The Syncwriter project was one of the most sophisticated endeavours to be met by the Workshop. Its success was largely down to Jonathan Gibbs who joined the department as a composer in 1983. Jon was one of those people who could apply his mind to almost any kind of problem and come up with a viable solution. Without doubt, his classical training at Cambridge had fine-tuned his faculties! To understand the requirement for Syncwriter, the reader needs to appreciate the technicalities of creating music and sound for television, much of which is described in the next three sections.
In the early days, source material came on 16 mm film at the television speed of 25 frames per second (as opposed to 24 frames per second used in the cinema). A copy of the film was usually provided, known as a ‘rush’ or ‘rushes’, having been developed rapidly and without too much concern for quality. There usually wasn’t an original sound track, although one could be recorded on a magnetic track at the edge of the film, positioned where the optical sound track appears on cinema film. In Room 13 (latterly Studio F) the ‘rush’ would be watched on the Prevost 4-plate film viewing machine.
The Prevost had audio heads as well as a standard film-viewing shutter mechanism. Being a 4-plate machine, it could accommodate two lengths of film, one on the upper plates and one on the lower, or one film on the upper plates and a ‘sepmag’ film on the lower plates. Sepmag film looked similar to standard film, but was magnetically-coated in the same way as standard recording tape. The viewing machine could therefore be used for editing together two films or for playing a film with sepmag sound. The latter would be synchronised to the film itself via the sprockets in the media.
The machine’s frame counter was set to zero at the start of the film. Ideally, there would be a visual countdown on the film before the content but this wasn’t always provided. As the film ran, the composer noted down the number on the frame counter at each scene change or cue. The later Steenbeck 6-plate machine incorporated a Roger Sharland Multi-Duty Counter that also gave timings in hours, minutes, seconds and frames, as well as in frames and other various measures.
Having created a cue list, the composer returned to the studio. Fitting the music to the cues was a tricky business, sometimes involving a mixture of instinct, luck, speed-changing and editing, as well as much work with a stopwatch. Indeed, Malcolm Clarke continued to use his stopwatch long after other composers had moved onto more modern methods. There were however some useful tricks available. For example, with a tape running at 7½ inches per second (in/s or IPS), one frame was equal to 0.3 inches of tape. Dick Mills had a special ‘rule’, calibrated in 0.3 inch divisions, bolted to the front of one tape machine: with this he could edit his material down to an exact number of frames.
On completion of the artistic work, the composer could send a quarter-inch tape to the customer ‘on spec’ or the material could be dubbed onto sepmag tape. The first option was perfectly acceptable for recordings under a couple of minutes in length. However, for longer material there was a risk of speed drift, particularly with tape machines of the time. Unfortunately, these employed synchronous motors that were locked to the rather imprecise frequency of the mains supply provided by the National Grid.
In the early years, dubbing to sepmag wasn’t easy: the composer trudged up to the Film Unit, copied the material across and returned to Room 13 to see if it fitted the picture. If a viewing machine were available, the ‘rush’ was loaded onto the upper plates with the sepmag tape on the bottom plates. When the film played the sound would, hopefully, be in perfect synchronisation with the picture. If it wasn’t, the composer could use a 16 mm editing device, incorporating an adhesive tape feeder, to modify the sepmag tape. And it wasn’t unknown for the original film to be edited to fit the sound!
If a film machine wasn’t available, or there were other problems, the composer could use the Acmade Picture Synchroniser, also in Room 13. This was similar to a small viewing machine, but only had a tiny screen in which to see the picture. It allowed the composer to shift the timing of the sepmag tape in relation to the picture film and then edit the sepmag tape accordingly.
By 1983, the Workshop acquired a Sondor Libra MO3 sepmag recorder, fitted into a trolley for use in any studio. This machine could use the centre and edge tracks (the edge track being on the opposite side to the sprockets) as well as a narrow ‘pilot’ track, located on the edge of the film beyond the sprockets. In addition, the Workshop obtained a Steenbeck 4-plate machine to replace the old Prevost.
When video recording became commonplace in BBC Television, the Workshop was supplied with Shibaden half-inch helical scan video recorders, although these were only used for playback. Despite being useful, this technology was a bit of a retrograde step. Unlike film, there weren’t any ‘sprockets’ to lock the sound and picture together and the machines didn’t have good speed stability.
