SMPTE EBU timecode by Phil Rees

Synchronisation and SMPTE timecode (time code) This document is a description (mainly technical, but not requiring extensive prior knowledge) of the SMPTE time-code synchronization system, especially the Bi-Phase Mark "LTC" timecode sync tone which can be recorded onto the audio track of a video tape or onto an audio tape. Other time code variants, such as VITC time-code are also mentioned. Some supporting information on the related subjects of colour television standards and field sequences is included. There is also a concise description of the binary coded decimal (BCD) representation as used in the SMPTE (or EBU) timecode (time code) data word. My original reason for exploring this subject technically was to develop a low cost MIDI tape sync unit, which was a frequent customer request, and a logical addition to our range of electronic music gadgets. Our Time Code tape sync unit, which supported SMPTE (including Drop Frame format) and MIDI Time Code, was called the TS1 and was for over a decade a successful and highly-respected product. Synchronisation systems When you require pieces of audio, video, or music technology equipment (eg. tape recorder and sequencer) to work together, you may need some means to make sure that they play in time with each other. This is called 'synchronisation' or 'synchronization', which gets shortened to 'sync' or even 'synch'. A magnetic tape-recorder track is the region of the tape that is scanned by one recording head element. It can, for example, carry an audio signal, that is an electronic analogue of a stream of sound. In order to use an audio track to record sync information, the information must be encoded into an audio-compatible signal, called a sync tone. An audio tape track on which a sync tone has been recorded is called a 'stripe'. It is desirable that the synchronised devices can join in wherever you start up the sync tone, even if it is not at the beginning of the stripe; this is called 'chasing'. The SMPTE/EBU timecode standard defines the predominant internationally accepted standard for a sync tone and it allows devices to 'chase' or locate to a precise position. In music technology applications, with equipment which is not timecode compatible, you may be able to use a tape sync unit by way of a tempo-relative stripe system. Examples of such systems are traditional FSK (Frequency Shift Key) and the Philip Rees proprietary FSKplus format. The SMPTE/EBU timecode standard Until the advent of video recording, there was mechanical sound to film synchronisation. This mainly relied on sprockets (a row of neat holes) in the film and in special sprocketed recording tape. Relative timing adjustments could be made by slipping sprocket holes. The same sprocket holes were used to maintain synchronisation. Video tape hasn't got any sprocket holes, so when video arrived an electronic equivalent was needed to take the place of mechanical methods of synchronisation. In 1967, The US Society of Motion Picture and Television Engineers introduced that which we call SMPTE ("simpty") time code. The audio sync tone version of SMPTE is called linear or longitudinal time code or LTC. By the way, there are also versions of timecode which can be inserted into a video signal or sent via a MIDI connection. On the former website tvhandbook.com, a Broadcast History Timeline had this entry for the year of 1969: "SMPTE timecode established to end the chaos of incompatible time codes for various editing machines". The original uses of SMPTE timecode include accurate video editing and synchronising film sound-tracks. The timing data in SMPTE takes the form of an eight digit twenty-four hour clock. The count consists of 0 to 59 seconds, 0 to 59 minutes and 0 to 23 hours. The second is subdivided into a number of frames, which may be varied to match the various frame-rates used around the world. The frame-rate is the number of times a second that the picture is updated so as to give the illusion of continuous movement. Frame-rate formats There are four standard frame-rate formats: The SMPTE frame-rate of thirty frames per second (fps) is often used for audio in America. It is used with the Sony 1630 format for CD mastering. It has its origins in the obsolete American mono television standard. The American colour television standard has a slightly different frame-rate of about 29.97 fps. This is accommodated by the SMPTE format known as thirty Drop Frame and is required for video work in America, Japan and generally the 60 Hz (mains frequency), NTSC (television standard) world. The number of frames in each second is not an integer, so an approximation is used; it is based on 30 fps, but two frames counts are dropped (skipped) at the start of every minute, except for every tenth minute (the minutes of every hour in which no frames are dropped are minutes 1, 11, 21, 31, 41 and 51). The EBU (European Broadcasting Union) standard of 25 fps is used throughout Europe, Australia and