8. Magic Roundabout


This period saw the installation of a circular studio based on the Yamaha DMP7 mixer and the Akai DP3200 audio switching matrix. As in many situations, the success of this scheme was down to a mixture of being in the right place at the right time and good project management.

The Mixing Element

The Yamaha DMP7 was first demonstrated in January of 1987. This small and very robust eight-channel mixer had surprisingly few controls and was entirely digital. Each input signal was fed into an analogue to digital converter (ADC), processed and mixed with signals from the other channels. The final stereo output was then converted into analogue form by a digital to analogue converter (DAC).

Unlike most mixers, the DMP7 included numerous sound effects. Better still, almost every mixer parameter could be controlled via MIDI. Furthermore, several mixers could be ‘chained’ together via a special Yamaha digital interface. This provided a feed from the main fader of one mixer that could be fed into the main fader of another mixer. Although, this meant that the operation of output faders on several linked mixers could be confusing, no input channels were used for the links.

Subsequent tests showed that six mixers connected digitally gave around -60 dB of noise, which was just about acceptable. But if analogue links were used, this worsened to around -50 dB. The device contained two microprocessors running at 4 MHz, as well as proprietary 23-bit digital signal processing (DSP) chips. The faders were motorised, allowing them to be controlled by the user or by the device itself. The faders employed a further set of low-resolution ADCs to provide a servo action.

The Switching Element

In July, whilst visiting the British Music Fair at Olympia, Brian Hodgson saw the Akai DP3200 audio switching matrix. This device, which had 32 inputs and outputs, was certainly more elegant than the jackfields used in the department’s existing studios. When tested at the Workshop in November, it was found to have an appalling control unit with its own display screen. However, up to four matrixes could be operated via the controller, providing a total of 128 inputs and outputs. This number fell to 96 if eight ‘tieline’ circuits were used between them, with ‘loop’ tielines between the first and last matrix.

Sadly, the matrix couldn’t mix two sources to one destination. Also, there was a 6 dB drop in output level when feeding unbalanced destinations, although this could be fixed by changing numerous internal links: the solution was to move these links and use the matrix with unbalanced circuits.

Further investigation by Mark Wilson revealed that the 4-pole XLR control connector on each matrix was really a two-way MIDI circuit. This discovery was a vital key: it meant that a new studio could be developed in which the mixing and destinations of audio signals could be controlled entirely by MIDI.

Later, in May of 1988, Ray Riley found that simple modifications allowed the DP3200 to control MIDI signals. Unfortunately, subsequent developments didn’t see his work developed.

Cogitation and Plywood

Considerable, and sometimes heated, discussions now took place between the author, Mark Wilson and Peter Howell regarding the new Studio F. This was to be located on the site of the old V8 (Film Unit Studio) and V7 (Projection Room), but there was uncertainty as to whether a 40-channel Soundcraft desk or DMP7 mixers should be used. Employing DMP7s with a Soundcraft Series 800 looked horribly like ‘mixed technology’ and certainly didn’t appeal to the author.

In the autumn of 1997, Peter tested a DMP7 in Studio B. He suggested that several DMP7s could be used to create the new studio, each assigned to a specific ‘area’ of operations. This idea then expanded into an arrangement of three DMP7s, each with associated equipment in the same vertical plane.

In the meantime, Mark Wilson was experimenting with HyperCard, a curious Mac program often used by Apple enthusiasts for unusual chores. This had a unique ability to create screen displays or ‘cards’ with pre-programmed buttons and boxes. Such cards could then be bolted together and used with other proprietary software. HyperCard also included its own language and could handle MIDI data, offering a real opportunity to control both DMP7 mixers and DP3200 matrixes.

Having obtained a budget of £50,000, the scheme’s proponents went ahead with a ‘lash-up’ where ‘bugs’ in the concept could be ironed out. This temporary installation would be in Room 11 (Studio H), the contents of which had been cleared during the previous summer. Early in 1988, several large segmental pieces of thick plywood arrived. These were arranged on top of a number of filing cabinets, together with large blocks of wood, to create a two-level circular console.

