The third UK release of 2005 is the 1971 Jon Pertwee story 'The Claws of Axos', which represents one of our most technically challenging restoration jobs to date...
'The Claws of Axos' is one of the many Pertwee stories which have had some or all of their original PAL videotapes destroyed and episodes subsequently recovered from North America as NTSC standards conversions. In the case of 'The Claws of Axos', episodes one and four were retained on their original PAL transmission tapes, but episodes two and three were recovered as 2" NTSC quad videotapes from TV Ontario in Canada.
This posed something of a problem. The NTSC tapes had originally been produced by the BBC in the mid-seventies using the BBC-designed CO6/508 Advanced Field-Store Converter (or BBC MK2B Standards Converter, as it was officially known - its sister, the MK2A, worked the opposite way, converting NTSC to PAL), an enormous and highly complicated standards converter which was used to convert the 625-line, 50 field-per second PAL video into 525-line, 60 field per-second NTSC video. To do this, it produced NTSC fields which were either down-sampled 625-line fields, or were a mix of two down-sampled 625-line fields. So there was a characteristic juddering of any moving object or any camera pans on the conversion, which was burnt-in to the NTSC video. When converted back to PAL through even a modern standards converter, that juddering is still apparent and in some cases is made even worse by the reconversion process. Something had to be done.
In late 1993, Ralph Montagu and Steve Roberts were invited to a meeting at the BBC's Research & Development department at Kingswood Warren in Surrey, to discuss various archive-related engineering projects. At this meeting they raised the question as to whether it would be possible to somehow 'un-pick' the conversions by utilising knowledge of the original conversion process and mathematically reversing it - a process they dubbed 'Reverse Standards Conversion', or RSC for short. This would benefit not only 'Doctor Who', but several other series such as 'Up Pompeii!', 'Doomwatch' and 'Vanity Fair'. Working purely from a sample of the NTSC conversion, R&D engineer Jim Easterbrook deduced the original conversion process and realised that it would indeed be possible to un-pick it. A demonstration of RSC , using a short clip from 'Up Pompeii' re-converted in software, was screened to an audience at the NFT in 1994 and showed that the technique could indeed bear fruit, although it would need further development and considerable investment in hardware to enable it to happen.
Unfortunately, 1994 was around the time that the BBC started to become rather more accountant-lead and internal accounting became the new way of doing business within the Corporation. Where previously BBC Research & Development could have had some latitude for investigating RSC as a philanthropic project which would benefit the Corporation, now it had to be accounted for and supported by cold, hard cash... which at that time the BBC Archives had precious little of. RSC, although seen to be a very worthy project, was just too expensive to justify at that time. Windmill Road did however find the funds to archive the NTSC quad tapes onto D1 digital videotape for future use.
Ralph Montagu continued to keep the idea of RSC alive and a few years down the line found an ally within Research Department in the shape of James Insell, a young engineer who happened to be a fan of the show. Insell realised that computing technology had moved on so far that it was possible to implement RSC cheaply as a non-real-time software process rather than in expensive real-time hardware, yet still provide results in a reasonable timescale. He had rather an uphill battle to convince his managers (who, having been bitten once in 1994, were understandably not keen to expend any more effort on the project), but, working on the project in his own time, he was supported by Montagu, who was eventually able to win some funding from BBC Worldwide. Insell researched the original documentation of the converter and suggested solutions to some of the remaining technical issuesbefore Jim Easterbrook was brought back in to develop his algorithm into a useable piece of software - and RSC finally arrived as a functioning process in 2004.
RSC currently runs as a Windows executable on an AMD Athlon MP 1800+ PC with a serial digital video input and output card. The NTSC episode on D1 is captured into the machine as a standard headerless video file in 525/60, with a separate audio file. The RSC software is then run on this file, taking around five hours to process a 25-minute episode, producing another headerless video file, this time in 625/50 format. This file is then played back out of the PC along with the originally-captured audio file (no audio processing is required as the length of the video remains constant) and recorded to PAL Digital Betacam.
At first sight, a raw RSC episode is both a revelation and a disappointment. The revelation is the motion rendition - instead of the jerky, juddery movement one is used to seeing on these episodes, motion is smooth and uniform, for all intents and purposes indistinguishable from normal PAL video. The disappointment is the amount of noise in the picture and the twinkling of horizontal detail due to the line interpolation losses. Noise is a major problem with RSC - it pumps in and out sequentially depending on how much processing had to be done to extract a PAL field from the NTSC video. However, after put through a somewhat aggressive noise reduction process and some vertical filtering, the results are much more acceptable and are really quite astonishing.
'The Claws of Axos' is being project managed by Ralph Montagu, who will hopefully be able to provide more detail on the specifics of the restoration shortly.
Mark Ayres has been responsible for looking after the audio side of things and here he talks about the processes involved in both restoring the soundtrack and mixing the commentary...
"The remastering of the soundtrack of "The Claws of Axos" threw up some interesting problems. As we know, episodes 1 & 4 still exist in their original PAL UK broadcast form, whilst episodes 2 & 3 only survive as NTSC conversions which have been subjected to our new "Reverse Standards Conversion" process. In addition, there is a reel of unedited studio footage of some scenes from episodes 1 & 2.
