The Society of Motion Picture and Television Engineers (SMPTE --- pronounced SIMP-tee) did us all a big favor. This professional organization writes standards for film and video recording systems used in North America. When they wrote the specifications for SMPTE time code, they made it possible for all manufacturers to use the same techniques for reading and writing time codes so that everybody's tape and equipment would work together. The main problem with SMPTE time code is that it was designed for professionals, to operate on expensive equipment. When home videographers and prosumers came along, the venerable SMPTE time code was about $800 out of reach, and pro VCRs designed to use it, about $4000 out of sight. NEC, Sony, Videonics, GSE, and a few others stepped in with a few incompatible "consumer" time code methods that were less eloquent but more affordable. Sony's RC (Rewritable Consumer) time code was built into its popular but discontinued CCD V801 camcorder and VX3 and CCD-V6000.

Before plunging headlong into the turmoil of time code, let's first review the science of editing and the purpose of time code.

Taking Your VCR's Pulse

For every picture recorded on a videotape (there are 30 per second), a magnetic pulse is recorded along the edge of the videotape. These control track pulses guide the videotape on playback so that the video heads follow exactly the right path to give you a clear picture.

When you select an edit point on your video player and recorder, your edit controller takes note where both tapes are paused. It then backs up both tapes a certain distance, counting control track pulses as it goes. If, for instance, both machines backspaced five seconds, the edit controller would count backwards 5 x 30 = 150 control track pulses. The controller would then stop both machines and start them playing, counting in the opposite direction 150 pulses. When both machines reached pulse number 0, the record VCR would switch from PLAY to RECORD, to begin copying the desired scene. The purpose of this preroll is to allow the tapes to be in motion at the time the edit is executed; any aberration in speed could cause a glitch in the picture.

Control track pulses allow VCRs to edit fairly accurately, but not perfectly. When the VCRs begin to move from a standstill, the tape is moving so slowly that a pulse or two might not register. If you have ever listened to an audio recording played very slowly, you noticed that some words become unintelligible. Similarly, the control track pulses become unintelligible while the tape machine is getting up to speed causing the VCRs to miss a pulse or two, making the edit point inaccurate by a few frames.

This is not usually a big deal; who's going to notice if an edit occurred 1/15th of a second early or late? There are some circumstances when it becomes a big deal, however. If, for instance, you rehearse your edit (putting the machines through the motion without actually making the recording, just see if you like the way the edit looked), then both machines slip a few pulses during the rehearsal preroll and another few frames during the actual edit, taking the edit even farther astray.

Animation is one situation where absolute frame accurate editing is necessary: When you render pictures and record them all on tape, you don't want to miss any and you don't want to leave any gaps between frames.

Time Code Versus Control Track

The big problem with control track pulses is that they all look alike. (I'm sure they say the same thing about us humans.) The machine can count them backwards and forwards, but if it loses track by one or two pulses, it will never know. If, on the other hand, each pulse had an identity of its own, like a serial number, then pulse #1153 would always be #1153. If a machine were told to place an edit at pulse #1153, it wouldn't accidentally make the edit #1152. The machine would speed up, slow down, or do whatever was necessary to reach #1153 at the right moment, to execute its edit on schedule (or would abort the edit for you to try again). This is one of the big advantages of true time code. It allows for frame accurate editing.

A few other advantages of time code: If you give your tape to someone to review and to select the desired scenes, they can write down the time code numbers which you can later use to find the exact places on the tape that they selected. If, on the other hand, you used control track pulses or an hours:minutes:seconds counter, the numbers wouldn't be exact, and depending upon whether you and they started the tape at exactly the same zero spot, your numbers may be ahead of or behind theirs.

Another advantage of time code editing is that when you remove a cassette from a machine and insert a new cassette, the machine will read the time code and know exactly where it is on the new tape. If you were using control track, removing one tape and putting in another would confuse the machine. The VCR wouldn't know if you were picking up at the same number where it left off, or whether to start at zero, or what. For the numbers to mean anything, you would have to rewind the tape to the beginning, zero the counter, and then run the tape forward to your edit point. And if you changed tapes again, you would have to rewind that tape to the beginning, zero it, and skim ahead to your edit point. This process wastes a lot of time, especially if you have many tapes.

