Ray Beckett relates his ...

DEVA NOTES

 I’m sure I am not the only one waiting for a viable replacement for R-DAT to appear. My experience with the format goes back to 1988 when I bought the first ‘professional’ DAT recorder to appear in the UK. This was the PCM 2000. It was built like a tank, offered phantom powered mic inputs (noisy), and, a timecode track which recorded, or attempted to record, timecode onto an analogue track that used the guardband between the helical scan area and the edge of the tape. Needless to say, the attempt to record reliable timecode on what was effectively a ‘Dictaphone’ was doomed to failure. Come to think of it now, the attempt to record 16-bit PCM reliably on a 4mm tape running through a scaled down helical scan transport originally designed as a piece of consumer electronics, and to do it day after day in the harsh conditions of film production, was also hopelessly optimistic.

Frankly, I was disappointed. I had just recorded two feature films using the PCM-F1 system. This system seemed even more optimistic in concept - use an encoder to convert 16-bit PCM to a black and white video signal that was then recorded onto a modified Betamax VCR. However, it worked, it weighed a ton, was luggable rather than portable, but it worked. Naively, I had hoped that the system would evolve into a four channel recorder lighter, and possibly with a reliable timecode track as well. Commercial rather than technical pressures won out and we had DAT. Over the years the transport mechanisms were refined and improved, power consumption was reduced, timecode was introduced into the data-stream and mic pre-amps got quieter.

However, we still have a format that is locked into 2 channel recording (except for the Stelladat II) and can chew tape on you unexpectedly, having cooked it first, thereby ruining your day. The number of threads in professional news groups dealing with which tape has the best block error rate, brand A or brand B, with often completely contradictory replies, do not exactly inspire confidence.

The Nagra has always been the benchmark for a purpose-designed reliable location recorder but when the Nagra D was introduced, I was disappointed again. Although a four channel 20-bit format, it is beyond the budget of most freelance sound mixers, and it is, again, luggable rather than portable.

Ken Loach is one of the most respected film-makers in Britain. He has been making films since the 1960s. Over this time he has always used a ‘documentary’ style of filming, shooting in script sequence, often with non-actors in leading roles. His style requires that both on and off camera lines are recorded at all times.

In early October shooting was completed on Ken’s latest film Bread and Roses. I have done five previous films with Ken using either a Nagra 4S T/C backed up with a timecode DAT or two DAT machines. There have been times during the last three films where four track recording would have been very useful.On the last film, My Name Is Joe, I used two DAT recorders running in parallel to get four tracks. This proved problematic as one of the DATs had no timecode and it was necessary to ‘match mods’ to align the tracks in post-production.

I was aware of the Deva through reading the internet news group rec.arts.movies.production.sound, particularly the posts by Jeff Wexler describing it’s use on a feature film. I rented a Mark I Deva and bought a DVD-RAM drive in April this year to test the system. I used it for two music recordings and was impressed by the clarity and openness of the sound and by the simplicity and speed of the back-up to the DVD-RAM disc. I decided that I would use the Deva on the next film, particularly as this was to be shot in Los Angeles and I would have good back-up should problems arise.

I contacted John Coffey at Coffey Sound and arranged to hire a Mark II Deva and three 2.2 Gigabyte removable hard drives. I chose this size as it is a good fit with the DVD-RAM discs I would be using for back-up mirroring. Playback, DVD-RAM back-up and sync tests were done at the transfer facility Sound Deluxe.

Ken Loach edits his films in what is now becoming the ‘old fashioned’ way: Picture and sound are cut on a Steenbeck. Sound is on 35mm Mag with timecode and user bits from the location recorder recorded on the balance stripe and a rough sound mix on the audio track(s). When the picture edit is completed the sound reel is passed through a device which reads jumps in the timecode and produces an edit decision list for the sound conform into a non-linear audio editing system (AudioFile in this case).

The big challenge of working with Ken is the knowledge that what is recorded on the day may well be all that is played out through the cinema loudspeakers. While this is true of many films, it is made more difficult by the fact that he will not add extra sound fx to a scene; being a realist film maker, only sounds that were there during the take will be used. Also, no looping will be done for the same reason.

The first scene of the film involved a re-enactment of an illegal crossing of the US-Mexican border near Tijuana. As Ken always shoots in a documentary style, a portable rig was needed. I had the Deva and my SQN mixer rigged in a shoulder harness. Sound pick-up was by an M/S stereo mic (Schoeps MK8/MK41) decoded to X/Y before recording onto Deva tracks 1 and 2. A radio mic on the leading actress was recorded onto track 3. This was a real test of the system as Ken shot the whole thing in one long take beginning with the characters emerging from the undergrowth, climbing a steep rocky embankment and running 150 yards before hiding again. Camera and sound ran with them. I was concerned about the effect that the jarring from my running was having on the Deva hard drive but no problems occurred.