There is a trend in our industry for clients and producers to clamor for the newest and latest technology, regardless of whether or not that technology will really improve the end product.
For example, all of us here at Equipment Emporium and Wilcox Sound Rentals recall with great amusement the enthusiam that a particular music video producer exhibited over using a DAT recorder for sync playback. This is going to be stupendous! Imagine, we’re going to shoot our video with digital playback! It’s going to be hot!
Well, of course we had to contain ourselves from laughter. Since the playback track is only a guide track, and does not ever appear in the finished product, it makes absolutely no musical difference whether one plays back from a standard NAGRA, a DAT or any other sync device. The sound being played back is only so that talent has something to hear on the set.
But this producer was not looking at the technical process of making music videos. Instead, buzz words and appearing trendy was at the forefront of his mind.
Case in point, recording SMPTE timecode on the audio track for shows that will be edited non-linear.
Having SMPTE timecode on the audio track that will match timecode on the picture is nice, but far from absolutely necessary. Considering the expense of purchasing or renting timecode recorders and slates compared to being able to use existing non-timecode equipment, one should definitely explore all of the post-production ramifications before blindly leaping into costly, albeit trendy, production sound decisions.
Did you know that up until only recently, TV shows such as Beverly Hills 90210 did not use SMPTE timecode when recording production sound? All audio was done with the venerable Nagra 4.2, and then transferred to non-linear digital for post. Why? Because it was cheaper to do it that way, and gave them the same results!
Here is what happens when audio is recorded with SMPTE timecode. Timecode is recorded along with production sound, on a Nagra IV-STC stereo recorder or a sophisticated DAT such as the HHB or Fostex PD4. Matching (jam sync’d) timecode may or may not be recorded on the film by means of in-the-camera keycode and an Aaton master clock module. A Denecke slate is filmed at the head of each scene, displaying a visual timecode as well as providing an old fashioned clapstick marker.
In post, the film is transferred to video in the telecine and then digitized into the non-linear editing system. Audio is resolved at the proper speed (slowed down slightly to match the picture slow-down created by telecine) and also digitized into the non-linear system. Using the timecode numbers as a beginning of the scene startmark or line-up reference, the editor performs a series of incomputer audio insert edits to sync up the dailies (matching up the picture and corresponding sync audio) for each take.
Now, examine what happens if no timecode is recorded on the audio during production. Just as before, the picture is loaded into the edit computer. Audio is resolved at the proper speed, and also digitized into the system. In order to sync the dailies, the editor goes to the picture start of the take (clapstick frame) and parks. Audio is advanced to the audio marker (the clapstick impact): and then the mark-in edit points are punched in.
Finding the start mark of the audio without timecode is easy. If one watches the visual waveform of the audio (the optical track), it is rather easy to locate the clapstick because it sticks out like the Washington Monument! With very little practice, an editor can sync dailies almost just as fast as with timecode, and at considerable savings of the production budget.
But without timecode, how does the edit computer keep everything in sync? The same way it always does, by means of its own internal timecode. Since most production timecode is discontinuous, it is only used for negative matching: the actual editing is done with a form of continuous timecode within the system.
It is true that without timecode, we cannot go back to the original production tapes and conform them with the negative for post. But why would we want to or need to? The audio coming out of the non-linear system is digital CD quality or better, far higher quality than we ever got off a Moviola. In the older days of tape splicing, we had to re-transfer and conform the audio in order to correct for all of the damaged sprocket holes, bad splices, and unintentional edits. But since our digital soundtrack is perfect, we do not need to return to the original tapes before moving on to advanced soundtrack building.
The only step a little tricky in this non-timecode audio process is resolving. When using timecode, we normally record on the set at 30 fps non-drop, and then transfer at 29.97 non-drop in order to compensate for the fact that picture is filmed at 24 or 30 fps (film speed) ends up being slowed down to 23.97 or 29.97 fps (film speed) in the telecine in order to be recorded onto videotape (which is 29.97 video speed).
If we use a conventional Nagra recording with a 60Hz sync signal, then we must transfer that audio into the edit computer at 59.94Hz. This can be very easily done by using an external sync box such as the TX-859.94 Crystal or a similar device. Just unplug the crystal jumper plug from the side of the Nagra and plug in the matching connector from the 59.94 external box: the play the Nagra back with resolver engaged as one normally would.
If recording with a conventional DAT recorder, the process is more complicated. Either the DAT recording must be transferred to an analog machine such as a Nagra and resolved as previously described ( either on a timecode Nagra or a 60Hz Nagra); or else the DAT tape must be played back on a special DAT studio machine capable of alternating its sampling rate to perform the required slowing down.
Some of the newer non-linear edit systems offer, or will soon be offering, a software routine whereby the end user can slow down the audio directly during the digitizing input process.
For some applications where a protection or storage copy of the audio is desired, it is possible to transfer the audio from the original Nagra (at the 59.94Hz speed) into a digital recorder such as the ADAT or DA-88 while adding an (arbitrary, not related to the picture) SMPTE timecode track for future identification or locating of audio. At the time that this protection copy of the soundtrack is created, the audio plus timecode is fed into the edit computer. This now provides a timecode for the audio so that the audio can later be conformed or reconstituted in the event of a computer crash of the edited workprint (assuming that a back-up disk is kept of the edit decision list).
The backing up of the audio track may not be necessary for most applications, but is an option that some end users feel more comfortable with.
Finally, it is important to remember that if the project is returning to the film medium (as opposed to being shown as video), one has to speed up the digital soundtrack when transferring out of the edit computer onto mag film. This can be accomplished making an immediate transfer to Nagra (59.94Hz or 29.97 fps non-drop timecode) or digital; and then playing back at the faster frame rate (60HZ or 30 fps non-drop). Or if the facility supports it, the mag recorder could be run at 59.94HZ, so that when the mag is played back at the normal 60HZ the audio will be back in sync with sprocketed picture.