The installation eventually consisted of an Akai matrix and original controller, six DMP7 mixers and a Mac II with an associated HD40 external drive. A special rack was constructed to hold the two main DMP7 mixers at the correct angle. Five MIDI mergers allowed real-time data to be recorded from the mixers whilst a MIDI Thru unit (effectively a MIDI distribution amplifier) fed MIDI from the computer to the mixers. The Mac II had a colour monitor, although the only thing that appeared in the 16 shades of colour was the Apple logo in the menubar. Fortunately, Peter Howell didn’t like colour anyway!

With everything under MIDI control, each synthesiser didn’t need a musical keyboard. In fact, keyboards would be difficult to fit in the console. Thankfully, many instruments came in convenient rack-mounting boxes. This meant that the Yamaha KX88, a full-size MIDI keyboard with a ‘pitch bend’ wheel and sockets for foot pedals, could be used as the master keyboard for the new studio.

During the tests it became clear that extra equipment, such as a Yamaha DEQ7 equaliser and a MLA7 microphone amplifier would be essential. In addition, the author had to create some sort of ‘monitoring system’ for switching loudspeakers and meters between various sound sources.


As usual, the problems were in the details: in particular, the aesthetic considerations of the console design. In one brainstorming session, Mark proposed that the cables of the desk could be exposed. When that didn’t go down too well, he suggested that they could be covered by curtains!

Fortunately for all, the finer aspects were developed by a friend of Mark, a furniture designer from Brixton. This was Jeremy Quinn, who had formerly worked in antiques, and, coincidentally, had gone to the same furniture-makers college as Steve Marshall (who was later attached to the Workshop as a composer). His ideas were very elegant: the console would be circular within but with a polygonal outer shape that accepted removable covers. These covers, attached by pairs of latches that could be released using a standard ‘carriage key’, gave the illusion of a circular shape and hid the cabling. Within the circle there were to be several cupboards (with doors) for various technical equipment.

The total cost would be £4,600. A top view of the design is shown below, as originally equipped.

The five ‘input’ mixers, working in a clockwise order around the desk, were assigned to synthesisers, S550 samplers, TX816 synthesiser, drum machines and the multitrack respectively. Most of these devices were located in a mini-rack positioned above the appropriate mixer.

Room with a View

By the spring of 1988, the Film Unit areas were cleared and the old walls on the site were demolished. The conduits for studio lighting were fitted to the original ceiling and then a new suspended ceiling was fitted. This consisted of a 600 mm ‘grid’ of acoustic tiles, ‘clipped’ into place to avoid vibration. To improve the acoustics, a six-inch layer of ‘bagged’ fibreglass was placed above the grid.

The author was very successful with the acoustic treatment of Radiophonic Studios. Traditionally, the BBC covered whole walls with acoustic absorbers: these were boxes of plywood with ‘pegboard’ fronts and an internal matrix-like structure. They came in two depths (although the ‘shallow’ version was ineffective) and with ‘few holes’ or ‘many holes’ on the fronts. In the early days, Alan Stokes advised that boxes positioned above or below the level of loudspeakers or the listener’s ears were pointless. By empirical methods, the author also discovered that small groups of boxes, with a narrow area of exposed wall between, were often as effective as a solid mass of boxes. In fact, the exposed wall and sides of boxes seemed to randomise the acoustic reflections and thus reduce room resonance.

The trick was to avoid complete symmetry. Therefore, a group of four boxes usually contained a pair of ‘few holes’ and a pair of ‘many holes’ boxes, with the same type of box on the diagonals. And a group of boxes on a facing wall would have the opposite kind of boxes on the diagonals. And, if possible, any gaps in the boxes on one wall faced a group of boxes on the opposite side of the studio. However, near the loudspeakers the boxes were usually arranged in symmetrical groups. John O’Connor of Cork Insulation (Woolwich) delivered seventy acoustic boxes for Studio F. This chap was a bit of a character: he used to play at the Prince of Orange ‘jazz’ pub in Rotherhithe.

Studio F was a rather irregular shape and included a structural column in the middle of the room, as well as beams in the ceiling and an iron strut in the northwest corner. Fortunately, there was sufficient space to fit the circular console between the column and the north wall. This left a kind of ‘relaxation’ area in the section nearest to the entrance door. The final layout of the room is shown below.