To give some kind of reference, video tapes were often provided with timecode ‘burnt’ into the actual television picture. This displayed the time of recording in hours, minutes, seconds and frames, usually in the form HH:MM:SS:FF. The odds were stacked against the composer since all tapes had to have the burnt-in timecode (they often didn’t), the timecode had to avoid blocking the view of ‘cues’ and all the cues used by the composer had to be visible on the monochrome screen (they often weren’t).
The Video Home System (VHS) format of video cassette recorder (VCR) did, at last, bring some relief to the department’s composers. The Workshop received its first VHS machines from Radio Rentals. Unfortunately, one of their earliest offerings had an odd feature: you could record onto a second linear audio track but not over the original sound track. But the later Ferguson 3V23, although only a domestic machine, performed superbly, despite being subjected to extremely heavy use.
Having acquired reliable VCRs, the Workshop considered more constructive ways of using timecode. Fortunately, the BBC’s Design Department came up with a timecode reader that accepted SMPTE (longitudinal) timecode and presented it on a large display with seven-segment LED characters.
Initially, the Workshop used one of the stereo VHS audio tracks for SMPTE timecode. Unfortunately, such analogue (linear) tracks weren’t entirely reliable and other options were considered in later years. However, if the VCR was played continuously you could copy timecode to a spare track on a multitrack machine, usually track 8 on an 8-track or track 16 on a 16-track.
So the composer now had a form of ‘sprocket’ that locked the multitrack to the time of the source material. The timing of any new sound or music laid onto the tape was assured, even if the machine wasn’t running at exactly the correct speed. Sadly, there was no such ‘sprocket’ to ensure the timing of the final work when mixed down onto quarter-inch tape. As it happens, most machines now utilised a tuned circuit as a speed reference, ensuring ‘frame accuracy’ over several minutes of playback.
All the Workshop’s studios were eventually equipped with BBC timecode readers, allowing the timecode from a VCR or multitrack to be easily observed by the composer. As usual, Murphy’s Law intervened: VHS tapes often arrived without SMPTE timecode, or with it on the wrong channel, with or without a ‘burnt-in’ display. The key problem remained: the VHS recordings required by the Workshop were non-standard and therefore weren’t always supplied.
By the end of 1983, Ray Riley had completed two examples of a new device known as the Timecode Memory Unit (TMU). This accepted timecode data from a standard BBC timecode reader and used it to produce other signals, including an audio output as a metronome ‘click’, produced at regular intervals and based around a count of SMPTE frames. It also could produce a ‘bleep’ audio output or start a tape machine at a time specified by several ‘push wheels’, each of which could be incremented or decremented for a given digit in the timecode.
To make this possible, a standard BBC timecode reader was fitted with a multiway cable and a 35-way ‘D’ connector. This was wired to the binary-coded decimal (BCD) signals that drove the reader’s display whilst the plug engaged with the TMU. The ‘D’ connector was also wired to the input of the timecode reader and to the click, bleep and tape start signals. Later, studios were equipped with two 35-way connectors, one for the timecode reader and another for the TMU (and eventually Syncwriter) with the audio and tape start connections extended into the main studio wiring.
To put the record straight, the author must point out that his photograph on page 53 of the book The Radiophonic Workshop: The First 25 Years involves a degree of artistic licence. The printed circuit board layout and components on display are for the TMU, all worked on by Ray Riley!
Although the TMU was useful, it didn’t solve the problem of relating a cue list to the time displayed on the timecode reader. The answer came with the BBC Model B Microcomputer, later replaced by the B Plus and Master 128 models. The BBC Micro was unique in many ways: firstly, it had a 1 MHz Bus connection for direct access to the processor’s data and address line. Secondly, it could be fitted with a customised EPROM that would override the machine’s startup software, replacing the appearance of the machine with something entirely new. Finally, in the realms of software, it allowed the use of *FX routines for special operations and also permitted ‘illicit’ but fast operations via machine code
Jon’s plan was this: the composer would use the BBC Micro to create a cue list that would appear on the screen as a ‘timeline’ that could be viewed at various magnifications. When used with external hardware, a ‘cursor’ would move along the timeline in step with timecode, switching into new ‘pages’ if a closeup view was in use. Hence a cue could be anticipated before actually getting to it.
The prototype hardware was assembled using a 3U rack fitted out with 64 and 96-way DIN connectors and a matching series of off-the-shelf Cube cards. The 1 MHz Bus connection was vital, since this gave the programmer direct control of the external hardware. However, the length of associated wiring was best kept to 500 mm or less. Appropriate hardware was included for receiving the BCD data from a BBC timecode reader, also circuitry for producing the standard click, bleep and tape start signals. In addition, the hardware was configured to produce a range of audio ‘clock’ frequencies for driving an analogue sequencer and was fitted with MIDI In, Thru and Out circuits.
