Skip to content

A Brief History of Time(code)

Share this Post:

I know you’ve seen the numbers, those little incrementing digits that appear on the video screen now and then — sometimes by accident, some-times on purpose. You’ve also run across them on many types of professional gear, including editing systems, videotape recorders, media servers, audio boards and even lighting consoles.

The numbers are called timecode, or more formally, SMPTE timecode, named for the national organization that developed and standardized the system in the late ‘60s — the Society of Motion Picture and Television Engineers. A timecode number is arranged into four groups of two digits each, separated by colons or semicolons (and yes, the punctuation does have significance). From left to right, the groups stand for hours, seconds, minutes and frames.

To simulate a real-time clock, the Hours column counts up from 00 to 23, the Minutes column counts up to 59, the Seconds column counts up to 59 and the Frames column (in NTSC video) counts up to 29. And, if you left your timecode generator continuously running, at 23:59:59:29, the whole thing rolls over — back to 00:00:00:00 and clocks up again.

Well Paul, that’s just peachy, but what do those numbers actually do — and why are they hanging around my user interface?

Cue Dots
By way of history, when I started editing professionally in the ‘70s and ‘80s, I cut programs using Ampex AVR-2 videotape recorders. These re-markable beasts weighed in at around 600 pounds each and used huge two-inch tape reels. The high-band images were beautiful, but the editing was fairly inaccurate because the machines relied on control track pulses for synchronization. (Essentially, control track pulses are electronic sprocket holes, for want of a better analogy.)

Editing was very linear (and rather labor intensive). To make just a single edit, you placed an electronic “cue dot” on the record machine at the point where you wanted to cut. This dot was recorded on an audio track, and it instructed the VTR to start a clean “insert” edit. Next, you found the desired “in-point” on the playback machine, and you reset both tape counters to zero. Then, you manually rolled each machine back five seconds and started a multimachine gang roll. The VTRs would roll up to the edit point together, lock horizontally and vertically, and if the stars were aligned perfectly in the heavens, you just might get the desired edit. If not, back to square one, hoping that the tape counters didn’t drift. If you wanted to trim the edit by a frame or two, serious skill was required.

Precision Guaranteed
And then a miracle occurred at our production house. Along came timecode, an electronic digital signal that was recorded on the audio track of the VTR.  Using a timecode generator, this signal was recorded along the entire length of the tape, and it essentially locked each frame of video to a single unique eight-digit timecode number. You could search to that number, cue to that number and, most importantly, synchronize to that num-ber. Suddenly, editing was frame accurate.

For example, if your “keeper” take started at 01:21:18:01 on playback reel 1, it was always that exact number. Even if the tape counter drifted, even if the reel was put away and brought back, that visual in-point was locked to that timecode. Using a computerized editing system, you could now control the VTRs remotely, preview over and over again to your heart’s content and trim frames with complete precision.

NDF and DF
But all was not well in timecode land. In NTSC video, the color television frame rate is 29.97 frames per second — not quite 30 fps. Because of this small discrepancy, editors soon discovered that if you cut a show to exactly an hour in timecode length (01:00:00:00), and compared this length against a real-time clock, the timecode program length was 108 frames (3.6 seconds) longer than an actual hour. The sponsors were not pleased.

To reconcile the difference, a special variation of SMPTE timecode was developed, called drop frame. The original version, now called non-drop frame (or NDF), is characterized by the colons between digits. Drop frame (or DF) is characterized by semicolons between digits.

In the drop-frame scheme, to run the timecode in sync with a real-time clock, two frames are dropped every minute, except for the tens-of-minute changes. For example, in NDF, 01:02:59:29 advances to 01:03:00:00. But in DF, 01;02;59;29 advances to 01;03;00;02. The two missing frames don’t affect the video at all because the timecode still runs sequentially. However, the net result is that after an hour of DF timecode, an exact hour of real-time has elapsed.

About That User Interface

Today, DF is the timecode standard in NTSC, and you’ll find that it’s the default setting on most professional gear that works with video clips. This includes editing systems, media servers, cameras and much more. For example, in addition to audio and video tracks, a timecode track is also re-corded each time you press the Record button on a DV camera. Besides “free run” mode, the camera’s internal timecode generator can be set to the time of day, or it can be preset to assist with multireel edit sessions (e.g., Reel 1 timecode starts at 01:00, Reel 2 starts at 02:00, etc.).

But why would there be a timecode interface on my lighting console? The answer is connectivity. If you look at the specs for any top-flight con-sole, a SMPTE timecode interface guarantees a high degree of control and synchronization with external gear — and that’s the name of the game with pro equipment. By adding SMPTE timecode to a console’s array of DMX512, MIDI, Ethernet and serial control interfaces, the LD can say yes when called upon to control a video device with timecode precision.   

Paul Berliner is president of Berliner Productions in Davis, Calif. — a company providing video production and marcom services to the broadcast and entertainment industries. He can be reached at pberliner@plsn.com.