Getting Around

During the summer of 1988, Ray Riley created new panels for the back of each DP3200 matrix. This contained 64 two-pole quarter-inch jack sockets, an exceptionally tight fit for this number of jacks and the space within the console. He also constructed the mini-racks that would contain the equipment above each DMP7. These were designed for standard 19-inch devices but didn’t use standard rack fixing methods. Instead, each had clip-on ‘runners’ that could be positioned at 1U intervals, allowing devices of any size to be quickly installed or removed, whilst non-standard equipment could be fitted by inserting a special shelf. In addition, a solid melamine-finished surface was provided at the top of each rack: this could be used for extra devices and prevented foreign objects falling into the rack itself.

The real problem with this studio was the monitoring system. In theory, a very simple switching system could have been used, but this would have needed at lot of wiring at the control unit. For better or worse, the author used a small control box connected to a much larger rack fitted in one of the cupboards. The latter contained various switching and VCA cards, mainly supplied by Canford Audio.

The VCR was to be housed in one of the racks, which meant that a normal infrared (IR) remote control couldn’t be conveniently used. Ray Riley therefore devised a clever solution: he modified the remote control so that the wiring to its buttons came out on a multi-core cable. This was then connected to a set of low-profile membrane switches that were literally glued to the console. The remote control itself was positioned in a rack exactly opposite the VCR, pointing at the VCR itself. The composer could now sit forward to watch the TV screen and comfortably operate the VCR controls.

During this period, Mark Wilson released an engineering version of his HyperCard software for controlling a matrix. This had expanded to include the ability to store programme notes and synthesiser voice listings. Mark also made his initial plan for the matrix connections.

A Yamaha DMP7D soon arrived for evaluation. This was similar to a standard DMP7 but only had digital inputs and outputs. This needed a ‘genlock’ approach to synchronisation of digital data, requiring a separate word clock cable to each device. The time for this machine would come later!

By the autumn of 1988, both engineers were heavily engrossed in cables, involving around 350 jack plugs and hundreds of metres of cable. Most of the wiring was created with 6-core and 8-core Klotz cable. This was known as Monocore, consisting of individually-screened single conductors, each with a separate insulating jacket, and enclosed within an overall outer cover. Apart from wiring to the stereo machines and multitrack, which was wired to phono plugs or XLRs, all connections were made to quarter-inch jack plugs. Any devices with alternative connectors were accommodated by inline adaptors obtained from Canford Audio or the nearest Tandy store!

The console now arrived in excellent fettle. Although created in medium density fibreboard (MDF) it was finished in a superb veneer. Jeremy had ‘limed’ this by painting it with emulsion, scraping off the excess and then applying a final wax polish. He also proceeded to make a combined housing and monitor stand for Syncwriter, as well as an extra desk and table tops.

Various minor details had to be dealt with. For example, a special box was required for the studio’s main fader, the output of which would feed the multitrack and stereo recording machines. A headphone box also had to be connected to the monitoring system, whilst a 12 volt distribution was required for the MIDI mergers and the VCR remote. As for the MIDI connections, the author remembers going down to the shops in the Edgeware Road in search of twenty 5-pole DIN couplers!

By Friday, November 4, 1988, the basic engineering was complete. Room 11 had been cleared, the effects devices were moved into the new studio and a new M16 MIDI matrix and control box had been installed. A few ‘last minute’ changes were required: for example, Mark had to insert a MEP4 MIDI Effects Processor into the MIDI feed to the TX802 in order to sort out its MIDI channels.

The grand opening by David Hatch, Managing Director Radio, took place at 11 am on November 29, 1988. He pushed a button that should have initiated a sequence of music and moving DMP7 faders, but nothing happened. Fortunately, Peter Howell quickly started a DAT machine that was on standby. It turned out that the Mac software had been ‘primed’ for too long and simply ‘timed out’. Anyway, the curtains whizzed open and there was then champagne and cake. George Crowe of Radio Projects looked suitably impressed. Dick and Malcolm’s photos of the event were, as always, embarrassing.

Further refinements continued to emerge. For example, Mark’s HyperCard package had expanded to control many other aspects of the studio’s operation and was now known as Cue Card. As a final touch, a photograph of Studio F was added to the ‘front page’ of his software. By January of 1989, Elizabeth Parker moved into her new studio as a ‘guinea pig’ and was bravely enduring ‘teach-ins’ with Peter Howell. Meanwhile, the old ‘lash-up’ was reconstituted in the Piano Room.

