct field order.
To do this, start by choosing upper or lower and clicking OK. In the Project window, option double-click (Mac), or alt double-click (PC) the footage. This will open the footage in a footage window. Scrub along the timeline at the bottom of the window and look for an area in the clip where there is motion on 1s. Step through the footage for several frames using the Page Up and Page Down keys. If the motion appears staggered, as if the frames are being displayed out of sequence, then the incorrect field order is specified. Go back to the File -> Interpret Footage -> Main dialogue and specify the other field order.

Specify the phase order from the Remove Pulldown pop-up menu.
As illustrated in figure 1, footage transferred with a 3:2 cadence will have a WSSWW repeating pattern. If you do not know the cadence of your footage, click on the Guess 3:2 Pulldown button. As long as the correct field option is specified above, AE will be able to analyze the frames and determine the correct phase order for you.

Note: Guess 24Pa Pulldown is specific for Panasonic 24p DV cameras, not for film transferred by telecine.

Set-up and render your footage without 3:2 pulldown.
After specifying the correct interpret footage options, click OK. Create a new composition that matches the newly interpreted footage by dragging the footage item to the New Composition button in the project window (see figure 4).


Fig. 4: How to create a new composition that matches your source clip's frame rate, resolution and duration.

Make sure the resolution pop-up is set for Full, and the layer quality switch is set to Best Quality (see figure 5). From the Composition menu, choose Add To Render Queue.


Fig. 5: The layer Quality Switch and the Composition Window Resolution pop-up.

In the Render Queue window, click the pop-up menu next to Render Settings and choose Best Settings. Click the pop-up menu next to Output Module and choose Lossless (see figure 6). Click the text next to “Output To,” specify a name and save location, then click Render. This will output your footage at 23.976fps with the 3:2 pulldown properly removed. These settings will output the movie in either QuickTime or AVI format using non-lossy compression, such as the Animation compressor. If necessary, you can choose DV or some other type of compression from within the Output Module settings. You can also specify frame sequence instead of QuickTime or AVI format. Click on the word “Lossless” to manually specify Output Module settings.


Fig. 6: The queued composition in the AE Render Queue window.

Guild member Ben Bardens runs a graphics studio and works as an editor/technical director in the Burbank, California area. He teaches After Effects and Photoshop at Glendale Community College and can be reached at bbardens@glendale.edu. Find out more about his classes at www.glendale.edu/~bbardens.

[ return to top ]