How SMPTE Time Code Is Recorded

SMPTE time code is a digital signal whose ones and zeroes assign a number to every frame of video, representing hours, minutes, seconds, frames, and some additional user/specified information such as tape number. For instance, the time code number 01:12:59:16 represents a picture 1 hour, 12 minutes, 59 seconds, and 16 frames into the tape.

If you were to listen to the digital signal it would sound like a wooden chair being scraped across the floor. The signal is made by a time code generator connected to your camcorder or VCR. It can be recorded during the original production or, in some circumstances, added later. During editing, a time code reader, associated with your videotape player, converts the digital signals back into useful information.

There are various ways to record time code, depending on your needs and the equipment used. Longitudinal time code (LTC) is recorded along the length of the tape in the form of a modulated audio signal. The signal may be recorded on a spare audio channel or, in the case of professional equipment, on an "address track" available for just this purpose.

One advantage of LTC is that it can be recorded after the original taping. Doing so on consumer gear, however, will erase your linear audio track. If you are adding narration or music to your recording, you may already be using your audio track and cannot sacrifice it for time code. Some industrial VCRs have stereo linear audio tracks, allowing you to record time code on the left track and audio on the right. HI8 and 8mm videographers can record SMPTE time code on their PCM (Pulse Code Modulated) audio tracks available on many editors. Like the industrial SVHS and VHS linear audio tracks, the PCM tracks have two channels allowing you to record audio on one and time code on the other. The time code and audio can be erased and rerecorded without touching the video.

Another disadvantage of LTC is that the raucous scraping sound of the time code sometimes bleeds into your adjacent audio track. You have to be very careful about the recording volume to assure that the time code signal is loud enough to work, but soft enough not to bleed into your audio.

Another disadvantage of LTC is that it cannot be read when the tape is still framed or traveling at very low or high speeds. This makes shuttling and jogging an LTC tape a bit more difficult, and slightly inaccurate; the controller must remember the last good code information it heard, then estimate how much the tape traveled coming to a stop. Furthermore, most consumer decks mute the audio when they leave the PLAY mode, which kills your time code signal as you jog or scan. Consumer gear is very unfriendly to time code.

LTC, because it is an audio signal, deteriorates rapidly from one generation to the next (linear audio tracks generally have poor fidelity). LTC time code seldom survives dubbing or editing. The situation isn't hopeless, however. When dubbing an LTC tape, you can play the time code signal into a time code generator that will refabricate a sharp new time code signal which it passes to the recorder. The time code numbers will be the same, but the time code recording will be an original (straight from the generator).

Vertical interval time code (VITC, pronounced VIT-cee) is a second way to record SMPTE (or any other) time code on a video tape. It places the code in the video signal in an unused area above the active picture, just below vertical sync. VITC offers several advantages: it doesn't use up an audio track, allowing you remix or dub your stereo audio at will.

Another advantage of VITC is that it can be read by the video players at speeds ranging from still frame to about 10 times normal forward or reverse.

VITC code also survives dubbing without the deterioration seens when dubbing LTC. Beware that some TBCs, if used during the duplication, may take it upon themselves to strip off the time code in the name of "correction."

One special requirement of VITC is that it must be recorded along with the original video signal on your source tapes. Since consumer camcorders don't have time code inputs, it forces you to copy your tapes and add this signal during the duplication process, taking you down one extra generation.

Since the time code is imbedded in the picture, you have to be aware that video insert edits will replace the time code along with the picture. In order not to have gaps in your time code, and to assure that time progresses smoothly (without skipping any seconds), the VITC time code generator must have jam sync capability (most do). This feature allows the time code generator to "listen" to the tape as it plays back, lining itself up with the numbers recorded on the previous scene. Then, when the VCR goes into the record mode, the time code generator knows what the last time code number was and picks up where the numbers left off.

Drop Frame Versus Non-Drop Frame

The NTSC color system doesn't create exactly 30 frames per second, but closer to 29.97 frames per second. So, what's a .03 seconds among friends, you ask?

If time code follows the color video signal, then it's a clock that is running slightly slow. In fact, at the end of one hour of non-drop frame time code, 1 hour, 3 seconds, and 18 frames worth of actual time will have elapsed, a discrepancy of 108 frames. For this reason, SMPTE time code allows you to choose whether you want time to be correct (but with some of the numbers skipped), or have a sequential number for every video frame (but not have the numbers accurately tell you the length of your show). Whatever method you use for recording your time code, that is the method you must use when editing your time code. If you mix the systems, you are in for a monster migrain.