Postscript One: Connections

Towards the end of 1992, several modifications were made to Studio F, including the removal of most of the ‘bridge’ circuits between matrixes, which increased the number of available inputs and outputs. These changes were made possible by the earlier introduction of digital links between the ‘input’ mixers and the ‘auxiliary’ and ‘main’ mixers. The final schematic is shown below.

For simplicity, other wiring to the main loudspeakers, alternative loudspeakers, headphones and audio level meter are omitted in this diagram, as well as the timecode circuits connected to the MIDI interface. Even so, the system still looks horribly complicated, although its inner workings were largely invisible to the user. The inputs and outputs of the four stereo recording devices (2TA: reel-to-reel, 2TB: DAT 1, 2TC: DAT 2 and 2TD: cassette) were connected directly to the monitoring unit. However, they were also wired via a 16-core cable (shown as two cables in the diagram) to Matrix 1.

The monitoring system allowed the audio output of the VCR or a ‘click track’ from a designated drum machine to be blended into the foldback circuit that fed a musician’s headphones. In addition, it provided a recording feed to the multitrack machine and samplers, accommodated timecode circuits to and from the multitrack recorder and accepted an overdub circuit for use during a recording session. The main output fader was also connected via the monitoring unit and was wired into Matrix 1. This matrix also gave access to the eight channel inputs of the Aux Mixer, as well as its ‘effects’ inputs.

Matrix 1 and Matrix 2 accommodated ancillary sources, the main studio output, effects devices, external circuits and the output of the MLA7 microphone amplifier. Matrix 2 was also concerned with percussion devices and the channel inputs and ‘effects’ inputs of Mixer 4. The channel and ‘effects’ inputs of Mixers 1, 2 and 3, as well as their associated sources, were accommodated by Matrix 3. For convenience, the auxiliary outputs of the ‘input’ mixers were combined into a single circuit,whilst the separate auxiliary inputs of these mixers were made available at the appropriate matrixes.

Postscript Two: Circular Resurrection

Following the demise of the Workshop in 1996, Elizabeth Parker, having spent 18 years gaining experience in the business, was determined to set up a private electronic music studio. This was a perfectly viable proposition, since in 1993 she had brought £50,000 of income to the Workshop. As a basis for this new studio, she purchased the original Studio F console for the princely sum of one pound, although being too large to fit in her existing home, it had to bide its time in the garage.

Her initial installation was created by Peter Howell, using those very convenient jackfields that have phono sockets at the back and corresponding jacks at the front. Almost inevitably, it suffered from hum loops and earth-related problems that took Rupert Brun three days to sort out.

In 1999, Elizabeth moved to her purpose-built studio in her new home at Long Ditton, where the old console was fitted-out by Rupert. The main mixer was a Yamaha Promix 01V, with faders assigned to the output of a Yamaha Promix 01 that was used as a submixer. These mixers were linked via a SP/DIF connection, similar in some ways to the paired DMP7s used in the original Studio F.

An Akai DP3200 matrix, controlled by a Peavey PC1600 MIDI Command Station, was retained for monitoring, connected to a Spirit Folio mixer for adjusting levels and equalisation. Timecode was directed to and from various devices using a simple push-button matrix, also designed by Rupert. Sound sources included Proteus 2, Proteus 3 and XL1 (all from Emu), a Roland XV5080, and two Roland S760 samplers. Additional audio processing was provided by a TC Electronic M2000 multi-effects device. The computer was an Apple Macintosh G4 running Mark of the Unicorn’s Performer.

Material generated on the system was usually mastered directly onto DAT, with timecode as necessary, and then transferred to CD-R if required by the customer. Alternatively, material could be created in multitrack form by laying down tracks onto an ADAT recorder with timecode. The resultant tape and machine could then be taken to the post-production studio for a final mix-down.

Unlike the over-engineered Radiophonic circular console, this installation reflected the commercial reality of creating electronic music in a competitive world. It was used by Elizabeth to compose material for all the major television channels, as well as for music used in movie soundtrack of a new version of Monty Python and the Holy Grail.


©Ray White 2001.