Over the past several years, it has become ever increasingly common to record production tracks with a SMPTE timecode reference instead of the traditional 60
Hz sync pulse. This article is intended as an introduction and overview of the use of SMPTE timecode in conjunction with Nagra analog recorders and DAT. The original timecode system is known as non-drop frame, since it assigns a progressive number to every video frame (0-29). This is an accurate way of
tracking individual frames, but caused a problem for video editors when they compared elapsed "real time" to "videotape time". Due to the fact that video
actually runs at 29.97 frames per second and NOT 30 frames, editors discovered that they could be off by 3.6 seconds at the end of a one hour show.
In order to "synchronize" the clocks on the wall with the elapsed time counters in the edit system, video engineers developed drop-frame timecode,
which works sorts of like a leap year in reverse. Two frame numbers are dropped or skipped every minute, except when the number of the minute ends in zero, such as minutes 00, 10, 20, 30, etc.
It is important to realize that the video frames themselves are not
deleted. Only their numerical labels are affected.
The correct timecode frame rate for video is 29.97 fps. Other rate options for film include 30, 24, and 25 fps.
If you are recording for a video shoot, then use whatever mode (drop
or non-drop) that the videotape recorder is using. Usually, video prefers to use the drop-frame mode, for the reasons discussed above. But always check with the engineer or camera operator just to be sure.
Video records at 29.97 fps, so that would be the correct timecode
speed for your audio recorder.
If you are recording audio for a film shoot, the settings are
Film editors generally (but not always) prefer to use non-drop timecode for keeping track of frames, since it eliminates confusion when converting from edited video back to film. So unless otherwise instructed, use non-drop timecode on film
The correct frame rate for recording audio that will be sync'd to
film is 30 fps, irregardless of whether the film camera is running at 24fps or 30fps! The reason for this is that the audio does not have to correspond to the film speed but rather to the video speed, since the editing is being done in video!
When the film is transferred to video (for editing), it is slowed
down by one tenth of one percent. Film shot at 24fps ends up at the equivalent film speed of 23.97 on the video monitor. Film shot at 30fps film sped ends up at 29.97 film speed in video.
For audio to remain in sync, it must be slowed down by the same
percentage. So if we record audio on the set at 30fps timecode, and then transfer it into the edit system at 29.97 fps (which also happens to be video sync) — the audio will end up in perfect sync with the picture.
When the editing is completed, the audio will be speeded up from 29.97 to 30 in order to match back up with the actual film for creating release prints.
But that is not your concern as a Production Mixer. The post production people and the film labs deal with that issue.
Therefore, unless instructed otherwise, the industry standard for
production sound (film shoot) is 30fps Non-drop.
If anyone tells you to use a different setting, make sure that you
get it in writing and witnessed! That way you cannot be held accountable for problems that may be encountered by the production company later on.
There are five different running modes for generating timecode.
The most basic setting is Free Run / Time of Day. That means that
the internal timecode generator is like a clock, denoting the actual time of day. The clock runs continuously, whether the tape is recording or not. This is convenient setting to use, since anyone who needs to note the timecode numbers only has to gaze at his or her own wristwatch. Timecode errors between the slate and the recorders are obvious and easy to detect. (It is interesting to note that some studios and producers object to time of day timecode because it is a permanent record substantiating overtime claims by
The next most common setting is Free Run / User Set. This is similar
to the above setting, except that the starting time for the TC generator is chosen by the user and does not correlate to actual time of day. Commonly, the Hours digits are used to signify sound roll number. Spare wristwatches can be reset to match the timecode generator, if you have people on the set who need to keep track of the code.
Record Run timecode means that the generator stops when the tape does. Numbers increment during the Record mode, but remain frozen in time during pause or stop. The elapsed timecode is sort of like a tape counter, and is an indication of how many minutes have been recorded. Users often use the Hours digits to indicate reel number, rather than an "hour" of the day.
The last two "modes" are External and Jam-Sync.