Once the prototypes had been proven, the author finalised the design and put the job of creating a final printed circuit board (PCB) to an outside company. The completed card was installed in a 1U case, complete with two power supplies, one giving +5 volts for the logic, the other +/-15 volts for the analogue circuitry. The rear of the unit had a 35-way ‘D’ connector for the timecode reader, a series of jacks for the audio signals and three DIN sockets for MIDI, all mounted directly onto the PCB.
The Syncwriter package was completed by taking the 1U case, with a second identical case containing two 5¼ inch floppy disk drives mounted side-by-side, and putting them into a 2U rack. This was then slung beneath a shelf on a Unicol stand containing the BBC Micro and monitor screen.
Both floppy disk drives were required for Syncwriter operation: one for the composer’s own data and the other for the workings of Syncwriter’s software. When Jon’s program was complete it was transferred to an EPROM that went into the ‘number one’ ROM slot in the BBC Micro. The machine’s normal startup ROM was then moved to another location. This meant that the computer started up as a dedicated Syncwriter machine, although it could be switched to standard operation if required.
To create the EPROMs, the engineering workshop’s BBC Micro was connected to an EPROM programmer via its serial port. This was also used for ‘backing up’ or updating EPROMs from other items of equipment. An ultra-violet EPROM eraser was used to clear the contents from EPROMs.
At the last moment, Syncwriter was made to read timecode independently, dispensing with the need for a separate BBC timecode reader. Ray Riley checked out a timecode reader ‘chip’ and the outside PCB company then designed a special ‘daughter card’. This plugged into integrated circuit (IC) sockets on the motherboard and the original ICs were inserted into the new card. Wiring was then added to the 35-way ‘D’ connector to receive the incoming timecode signal.
Syncwriter was a great success and an indispensable aid: Jonathan continued to update and refine the software, even after he’d left the Workshop. By 1986 four extra Syncwriter units were built, making total of seven. At the same time, the two old prototypes were scrapped and the software updated to work on the B Plus computer. In November of 1987, Jon persuaded Syncwriter to produce MIDI Timecode (MTC) and two months later it could produce MTC and MIDI Clocks at the same time.
By 1990, Macintosh software had overtaken Syncwriter and the units went to Redundant Plant. Immediately this happened, one of Roger’s clients noticed the lack of a timecode-triggered tape start! Long after everyone had forgotten Syncwriter, the author received a phone call from an engineer in the BBC Regions who had rescued a scrapped unit: the author’s attempted removal of the writing on the front panel had failed. To keep them happy, Jonathan generously sent them the latest software and the author provide them with circuit diagrams. No more was heard: perhaps they’re still working on it!
The Electronic Baton was a separate project to Syncwriter, although also related to timing. It was designed to help a ‘real’ musician perform at a required tempo. Traditionally, a musician was fed a ‘click track’ via headphones, either played from a multitrack recorder or created by a drum machine. Understandably, many musicians didn’t like the headphones or found the clicks intrusive. Therefore the Electronic Baton was created in an attempt to emulate the actions of a conductor’s baton.
The Baton consisted of three parts: a sending unit, a receiving unit and a display box that could be attached to a microphone stand. The sending unit was conveniently connected to the area containing the receiving unit, along with the musician, via normal audio tielines. The sender had four buttons, arranged in the shape of the operator’s hand. These buttons were Hall-effect devices, similar to those used in high-quality computer keyboards. Within the unit were four ‘555’ oscillator cards: whenever the operator pressed a key, a tone of an appropriate frequency appeared at the sender’s output.
The receiving unit contained a tone decoder ‘chip’ that only recognised the distinct frequencies produced by the sender. The decoder produced four separate output signals, wired to the display box via a 7-way XLR connector. The display box contained four light-emitting diode (LED) arrays, arranged in a pattern similar to a number ‘4’, following the movement of a traditional baton.
Operating the buttons on the Baton was a bit tricky on the fingers. To get around the problem, Peter Howell used the Fairlight CMI to generate a sequence of suitable tones for playing directly into the decoder. In fact, any source of correct frequency tones could be have been used.
This wasn’t the Workshop’s first use of tone signals for control purposes. Years earlier, Dave Webb created such a system to control a remote Studer A80 tape machine. It consisted of a sender box with buttons and a decoder unit that plugged into the machine’s remote control socket. Unfortunately, since the audio signals only went one way, no indicator lamps could be provided on the control box.
©Ray White 2001.