by Maxwell Steer © 1995
Dubbing is the art of the possible - an opportunity to fasten the wings on a production so that it soars into the air - or a last chance to wallpaper over the cracks of a rickety structure to stop it falling apart.
Either way, dubbing (or rerecording, as it's generally known in the US) is constrained by many technical and artistic factors, not least that it often takes place when all the money has been spent! As Graham V Hartstone, Head of Post-Production at Pinewood says "Dubbing always occurs at the tail end of the process. They'll spend as much time and money as they've got rewriting and reshooting, but when it comes to dubbing they expect the mix to happen right the first time."
To some extent dubbing is the Cinderella art of film-making, it's all but invisible to the general public yet the final responsibility for the sound of Prince Charming's approaching footstep rest firmly with Cinders rather than with the Ugly Sisters of picture or sound editing.
So what is a dub?
The term originated in the US and may have applied originally to 'doubling' or copying the Vitaphone sound discs that were the first successful attempts at synchronising sound with picture. Or it may have originated with 'doubling' or post-synchronising the actors' voices, something that was necessary very early on. The rush to convert to sound was so intense that even films already shooting were literally converted to sound overnight. One such was The Canary Murder Case where the star, Louise Brooks, refused to reshoot the silent scenes so that another actor was brought in to 'double' her voice and match it to the sound scenes.
Whatever it's origins, dubbing as much as any other aspect of production has always been governed by available technology, and therefore any consideration of the art of dubbing has to take account of the development of film.
Bob Allen of the Association of Motion Picture Sound has been researching early audio patents and has found that many current ideas go right back to the early years of the century - for instance there's a patent for a radio mic as early as 1917, and optical stereo was patented in the early 20s, six years before combined optical prints were first available. But as Bob Allen points out "most developments had to wait, first for transistors & then for digital, to make them possible."
The Sound of Silents
During the silent film era there were some directors, notably DW Griffith, who actively did not want sound. Griffith believed that wordless film was a truly universal dramatic medium capable of touching hearts the world over, and took immense care to send round complete scores to accompany his pictures. On the other hand, one of the strongest motivations for the development of sound film was that the top grossing stars of the first two decades of the century were vaudevillians and opera singers who would emote for the camera at the top of their (silent) voices accompanied by a pit band playing their tune!
Experiments with synchronising sound-on-film were first made by the inventor of film himself, Thomas Edison. In 1895 he demonstrated his Kinetophone, but available technology hadn't really licked amplification at this stage. He was followed by Gaumont's Chronophone and Nolan's Cameraphone. It wasn't until 1923 that a tenacious inventor, Lee De Forest, demonstrated Phonofilm, the first viable optical sound-on-film technology.
But unfortunately for him a rival system track was being developed by the cable /telephone giant Bell/AT&T. Their subsidiary Western Electric (later Westrex) had developed a 16 inch shellac disc revolving at 33.3 rpm which could record 9 minutes of sound, and became known as the Vitaphone system. The slower speed allowed a significantly better signal/noise ratio over the consumer standard 78 rpm. After a spectacular debut in 1920 when a recorded speech by President Harding was relayed to packed auditoria in New York & San Francisco the search was on for a viable technology to harness the invention to film. By 1926 the problems had been solved and Warner Brothers had acquired an exclusive licence for the Vitaphone sound-on-disc system.
1926 was quite a year. A few months before Warner's first 'sound' film, Don Juan, national radio broadcasting had started in the US, and Fox's Movietone news was to follow hard on its heels. Tho Don Juan featured only music and sound effects it electrified audiences that summer. However it was the legendary Jazz Singer that hit the jackpot and was still grossing $100,000 a week per theatre a whole year later, catapulting Warner's into the major leag with its stock rising in value 600% in two years.
In fact there were only four 'talking' segments in The Jazz Singer, each of one disc's length, which was conveniently just inside the duration of a reel of 35mm film. The projectionist had a cue frame on which to start the Vitaphone disc, and even if it all went astray such was the novelty that nobody cared.
Strictly speaking, to coin a phrase, the honour of the first talking picture should go to Fox. A month before The Jazz Singer they'd released the first Movietone newsreel using Photophone sound-on-film technique, a rip off of De Forest's system who, in time-honoured fashion, died in poverty. The sound was merely voice-over rather than sync sound (which was not really perfected for another 5 years) but the flexibility of this method came into its own the following year when Lindbergh's historic transatlantic flight was shown in cinemas nation-wide within 24 hours.
Sound caught on like wildfire. But at this stage mixing (optical) soundtracks remained technically impossible. Sound tracks consisted of only a single layer of sound at a time, be it dialog, songs with orchestra or effects, therefore cinema orchestras were still needed to provide incidental music. Dubbing was little more than joining together raw 35mm optical sound neg and covering the joins with 'blooping ink' to minimise interference to the optical track before the film labs created a combined optical print. The sound track was located between the picture image and the sprocket holes or perforations as it is today.