wherever the mains frequency is 50 Hz and the colour TV system is PAL or SECAM. The remaining rate of 24 fps is required for film work, it is rarely used for audio. In audio post-production, SMPTE has been adopted for machine synchronisation and as a reference of tape position. Essentially, the choice of frame rate for audio work is usually arbitrary. The modulation scheme of SMPTE LTC A bit is a binary digit (which can take only the value 0 or 1). The value of a bit is often represented by the 'off' and 'on' states of an electronic switch. When the sequence of binary digits must pass through a channel designed for analogue audio signals the 'on' and 'off' states ('phase') cannot be reliably distinguished. However, audio frequencies are reproduced faithfully; to exploit this property, SMPTE Longitudinal TimeCode encodes its data into the rate (frequency) of electronic state transitions. The binary value 0 is represented by a single transition at the start (or 'boundary') of the bit-period (or bit 'cell'). The binary value 1 is represented by a two transitions - one at the start and the second in the middle of the bit period. This scheme is called 'Bi-Phase Mark' and closely resembles frequency modulation ('FM'); that is, a stream of binary ones is represented by a burst of audio at double the frequency of a burst of audio which is used to represent a stream of a binary zeros. This straightforward scheme of modulation is amazingly robust and compatible with a range of real world audio channels, including tape recorder tracks. To ensure good performance in audio channels, the rise-time of the waveform is also specified. The data organisation of the SMPTE LTC frame The datastream for each video frame of Longitudinal TimeCode consists of eighty bit-periods (bit cells). At the original American frame-rate of 30 fps, the bit rate would work out to 30 x 80, that is 2400 bits per second. The frequency for a stream of zeros would be 1.2 kHz and for a stream of ones it would be 2.4 kHz. The corresponding bit rate for the European 25 fps frame-rate is 2000 bits per second. in this case, the frequency for a stream of zeros would be 1.0 kHz and for a stream of ones it would be 2.0 kHz. All these frequencies are safely within the audio range, so the SMPTE LTC sync tone waveform can be recorded easily on any half decent audio track. In each frame, twenty six of the eighty bits carry the SMPTE time or 'address', in binary coded decimal. In the diagram above, these Bits are shown as FRAME UNITS, FRAME TENS, SECS UNITS, SECS TENS, MINS UNITS, MINS TENS, HOURS UNITS and HOURS TENS. The BCD digits are loaded 'least significant bit first'. Thirty two bits are assigned as eight groups of four USER BITS, also sometimes called the "Binary Groups". This capacity is generally used to carry extra info such as reel number and date. The User Bits may be allocated howsoever one wishes as long as both Binary Group Flag Bits are cleared. The User Bits do not usually vary in the course of a timecode stream. The last sixteen Bits make up the SYNC WORD. A timecode reader uses these Bits to find the frame boundary, the tape direction, and the bit-rate of the sync tone. The values of these Bits are fixed as 0011 1111 1111 1101 The Bi-Phase Mark Phase Correction Bit bPMPC is Bit 27. This bit may be set or cleared in order that every 80-bit word contains an even number of zeroes. This means that the phase (or logical sense) of the pulse train in every Sync Word will be the same. As the LTC standard evolved, during the 1980's, timecode was being heavily used in 1" video editing. Many of the early processor systems used to synchronize the playback and record machines to an editor required that the timecode Sync Word to be properly timed to the vertical interval (per the spec). However, because of tracking (or other interchange) problems with some tapes, it was necessary to manually track those tapes for optimum playback. At the same time, not all VTRs were equipped with timecode boards, and not all of those that did performed a regeneration of the off-tape timecode (which provides a properly phased and shaped signal to the outside world). As a consequence, as the tracking was varied, the timing relationship of the Sync Word varied as well. Much to the dismay of many editors and tape operators (spoolers), it was learned too late that the timecode readers in the editor could be off by a frame if the tracking was moved too far away from spec. The Ampex VPR-2/2B, the work-horse of the industry at the time, had a timecode waveform display. Using it to verify the position of sync word was complicated by the constant toggling of the polarity of the waveform due to the varying number of zero's per frame. The Bi-Phase Mark Bit was added to make finding the Sync Word visibly easier. By the time the standard was fully adopted, most newer machines had already incorporated the regeneration feature which made the Bi-Phase Mark Bit obsolete. Prior to the standard being adopted,