External refers to continuously reading timecode from an External
source and re-generating it onto the tape. If the External code should stop or be intermittant, the code being recorded would also be in error. (Many recorders are programmed with a self-protection that would automatically jam-sync to the last good code and begin generating new code in the absense of the External code.)
Jam-Sync (a.k.a. Set from External) means that the recorder synchronizes its internal timecode generator to match the starting numbers from an External source. When the connection to External is released, the internal code will keep
in step with the external timecode source for a few hours or longer, depending on the accuracy of the timecode generators in question (recorder and source). Of course, Jam-Sync only makes sense in the Free-Run timecode modes.
There's not much point in recording timecode onto the audio if there is no timecode reference on the picture.
In the case of video, timecode is normally recorded onto the videotape even if audio is recorded separately. Jam-sync all of the timecodes, and that
will take care of basic sync (video word sync is another matter, but engineers take care of that on the big shoots). In addition to jam-sync, use either a traditional or timecode clapstick just for protection.
If there is no timecode being recorded onto the video (i.e. prosumer
camcorders), then you should use a timecode slate so that the editor can line up the picture with the audio.
When shooting in film, a timecode slate should be used for matching
up picture with audio.
The standard of the industry is the timecode slate manufactured by
Denecke, Inc. Older versions of these slates were "dumb" slates and could only
display the timecode being fed into them via a cable. Early on, mixers began using Comtek transmitters and receivers (similar to a radio mic) to send the timecode from the recorder to the slate, thus eliminating the awkward cable.
Soon after, Mike Denecke came out with his portable sync box, which
was a timecode generator that could be attached to the back of the slate. The addition of a self-contained timecode generator makes a slate into a "smart slate".
The sync box could be easily jam-sync'd to the recorder, and could
hold sync for half a day. Of course, since re-jamming the timecode is so simple, mixers seldom wait that long.
The latest version of the Denecke slate has a sync box built-in.
Obviously, when you are jam-syncing a smart slate to the recorder,
you must use a form of Free-Run timecode, since the slate would have no way of knowing when you are in Record or Pause.
When doing sync playback, as in a music video, the timecode slate
needs to display the code of the soundtrack being played back. Therefore, the slate must function as a dumb slate and receive a timecode feed from the recorder (playback source). This feed could come via a connecting cable (awkward!!!) or from a Comtek transmitter system (which is how it is normally done).
The Nagra IV-STC stereo timecode recorder is an analog reel to reel machine. The tape runs at 7 1/2 inches per second while a timecode track is
recorded down the center of the tape. During sync playback, a device known as a sync stripper isolates a sync pulse from out of the timecode signal and sends that signal to the resolver unit, which controls the precise speed of the tape by comparing that signal to a reference signal. Changing the setting (rate) of the timecode will cause a change in the playback speed. For instance, a tape recorded at 30 fps in the field can be resolved at 29.97 fps to slow it down in order to sync with a film-to-video transfer.
But digital recorders work differently. Audio is sampled at a precise speed of 48,000 times per second. Audio is then played back at 48,000 times per second. To change the speed of the audio is not the simple matter of turning the tape reels slightly faster or slower, as in the case of the Nagra. Instead, complex electronic circuitry must be capable of re-sampling the digital audio at a different rate, a task much more complicated than it sounds.
Timecode recorded in digital tends to be cosmetic in nature, and is not used to control the sampling rate of the machine. It is possible in digital to even change the output timecode of a pre-recorded tape without changing the speed nor the original timecode of that tape!
In order to change both the playback speed of the tape as well as
the timecode, two tasks must occur. The recorder must alter the sampling rate, and a new timecode must be generated, based on the sub-code of the tape. Merely re-setting the sampling rate (on those machines capable) or re-setting the timecode by itself may not achieve both tasks. Each problem needs to be adressed individually in the machine set-up.
Timecode recording techniques are simple in theory, but can get tricky in practice. Do not attempt to take on a timecode recording assignment on
your own without spending a few hours being checked out by an expert on the idiosyncracies of the specific hardware package you are planning to use.