Interestingly, the Laurel & Hardy films of 1931/2 seem to have been the first to dub music under the voice track. Almost certainly this was done by optical superimposition with continuous, pre-set levels. So great were the problems of mixing that initially optical sound emulated visual procedures, where in order to make dissolves, the outgoing scene was faded, then film wound back in the camera and the incoming scene faded in as a double exposure.
At first standard movie film was used for sound but the physical characteristics of grain introduced so much noise that special emulsions were soon developed to shift the noise up the spectrum. These were further softened by lightly exposing or 'pre-flashing' them. Another improvement came early on from biased recording.
Sync sound was initially a very hit and miss affair, posing horrific problems for post-production. Recordists divided their least favourite actors into boomers and essers. The first mics were fundamentally the same carbon omni-directional transducers used in telephones. (In the 1928 film The Lights of New York a prop candlestick telephone on the desk was the mic recording the voices.) They got very noisy in wet weather, suffered terrible 'fall off' if the subject was more than two feet away and, in days of unscreened leads, needed to be as close as possible to the recorder. Think of the scene in Singing In The Rain where Debbie Reynolds has to overdub a performer whose pearl necklace comes across louder than her voice, and the business of having to sing into the bushes! No wonder the divine Louise Brooks couldn't be bothered with it.
The War of the Words
The main complication of the first few years of sound-on-film was a fiercely fought battle between two rival systems of optical encoding. While some major studios such as MGM followed Fox, by now 20thC Fox, in pioneering Western Electric's Variable Density system, where the sound was recorded as horizontal lines (resembling a barcode) closer for higher frequencies and varying in density according to volume RKO & Republic followed Warner Brothers in adopting RCA's Variable Area, which created an analog waveform of varying amplitude. Vitaphone withered as sound-on-film quickly took over. Owing to an old, cleverly worded patent-exchange agreement between RCA & Western's parent AT&T, RCA had an option to licence AT&T patents. When they saw the technical superiority of the Variable Area system RCA realised it had much to gain by persuading the majors to modify their equipment.
The art of dubbing was then, as now, to squeeze the maximum amount of audio content onto the film track so that optimum signal/noise ratios were maintained in the cinema which were often little more than 'corridors' with a loudspeaker at one end. The average bandwidth of early optical sound was 100Hz-4kHz(!) tho this expanded during the following decade to a potential 30Hz-10K, but the dynamic range still remained under 30dB. In practice dubbing mixers worked to a standard evolved by MGM's Head of Sound Douglas Shearer (brother of Norma) known as the 'Academy curve' or 'Academy roll off' with a floor of 125Hz, and a slope above 4kHz tailing to -15dB at 8kHz, with an effective ceiling of 9kHz. This remained a de facto mono standard until the Dolby A in 1970s
Dubbing as we know it is generally reckoned to have started around 1930. The film generally credited with pioneering sound mixing is Rouben Mamoulian's Applause. Mamoulian experimented with editing all the sound on two interlocked 35mm tracks which was really the beginning of standard film tracklaying/dubbing practice. Later, to achieve 'unreal' sounds in Dr Jekyll & Mr Hyde he experimented with a number of photographic techniques creating 'sound' directly onto the optical negative.
In the early '30s production sound mixing consoles had 4 channels each with an on/off switch and single rotary fader or potentiometer. A total of 8 controls! Dubbing consoles were often two production mixing desks placed side by side. Synchronisation procedures were at first limited to four tracks. Standard 'dubbing prep' was to lay up dialog on track one, leaving three tracks to be shared between Music and Sound Effects as appropriate. The staggering achievements of the Busby Berkeley and Astaire/Rogers films were all made under limitations unthinkable today. Fred Astaire in particular had to pre-record his tap steps exactly as he would afterwards dance them. (In England Sound Effects tracks were always known as 'FX' but are now increasingly called by the US term 'Foley' after the legendary footsteps expert Ed Foley of Universal.)
One of the first functions of dubbing mixers was to match the sound perspective to that of the shot. Tho initially little more could be done than varying levels, it was aided within a few years by a number of technical improvements including the introduction of dynamic mics. Who knows?- perhaps John Gilbert's career as a screen Romeo would not have come to such an abrupt end if EQ had been available to tweak down his 'white' voice?
Optical sound was capable of very high quality and the conductor Leopold Stokowski, a technical as well as musical innovator, who recorded the music for Fantasia, insisted on recording his Philadelphia Orchestra on 35mm even after 1/4" magnetic tape became available. By the mid '40s dubbing consoles of 8 channels were normal, and 10 or 12 'not uncommon'. Despite considerable research I haven't been able to establish when EQ was first introduced, but from the films of the period at least Hi-pass/Lo-pass filtering must have been a feature of the larger desks.
Until the German invention of magnetic tape became available after WWII and its flexibility gradually loosened the RCA /Westrex duopoly all dubbing consoles were built and owned by the patentees and licensed to producers who, even in 1930, had to pay a fee of $500 a reel for their use É negotiated down from $4000.
