2005's last DVD release is the extremely popular 'City of Death', which pits Tom Baker's Doctor and Lalla Ward's Romana against the sole survivor of the Jagaroth race in a fast-moving and witty story which features location filming in Paris. 'City of Death' owes much of its success to being co-written by the late, great Douglas Adams and achieved the highest ever viewing figures of any Doctor Who story. Released in the UK and US for the first time on DVD in a lavish two-disc edition, 'City of Death' is bound to end up in many Christmas stockings on both sides of the Atlantic...
'City of Death' was shot in studio on 2" quad videotape and on 16mm film on location, and as such is typical of the production methods employed on Doctor Who around this time - the Paris location not withstanding! The quad transmission masters reside with the NFTVA and as usual the restoration work was done from D3 digital videotape copies, with the work being carried out by the RT's Jonathan Wood.
"As usual the D3s were dubbed to Digital Betacam in component form via the BBC's Transform PAL decoder which gives the best possible source material to work from free of PAL video artefacts. The next stage was to begin processing the pictures in the usual manner, that is a shot-by-shot re-grading (mostly of the film inserts) coupled with an appropriate amount of digital video noise reduction. However this process was revised shortly after the session began...
The story begins with some 35mm model effects footage intercut with studio material on videotape which didn't pose too much of a problem although it was noted that the Quantel picture shake effect used on the cockpit shot during take-off brought in the edges of the active picture to quite a degree. Unforseen though was some instability on the 16mm film inserts which seemed to get worse as the story progressed, so that by the time the 2nd insert with the Doctor and Romana sitting outside a café came up the amount of sideways jitter was quite objectionable. This may have been a fault in camera, dodgy negative stock or a fault during the printing process. Either way, some stabilisation was going to be required and a tape of the film insert material was given to Ian Simpson to process on Furnace running on a Mac G5. It was decided to hold off adding any DVNR to these 16mm inserts at the grading stage in case the instability reduced the effectiveness of the noise reduction, but MKII telecine field-flicker was removed in the usual way by filmising the material through a Snell & Wilcox ARC-100.
The studio material was generally good noise-wise although the recording block for the café interior scenes and the dark cellar shots suffered more than others with the DVNR being increased accordingly. Colour balance was quite good but there was a yellow bias to some shots in the main room at the chateau on occasions, also the odd camera was slightly out of alignment. Playback levels from the D3 were lowered during the dub on some of the video effects shots in the time vortex such as the chicken/egg and the death of the professor as there were high video levels on some channels which would clip on the input to the grading system. There were also differential chroma phase errors built-up during the creation of the analogue effects - these are not evident however on a 4:2:0 system such as DVD.
The stabilised film inserts were returned requiring a small amount of zoom via the ARC to hide the now shifting edges of the picture, with image content remaining stable, although some jitter within whole frames remains as there was an amount of top to bottom warping ("rubber pictures"). After this the inserts could finally be processed with DVNR to reduced grain and fine dirt & sparkle. All four episodes were loaded into the disk based Scratchbox retouching system for manual clean-up of tape drop-outs, line flashes, level bumps, film dirt, sparkle and cement joins. There were a fair number of tape defects amounting to some 300 fixes, while the film de-blobbing work totalled about 1400 manual fixes. Episode recaps were re-inserted from material from the preceding episode (fewer tape generations) and the opening titles we re-made as usual from the master film transfer with the original captions keyed on again."
Whilst Jonathan was working on the pictures, Mark Ayres was getting stuck into the audio...
"The audio remastering for City of Death was always going to be a pretty standard affair. There was a fairly high background noise level, but it was consistent, and therefore fairly easy to remove with some broadband noise reduction. The few electrical clicks were removed with a de-click plugin. Dropouts and bumpy edits (often on film sequences) were repaired, and the theme music replaced as a matter of course, necessitating careful rebuilding of a couple of the episode openings. Levels were fairly consistent, but with some quite high peaks which would have been limited on original transmission - I have left them alone, as the higher dynamic range of DVD can take them. I did reign the level back at a couple of places where dialogue had been pushed unnaturally to lift it over the music, dropping the whole mix a couple of decibels at these points helped to sell the shots much better while keeping the perspective.
There were two places where I made bigger repairs, both in episode two, and both for faults which I had noticed, and which had bugged me, even on first transmission in 1979 (when I was eighteen years old!). Firstly, at about 14:25, where Hermann is demonstrating removal of the glass around the Mona Lisa, there is a cutaway to the Countess. This was a dropped-in shot, and studio noise dipped in and out on the cuts, making this very obvious: I have now continued the studio noise (and the sound of the glass cutter) over the edits. Secondly, at about 22:20, as the Doctor steps into the "renaissance sunshine" in Leonardo's cottage, the grams operator had not stopped the incidental music master after the previous cue, and there was a brief blast of the beginning of the next cue before the fader was pulled down. This has been removed - and was a tricky repair as there is some distinctive birdsong audible at this point."
Unfortunately neither Tom Baker or Lalla Ward were able to take part in the commentary or featurettes for this story, but a full extras package has been put together to complement the main feature. In fact due to a surprise find well into the production of the features, the extras package was rather too full! Not wanting to drop any of the existing features but wary of the possible reduction in picture quality of putting too much material onto one disc, BBC Video agreed with our request to split the package across two discs. This means that the remastered episodes can be presented on a dual-layer disc of their own, ensuring that the quality will be as good as it possibly can be.
The surprise find mentioned above is three reels of Shibaden 1/2" videotape, containing timecoded copies of some of the studio recordings for 'City of Death'. We are very grateful to Jan Vincent-Rudzki for the loan of these tapes, which have never been seen before and which offer a fascinating glimpse into the story's studio sessions. Finding the tapes turned out to be the easy part... Actually getting any playback from them was another story in itself!