Non-drop frame mode is used primarily by people making videodiscs, CD-I or other interactive programs where every frame needs to have a sequential identity so that the frame can be accurately located. The more popular method, drop frame mode, is used by most editors because it provides an accurate listing of elapsed time.

The drop frame system is rather ingenious yet arcane. The machine drops two frame numbers every minute on the minute except each tenth minute. No actual picture frames are dropped, only certain frame numbers. For example, in non-drop frame mode, your counter goes from 01:12:59:29 to 01:13:00:00. These are the SMPTE numbers showing hours, minutes, seconds, and frames. After the 29th frame comes the next second and the zero frame. In the drop frame mode, the changeover looks like this: 01:12:59:29 to 01:13:00:02. Notice the frame number 00 and 01 are dropped. These occasional dropped frames add up to 108 at the end of an hour, making time code "time" and clock time match exactly.

Editing with Time Code

When you go out on a shoot, you have the choice of creating time code while you shoot, or adding it later. If recording time code during the shoot, you need video equipment capable of accepting an input from the time code generator. Incidentally, if you stop recording for a reason, the generator must be told to stop, and jam sync itself when you start recording again. This will assure that the time code stays continuous and sequential. Although it is possible to edit tapes that don't have sequential and continuous time code, if you wish to use nonlinear editors (computers) to edit, you will need continuous, unbroken time code.

In the analog world you could get this by copying your source tapes while adding time code. Here you can add the time code to the audio (hi fi or LTC) channel or to the video (VITC). VITC is better because it avoids the complications with the audio channels. And while we are at it, let's make 2 copies of the source tape. This second copy will also have time code recorded on it but will have one additional feature: the time code numbers will appear visibly in the TV picture. This is called a window dub. Now you have two copies of your program, one with invisible time code to be used for editing, and the other with visible time code handy for you or your client to view, perhaps at home. These window dubs are for logging purposes only, so the tape can be VHS, and you don't mind if it gets beat up. When the desired scenes are selected, you or the client can jot down the time code numbers, or if using a computer and non-linear editing software, it's possible to just press a mark button and have the computer log the numbers.

If you use digital recorders, you could digitally copy your tapes, adding clean, continuous time code automatically as you went. If working in High Definition, where nonlinear editors are very expensive, you would mark all of the in and out points from the desired scenes, and you end up with an edit decision list (EDL). You can bring this list to a professional editing shop that will now perform the edits for you using your "good" copy of the source tapes. You could have the editing machine just follow the numbers, and you could come back in a half hour and see the finished production, or you could watch as the process unfolds. Usually a producer likes to see each edit, possibly trimming or rethinking a few after seeing how they fit together.

Having the prepared list is an important step to saving money in the editing suite. Professional on-line editing suites charge up to $400 per hour. You don't want to spend that kind of dough thinking deep thoughts about your show's content. Nor do you want to spend $7 minutes discussing with your client which scenes to leave in and which scenes to take out. The edit decision list brings you so close to "perfect" that all you have to do is tweak the program a little and maybe add a few special effects.

List Managed Edits If you use a nonlinear editor, the computer gives you an opportunity to "save project". This is not the same as recording the final show to DVD or tape. This is the saving of the internal database that was used to create your show from all the pieces. It is often an edit decision list (EDL).

Edit decision lists embrace effortless changes in the final program. Months or years later, if the raw material is still available, scenes can be moved, deleted, added, and previewed easily, simply by moving them (their time code in and out points) to different parts of the list or timeline; it's like using a word processor. More importantly, if a client asks you to add or remove a scene, lengthening or shortening a finished show, you simply change the EDL or timeline and let the machines re-edit the show from scratch. Those of us who have re-edited a finished show the hard way know how little joy there is to this process. But without time code, this is the only way to re-locate all the scenes and reassemble them. There's a saying, "Machines should work, people should think." It applies to using EDLs.

By saving the EDL along with your tapes, you can easily update a show adding new or changed information a year later. Again, you add or change a few lines on the list and the editing machines suffer putting the whole show together while you jog or sip tea.

Life is easier when every frame of video has an identity of its own. SMPTE did us all a big favor.
 About the author  About Today's Video 4th. ed.  Return home