England's oldest specialist film sound facility is De Lane Lea, based in Soho. Still at the forefront of innovation, De Lane Lea is progressively converting to digital operation. Norman Brown, now their technical director, remembers working as an installer for British RCA in the '50s. "They had very crude bass and treble passive filters with low and mid frequency shelfs. All the R&D was done by the studios. The manufacturers wouldn't spend anything on development."
His colleag Hugh Strain, De Lane Lea's senior dubbing mixer, recalls starting in 1949 as a trainee in the sound department at MGM's Boreham Wood studio in where his first job was "using a fine brush or pen to remove dirt on optical prints in order to reduce noise. At that stage we were still recording music optically on 200mm Variable Density 'push pull' tracks which were reduced to 100mm in final version. The dubbing process [in Britain] involved converting the 35mm optical tracks into electrical waveforms, mixing it and then reregistering the waveform on 35mm optical negative.
"MGM had a 24 way desk at Boreham Wood with some primitive Dialog EQ which could be used to lift the voice. I think the thing that's changed since then was the staggering number of voice[-over] loops an actor could get thru in a day - 220 was not uncommon. The big Hollywood stars were completely used to it."
Taking Stock of Maggie
However by 1952 magnetic film was used for 75% of Hollywood recording, music and dubbing. Initially editors disliked mag tracks as they had become so skilled at 'reading' optical tracks that they complained having to use a 'squawk box' to review the audio slowed them down and continued to demand an optical 'work track'. In 1950 Paramount Head of Sound Loren Ryder had won a technical Oscar for converting the whole studio to sprocketed 17.5mm magnetic audio, the medium used for The 10 Commandments.
Hugh Strain recalls that the first feature film he dubbed using mag film was Moby Dick. "Prior to that you still had to dub complete reels in a single pass because, like optical stock, in the early days of mag tracks it was impossible to drop in or out. In the old days you didn't spend ages discussing things - you'd do five takes, maybe print them all, then leave it to the editor to decide."
Magnetic film also permitted multi-track recording. Initially three tracks was the favoured standard, but today 3, 4 or 6 tracks are options. An immediate economic advantage of a triple track mag master was that it was no longer necessary to go thru the whole dubbing process from scratch when making foreign language versions, since the Dialog, Music and Effects (Foley) were recorded as discrete tracks on the final mag master. Any subsequent replacement of dialog left the 'M&E' tracks untouched.
Magnetic recording very soon led another significant development: reversible drives. Immediately nick-named 'Rock & Roll' dubbing, this was far more quickly adopted in Europe than in the US where it became known as 'back up recording'. Rock & Roll was invented in about 1960 by M Ninnie, a former Westrex engineer, at Studio Boulogne in Paris. A redoubtable innovator, M Ninnie also perfected the electronic problems of 'dropping in', "but at this stage", observes High Strain, "once you'd dropped in you still had to continue the take until you got to an obvious 'out point' such as a gap in the dialog track or a door slam. Within a year most people had found a way round the problem tho 'outserting' wasn't available electronically until Sonifex produced their DDL recorders ten years later." Among M Ninnie's other inventions were a cueing device using an 18thC clock mechanism for post-synchronising dialog with balls descending a tube. (No doubt as a salutary reminder to directors who wished to keep them in the air!)
"Rock & Roll may speed up dubbing," Norman Brown adds caustically, "but it slows down the output. Now you can do drop-ins an awful lot of time is wasted by going back over things, second guessing them. The effect of advances in technology is that they allow directors to keep their options open right up to the final mix, and that means that instead committing to sensible decisions earlier they keep defering them, and in general that slows the dubbing process down because they're forced to make their minds up about all sorts of things they could have sorted out earlier on."
Like many other film dubbing studios De Lane Lea is progressively integrating 'digits with sprockets', an operation supervised digital expert Nick Church. Already they have a number of 16 track Audiofile mainframes whose M16 terminals can be remoted equally from Dubbing Prep rooms and Studios enabling an editor's pre-programmed mixes to be used 'as is' or remixed by the dubber in the context of additional material.
Nick is particularly enthusiastic about a new desktop alternative premixer, TimeLine's Studio Frame. A PC-based system it replicates all the functions of Audiofile but also prints frame-accurate dubbing charts in both horizontal and vertical formats as well as itemising effects as separate 'spotting sheets' - a vast saving of labour. His favoured medium for replacing sprocketed film is the Akai DR8 recorder reading/writing both DR8 Hard Discs and DD1500x magneto-optical (MO) drives. He explains "the DR8 is the one digital multi-track slave system which reads time-code in reverse, unlike the Tascam DA88 or ADAT. This means that the dubbing mixers can save time by adjusting their EQ as they roll back. We're installing five of them in our new Studio One, so that'll give us 40 tracks of DR8 in addition to the 72 we already have of 35mm, plus AudioFile or Studio Frame."
Continuation of article