The three tapes contained a total of nearly four hours of material from the studio recordings of 22 May and 4 & 5 June 1979. These would have been recorded to allow the director to make offline notes for the editing sessions. Tapes were often recycled but, by this time, VHS was starting to make in-roads and U-matic was also used increasingly as both formats were much more reliable than Shibaden. This is probably why these tapes have survived while studio recordings of much later stories, on VHS, have been wiped over with other material.
We have access to a working Shibaden recorder and planned to make two or three transfers of each tape with different tracking settings, in order to combine the best sections from each pass. Unfortunately things did not go smoothly on the first session. None of the tapes was stable and the JVC timebase corrector we were using was unable to lock onto the picture for more than a few seconds at a time before outputting a few seconds of black. To compound matters, one of the tapes developed a tendency to snag on the heads which resulted in several short sections (of non-unique material, fortunately) being lost but, more seriously, the retention pegs on the recorder's feed hub were sheared off.
At this stage we investigated having the material transferred at a commercial facility, only to discover that the only company able to deal with Shibaden tapes did not currently have a working recorder either. It was tempting to abandon the transfer, but from what we had seen initially, we knew that there was some amazing material on the tapes and we felt that we should make one more attempt to obtain a playback. We were able to replace the broken tape hub using spares from a standby machine and we also discovered to our comparative delight that a FOR-A timebase corrector was able to lock to the very unstable playback more reliably than the previous model.
Two transfers were made of each tape onto DVCam cassette via the TBC. We had found that the recorder would lock up in different places on different playbacks, so with luck we would be able to knit together a relatively stable version in the editing suite. Two of the tapes had fairly clean but very unstable replays, and the third had a more stable replay but was plagued by three noise bars over the picture which could not be removed by adjusting the tracking.
Once the DVCAM tapes were imported to computers we were able to analyse the problems we faced. Ideally, one field would have been stable and removing the bad field would have resulted in a good, filmised picture. Unfortunately the phase of the relatively stable field would change every few frames. Sometimes loss of signal would lead to up to ten frames being missing on one of the fields and the surviving field would not always be clean. There was no way that all four hours of material could be restored, so approximately thirty minutes was identified for further work. These sections were de-interlaced into versions made up of each field, from both tape passes (so in effect we had four versions of each section) and the best frame identified. During the combination process, disrupted fields were aligned roughly with the help of the burnt-in timecode on the tapes, and all but the worst "picture rolls" were fixed.
We now had versions of each section using the best existing fields, which were still very unstable with vertical jitter of +/- 40 pixels and occasional horizontal jitter. The Foundry's Furnace plug-in was used to stabilise the shots further. The next stage was to clean up the stabilised sections, for which Commotion was used to lessen the considerable dropout and rotoscope-in missing picture information. Badly damaged sections were replaced using Twixtor and Furnace where possible. Heavy adaptive noise reduction was then applied to remove the noise present on the recordings and, finally, the sections were VidFIRE processed to return the original video look that was lost when we had to lose the unstable fields.
All told, the restoration took around 140 man hours and is not something we would hope to have to repeat - very definitely a labour of love! However, in the end we successfully retrieved around twenty minutes of previously unseen behind-the-scenes material for the DVD from tapes which, initially, appeared to be unplayable and unusable.
After hearing that we had been allowed a second disc, we were able to relax somewhat and increase the size of the extras package that we had planned.
First up is a commentary, featuring actors Julian Glover and Tom Chadbon, along with the story's director, Michael Hayes.
Paris in the Springtime is a 45-minute featurette, written by Jonathan Morris and produced by Ed Stradling. It goes behind the scenes on the production of the story, with a particular emphasis on the contribution of Douglas Adams. Appearing in this informative and amusing featurette are Douglas Adams himself (courtesy of two separate archive interviews shot by Kevin Davies), Julian Glover, Catherine Schell, Tom Chadbon, Michael Hayes, Anthony Read (former script editor), David Fisher (writer of 'A Gamble with Time', the original story which was reworked into 'City of Death'), Pennant Roberts (director of Adam's earlier Doctor Who story, 'The Pirate Planet'), and new series writers Steven Moffat and Rob Shearman, all linked together by narrator Toby Longworth. A particular highlight is the retelling of 'A Gamble with Time', using a beautiful set of illustrations produced for this featurette by artist Jason Lythgoe-Hay.
Paris, W12 allows the viewers a rare 20-minute look inside the studio during the recording of the story, courtesy of extracts from the three Shibaden studio tapes mentioned above.
Prehistoric Landscapes is a short montage of the model landscapes and spaceship effects used in the story, most of which never saw the light of day. Similarly, Chicken Wrangler comically reveals the pitfalls of trying to film live chickens for one of the effects sequences, along with a rather too animated model chicken! Most of the material used in these two featurettes came from private collections and an astonishing amount of it was the original 35mm camera negatives!
Eye on... Blatchford is one of those little regional-interest programmes that the BBC does so well. The programme follows a day in the life of Sardoth, the Second-to-Last of the Jagaroth, as he tries to balance the responsibility of saving his race with his own attempts to fit into life in the rural village of Blatchford. It's not easy when you're green. Oh Mummy, they're at it again...
PC and Mac users will be able to enjoy the entire 1980 Doctor Who Annual in PDF format, courtesy of Steven Bagley.
An 8-minute Photo Gallery highlights cast and design photographs from the story, as well as studio floorplans and set models. Dr Martin Wiggins has supplied the production notes and there's no shortage of interesting Easter Eggs to hunt out and enjoy.
Copyright Steve Roberts, Jonathan Wood, Mark Ayres 8 August 2005