In order to keep the quality as high as possible, project coordinator Ralph Montagu had re-laid the studio scenes into the first two episodes, and the audio was supplied to me "split-tracked" - the original episode sound on one track, and the replaced shots "checkerboarded" across another two. My first job (after the standard replacement of opening and closing titles music) was to conform these edits to create a new master soundtrack. Differing noise-reduction levels were required on the different sources, and subtle equalisation evened out any multi-generation tonal differences. Some sequences could be taken entirely from the studio tapes, others had to use the final master track where music, effects, or radio chatter had been added. The transitions needed to be carefully timed, especially where music comes in or goes out on an edit.
Clicks and pops were removed as is standard, and a few troublesome edits were eased.
At 3 minutes and 17 seconds into part one, there is a scene where Jo and Bill Filer walk in on a meeting between the Doctor, The Brigadier, and Mr Chinn. This was an edit point in the original, and there is no sound of the door opening - we cut to it once it has opened. This had always troubled me as a sound editor. On this master, having the studio sequence available has enabled me to do an overlap edit so that we can now hear the door opening before the cut. It's subtle, but it sounds better!
Towards the end of episode one, after the Master has introduced himself to Filer in the Axon ship, we cut back to the Axonite demonstration. I had always wondered why there was a very uncomfortable sound dip at this point - and boosting the sound to counteract it exposed the reason: a very loud cough from a member of the studio crew. This cough has now been removed, and the sound dip smoothed out, something that would not have been possible without a lot of effort in 1971. Similarly, at the start of episode 2 I have slightly re-laid the cliff-hanger reprise so as to disguise two very poor audio edits (it sounds as if they actually tried to copy the cliffhanger ending of episode 1 to the beginning of episode 2 to save time, but left the episode 2 audio faded up as well - this is now not an issue!).
Episodes 2 and 3 presented their own problems, being sourced from the NTSC transfer. Both are slightly lacking in high frequencies, and this has been counteracted with some equalisation as far as possible. More troubling is that episode 3 is very "crushed", especially in scenes set in the reactor control room - there is a low bass throbbing sound effect in these scenes (courtesy of Brian Hodgson and the BBC Radiophonic Workshop) that is mixed quite loud - it obviously hit the compressors quite heavily during recording and the problem was exacerbated in the NTSC transfer. Again, this has been treated a little bit to ease the worst, but it is still not perfect, and never will be.
Levels in general have been evened out so as to be more consistent across the four parts. On occasion, the audio does peak quite loudly on these episodes, and would have caught the limiter on transmission. Here, though, except in a couple of instances which were particularly troublesome, I have tended to leave the peaks unconstrained - DVD can take the level, and it's nice to have a bit of dynamic range!
The remastered episodes were bounced to mono AIFF audio files and burnt to a DVD-R with the commentary tracks. This DVD-R was sent to the picture team, where the files were dropped onto the timeline in Final Cut Pro alongside the remastered pictures for playing out to the final DigiBeta master tape.
Commentary mixing also came to me, and I had thought it would be quite a quick job as the commentary had been recorded to a synchronised DAT cassette at Television Centre.
Of course, what I had forgotten was that the commentary was recorded last year, to pre-restoration pictures. Hence the sync drifted widely where duff frames had been removed from the new master. Trying to resync this manually would be a recipe for migraine, so editor Dave Hawley supplied me with his EDLs (Edit Decision Lists), and two hours later I had the thing synchronised properly.
It was then that another problem revealed itself. The commentary was recorded to DAT using roll-back and drop-in where corrections or re-takes were required. Sadly, there seemed to be a problem with these master DATs, which all exhibited frequent digital distortion and "zipper" noise, and there was one total dropout. Luckily, they had also run a continuous snoop tape, and this was fine - so I suspect that either (a) the main master DAT machine was faulty or (b) it was running out of digital sync though still chase-locked to timecode. Either way, I then had to conform that snoop tape to my previous re-sync - another couple of hours. This proved beneficial in a couple of areas: firstly I was able to improve on a couple of the drop-ins, and secondly I was able to retrieve a couple of nice comments which had otherwise been erased and which, with careful editing, could now be accommodated.
The final commentary tracks were bounced to stereo AIFF files for final synchronisation and copying to the master DigiBeta.
The extras for this release include a commentary with stars Katy Manning, Richard Franklin and the show's producer, Barry Letts. Director Michael Ferguson is interviewed about his work on this and other Doctor Who stories in the latest of John Kelly's Directing Who featurettes. Who Was Doctor Axon? takes a look behind the scenes at Kingswood Warren, home of BBC R&D, in a ten minute featurette presented by Jack Pizzey, who spans five decades to bring us the story of standards conversion and RSC... and the mysterious Doctor Peter Axon! Richard Bignell's latest installment in his popular Now & Then series sees him on location at Dungeness, in a featurette narrated by Katy Manning. Richard has also cut together a fascinating look behind the scenes on the studio block for the story (still called 'The Vampire from Space' at that point!), which includes an optional subtitle information track. Rounding off the disc are the usual photo gallery, information text and maybe even the odd Easter Egg!
Copyright Steve Roberts, Mark Ayres, 14 February 2005