Using Timecodes (Part 1)
An Introduction To Timecode Synchronisation
The release of the new Fostex synchronisers reinforces the importance of timecode in today's recording process. So, to help you familiarise yourself with their applications, Francis Rumsey begins a new series that should tell you all you need to know about timecodes.
Until recently, the complications of working with timecode had not permeated the semi-professional audio market to any great extent. Now that manufacturers like Fostex have begun to introduce astoundingly cheap synchronisers, it would seem that it is time for some advice about how timecode can be used, and the many pitfalls which may be encountered!
In all but the smallest operations, the use of timecode in video recording is universal, providing an absolute positional reference in the recording and a means of locating one frame of video precisely. Timecode relates more easily to video than to analogue audio, because it is a frame-counting code; in other words, the timecode recorded on a video tape is locked to the frame rate of the video and counts sequentially at that rate.
Analogue audio recorders do not use a segmented format because the sounds are recorded in a continuous magnetisation pattern along the tape, rather than in helical scans as on a video tape. This means that any timecode recorded on the audio tape will be imposing a 'framed' pattern on the recording, providing a means of locking it to other 'framed' formats.
Timecode is used in video editing to locate edit points and relevant scenes, as well as to synchronise the VTRs during the preroll (the period preceding the edit point). It is also widely used in multitrack post-production of videos, for locking audio machines to the master picture so that sound tracks can be laid down in synchronisation.
Twenty-four track audio studios often use timecode and a synchroniser to lock two multitracks together, providing them with forty-six tracks of audio capacity.
There are a number of terms which need to be defined before proceeding in any detail, so here's a short glossary:
MASTER: The source of timecode to which other machines in the system are referenced. In a video post-production suite this will normally be the VTR, for reasons which will later be explained. The master does not normally receive speed control information from the synchroniser.
SLAVE: One of a number of machines which are to be locked to the master by changing their speed so that they run without drifting in offset.
SYNCHRONISER: The unit which reads timecode from the master and slaves, compares the positions and controls the slaves so that they stay in the right place. Synchronisers vary enormously in the number of machines to be controlled and the facilities offered.
CONTROLLER: Some synchronisers are just 'black boxes' which lock two machines together. In order to control these parts of the system intelligently, a controller is used which interfaces with a number of synchronisers to order their actions. This idea is becoming more widely used, as it allows for considerable flexibility in system configuration.
INTERFACE: Just as people from different countries speak different languages, so do machines from different manufacturers. A synchroniser will need to be equipped with a number of possible interfaces in order to allow correct connection to various machines. Just as musical instrument manufacturers are standardising on the MIDI interface, so audio and video equipment manufacturers are aiming towards a universal remote control interface protocol.
OFFSET: It is highly possible that the machines to be locked will not have the same absolute timecode position (one might be at one hour, and the other at ten hours). An offset must therefore be introduced by the controller to ensure that the machines lock at this distance, otherwise the slave would simply run off the end of its tape trying to locate to the master position.
Synchronisation involves remote control of the tape transport and capstan speed control. This is in order that the synchroniser can roughly locate the tape to within a second or two of the master, followed by placing the transport in 'Play' mode and varying the capstan speed to trim the lock. It is important that the tape speed is not varied audibly after the initial lock period, as this would ruin the production, although it will be necessary to maintain the long-term offset so that the slave does not drift relative to the master. Designers go to considerable lengths to optimise the lock parameters so that tight enough lock is maintained without transferring fast speed variations in the master to the slaves.
The locking of, say, two multitrack tape recorders requires that very accurate lock conditions are maintained: of the order of fifty microseconds between transports. This is much more stringent than on audio to video lock, because any offset changes between the two recorders will result in phase changes between audio tracks, which may be undesirable. For this reason, it is not recommended that a stereo pair of tracks be recorded between two machines (ie. Left track on machine one and right track on machine two), or that any phase-sensitive pairs are treated likewise. Most good synchronisers specify lock conditions to an accuracy of +/- fifty microseconds, which is one cycle of a twenty kilohertz tone, so you can see that it is best to split different types of material sensibly between machines; for example, rhythm tracks on one and melodic/harmony tracks on the other.
Locking the two machines together requires that one track on each tape must be striped with timecode, which will obviously reduce the audio track capacity, and it's important to realise this because two four-track recorders would only create six tracks when locked together. Another point to consider on these lines is that timecode crosstalks, very badly between tracks, as it is right in the most sensitive frequency range of our hearing, and this may impose the added restriction of a guard band (an empty tape track) left between the timecode track and the next audio recording track.
This will depend greatly on the performance of the tape recorder, and the only way of finding out is to try it.
1 Check for crosstalk between timecode and adjacent audio tracks. If necessary, leave one track free as a guard band.
2 Record timecode on one of the outer edges of the tape, which is either the lowest numbered track or the highest. Convention usually results in the highest numbered track (eg. track sixteen on a sixteen-track recorder). In this way crosstalk will only go one way, ie. onto track fifteen, and it also saves putting audio on an edge track where it could be corrupted by edge damage.
3 Try to record something on the track next to the code which can be equalised to reduce the timecode crosstalk (eg. don't record violins!). But don't record high level bass transients on the adjacent track, as these will tend to spread on the tape which may corrupt the timecode.
4 Experiment with the recording level of the code. If it is too high, it will suffer distortion and the shape will become difficult to decode. If it is too low then it may be below the decoder threshold in the timecode reader. Aim for a recording level about three to six decibels below the reference zero on the tape machine's meters, and if you can adjust such things as bias and equalisation, reduce the bias a bit and wind down the high frequency EQ to obtain the best square wave on the replay side (off tape) when viewed on an oscilloscope.
5 Do not edit the tape physically once continuous code has been recorded. Discontinuous timecode really confuses synchronisers and will give you all sorts of problems. If you intend to do any editing, do it first, and then record the timecode afterwards.
6 Be careful at what speed the timecode is recorded: many people forget that they have changed the replay speed of the tape machine since they recorded the code!
7 After recording the code, put the relevant track in a 'Record Safe' mode to prevent accidental erasure of the code track.
When your timecode track is replayed, it will have a fundamental frequency in the region of 2-4kHz (although the harmonics needed to reproduce the square wave shape will be higher) at normal play speed. If the tape is travelling at high speed, however, perhaps in a search or rewind mode, the bandwidth of the replay amplifiers will not be wide enough to replay accurately the shape of the time code wave, as the high frequency harmonics will be rolled off. To get round this problem, most synchronisers count pulses from one of the timer rollers on the tape machine when the tape is in fast wind modes, dividing them by relevant amounts so that a suitable frequency can be obtained. There are problems associated with this in that the synchroniser will not register any discontinuities if they pass during a wind period, but it works satisfactorily most of the time.
The alternative to pulse counting is to modify the replay amplifier on the tape recorder so that its frequency response goes up to, say, 200kHz, and then to use a high speed timecode reader in the synchroniser, although this is a more expensive solution. Some tape recorders with dedicated timecode tracks actually reshape the timecode before outputting it. This makes it suitable for copying directly to another machine, but do not copy timecode straight off tape onto another machine without reshaping first, because it may become too distorted to read.
Next month: Timecode standards, timecodes with video, and editing using a synchroniser.
Feature by Francis Rumsey
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!