Analogue Wayback, Ep. 4

Divide 24 into 60.

Back in the 1970s and as late as the early 80s, film editors working on film ruled the world of high-end regional and national TV commercials. They had the agency connections and, as such, controlled the workflow in post. Most of the top commercials were shot on film, so after the lab processed the negative the editors (or editorial companies) took it from there. Often they would subcontract and outsource any further lab services, video finishing, and the sound mix as part of their arrangement with the client.

Timecode-based video editing was just starting to grab hold in the mid-70s. Early attempts at creative, nonlinear (“offline”) editing systems were commercially unsuccessful. What stuck were the linear (“online”) edit suites built around switchers and  2″ VTRs. Occasionally producers, directors, and agencies would go directly into an online bay and cut start-to-finish there. But at the highest end, spots were typically edited with the agency folks sitting around a flatbed film edit bench, like a KEM or Steenbeck. Online bays and editors were considered technical and used to conform the edit, thus replacing the work previously done by labs to finish and distribute those TV spots.

Frame rate and interlaced – the two flies in the ointment

This was the world of interlaced NTSC in the US (PAL in much of the rest of the world). There was no HD, progressive scan, or 24p video (23.976). NTSC operated and continues to operate at 29.97fps using interlaced fields – i.e. two unique images (fields) per frame, each containing half the content of that whole frame. Therefore, NTSC is really 60 images per second, or expressed in technical terms, 59.94i. Film is typically shot at a true 24fps (or now offset to 23.976) where each frame is a complete image. This presents the dilemma of converting 24 film images into 60 video fields (let’s keep it simple and skip the 23.976 vs 24 and 59.94 vs 60 discussion).

Every four film frames (out of 24) corresponds to five video frames (out of 30). Telecine engineers devised the scheme of 3:2 pulldown to transfer film to video. The pattern of corresponding frames or cadence is 2:3:2:3. Consecutive film frames are alternately scanned for two or three video fields. In this cadence of so-called whole and split frames, the first film frame in the sequence covers both fields of the first video frame. The next three film frames continue the cadence until we reach the fifth film frame/sixth video frame. At this point the film and video frames sync up again and the cadence repeats itself. If you park on a single video frame and see only one film image, that’s a whole frame. If you see a blend of two film images, then that’s a split frame.

The hybrid editing workflow model

This 2:3:2:3 cadence is how film aligns with interlaced NTSC video, but there’s also the issue of different counting methods. Video editors work in SMPTE timecode – 00:00:00:00 – hours, minutes, seconds, frames. Film editors work in feet and frames – 0000+00 – 16 frames/foot for 35mm and 40 frames/foot for 16mm. Where’s the rosetta stone to go between them? 

Since transfer of the negative to video was already possible in the 70s, a hybrid editing model developed. Shoot film – transfer to video – edit film – conform the film edit on video. That approach goes like this. Workprint is created by the lab from the negative. This workprint is cut by the film editor. The negative itself isn’t cut (at least in the case of most commercials). The key is to get the proper cross-reference between the two forms of media. To do that you need a common starting point for each film roll.

The key

The lab punches a hole at the start of each negative film roll and also adds edge numbers for feet and frames. The hole punch becomes 0000+00. During the film-to-tape transfer of the negative, the telecine operator/colorist syncs the hole punch to a zero timecode value on the videotape, such as 01:00:00:00. Now there’s a common starting point. 0000+00 = 01:00:00:00. In the whole and split frame cadence (2:3:2:3) the first, second, and fourth film frames have matching whole frame video frames (:00, :01, :04). The third film frame is split across two video frames (:02, :03).

Next, the film editor works through the creative edit of the commercial. When the cut of the workprint is locked, a cut list is delivered to the online edit facility. This is similar to a video EDL (edit decision list), except that events are listed in feet and frames instead of timecode. From the online video editor’s point of view this foot/frame number isn’t an absolute value, but rather an offset from the common 01:00:00:00 start point on the videotape transfer.

For each edit, the offset between the two must be calculated in order to get to the correct matching shot on the video tape. It really doesn’t matter whether the source image lands on a split or whole frame timecode, because the source is unedited camera original. However, it does matter in the timeline, because the duration on the film clip determines the out-point of the video edit. In order to maintain the proper rhythm of the film editor’s cut, certain frames will be ambiguous. The video editor will have to make a judgement call whether to trim one frame forward or backwards.

A cheat sheet to the rescue

I worked as on online editor in Jacksonville in the 1970s. One of our clients was the largest ad agency in Florida. Their creative director had been a film editor in an earlier life. Any time large campaigns for one of their brands was produced, he would lock himself into a cutting room at the local film lab and edit the spots on their Steenbeck. It then became our job at the post house to translate that cut into a video master based on the steps I’ve just described. 

Fortunately, my boss was experienced in both film and video editing. He took it upon himself to create a handy conversion chart that cross-referenced 10 minutes of film (the approximate length of a film roll) between a foot/frame count and timecode. He even marked all of the ambiguous frames. Using this chart, it was easy to take a film number from the cut list, look up the timecode on the chart and type that into the edit controller. Preview the edit, trim as needed, and commit to the edit. Rinse and repeat.

I’m sure that this sounds rather old school to most readers, but it’s one of those arcane skills that still has validity today. Film is not quite dead yet. Avid still integrates some features of cut list conversion. As a business model, film editorial companies of the past are top commercial edit shops today, using Media Composer, Premiere, and Final Cut in place of Steenbecks and KEMs. So while the technology may have changed, many of the concepts haven’t.

©2022 Oliver Peters