JAN
IVARSSON:
A
SHORT TECHNICAL HISTORY OF SUBTITLES IN EUROPE
Mechanical and thermal processes
Dedicated subtitling equipment
It
was not long after the invention of film that efforts were first made to convey the dialogue of
the actors to the audience. They started with what we now call intertitles: texts, drawn or printed on
paper, filmed and placed between sequences of the film. They were first seen in 1903 as epic, descriptive
titles in Edwin S. Porter's Uncle Tom's Cabin. (The technique may have
been invented by cartoonist and filmmaker J. Stuart Blackton.) The titles were
from 1909 on called sub-titles, as they were
used in the same way as subtitles in for instance a newspaper. Early, but
rarely, the subtitles were placed in
the
moving image, for instance as in
Porter's College Chums (1907) or the French films Judex
(1916) or Mireille (1922).
In
the era of intertitles, it was easy to solve the translation problem. The
original titles were removed, translated, filmed and re-inserted. Or a
speaker was used to give a simultaneous interpretation of the intertitles, the
French bonimenteur or the Japanese benshi.
In
fact, the very first “subtitles” in the modern sense saw the light of day already during the
silent film era. In 1909 M. N. Topp registered a patent for a “device
for the rapid showing of titles for moving pictures other than those on the
film strip”. With this method the projectionist, using a sciopticon (a
kind of slide projector), showed the subtitles on the screen below the
intertitles. However, this was never much more than a curiosity, although
similar techniques, with the titles on a film strip instead of slides, have
been used from time to time up to the present day (Brant, p. 30).
From the year 1927 on, with the invention of sound film, the audience could
hear the actors, so the titles inserted between scenes disappeared and the
problem assumed new dimensions. Of course, one could make several language
versions, or have the film post-synchronized (dubbed) in another language.
However, some film producers and distributors found this technique complex
and expensive.
Why
not use titles as before, inserting them in the picture? They thus became what
we now call subtitles, and since this technique is comparatively cheap (subtitling
only costs between a tenth and a twentieth of a dubbing), it became the
preferred method in the smaller language areas, such as the Netherlands
and the Scandinavian countries.
In
the early days of film subtitling the main problem was to place the subtitles
on the distribution copies, as the negative was usually in safe keeping in the
country of origin. Norway,
Sweden, Hungary and France quickly took the lead in developing techniques
for subtitling films. However,
"the
first attested showing of a sound film with subtitles was when The
Jazz Singer (originally
released in the US in October 1927) opened in Paris, on January 26, 1929, with
subtitles in French. Later that year, Italy followed suit, and on August 17,
1929, another Al Jolson film, The
Singing Fool, opened in Copenhagen, fitted with Danish subtitles." (Gottlieb, p. 216)
To
start with, attempts were made to breathe new life into the technique invented
in 1909, i.e. manual projection of slides with printed texts directly onto the
screen, but very soon, methods of copying photographed titles on to the film
copy itself came into use. A frame containing the title was kept in position
while the film negative and the positive print strip were fed forward and
exposed.
Later
on this process was made automatic. Exposed “blank” frames were inserted
between the title frames and the titles were fed forward by means of a counter
to ensure that the subtitles were the right length and came in the right place.
One
problem with the method was that, since the original film negative was usually not
available, it was necessary to re-copy the whole film to obtain a new negative, with a consequent loss of focus and substantial increase in the
noise level—a serious drawback in the early days of sound films.
But
sometimes the film negative could be obtained, and it was soon realized that,
if a large number of copies were required, the most efficient method was to
photograph the titles onto a separate film of the same length as the original, with the in and out cue frames synchronized with the sound.
The
film negative and the roll with the titles were then copied simultaneously,
an operation which took much less time than repeating the slow exposure
procedure frame by frame.
In
1930 a Norwegian inventor, Leif Eriksen, took out a patent for a method of
stamping titles directly on to the images on the film strip, first moistening
the emulsion layer to soften it. The titles were typeset, printed on paper and
photographed to produce very small letterpress type plates for each subtitle
(the height of each letter being only about 0.8 mm). Later, in 1935, a
Hungarian inventor, O. Turchányi, registered a patent for a method whereby
the plates were heated to a sufficiently high temperature to melt away the
emulsion on the film without the need for a softening bath. However, both
these processes were difficult to control and results often erratic, with
poorly defined letters. Despite the drawbacks, this technique has been used by
some film laboratories in eastern Europe, Asia and South America up to the
present day.
In
1932 R. Hruska, an inventor in Budapest, and Oscar I. Ertnæs in Oslo (later
in Stockholm) simultaneously took out patents on an improved technique for
impressing the titles directly on the film copies.
A
very thin coating of wax or paraffin was applied to the emulsion side of the
finished film copy. The printing plates were placed in a kind of printing
press, into which each plate was fed and heated to a temperature of nearly a
hundred degrees and one by one pressed against the paraffin coating at the
bottom of the frame which corresponded to the beginning of the dialogue line.
The paraffin under the letters melted and was displaced, exposing the
emulsion. This process was repeated with all the frames on which this subtitle
was to appear, corresponding to the duration of the speech. The same
procedure was carried out with the next subtitle and so on throughout the
film.
After
the printing process the film was put through a bleach bath, which dissolved
the exposed emulsion, leaving only the transparent nitrate or acetate film.
The etching fluid and the paraffin were then washed away. This process
produced clearly legible white letters on the screen, although the edges of
the letters were slightly ragged due to the variable consistency of the
paraffin and variations in the penetration of the etching fluid.
Later
on, this process too was automated by means of a counter, which fed the plates
forward, counted the frames on the roll and ensured that the subtitles came
in the right place and were of the right length.
This
was the cheapest process when less than ten copies of a film were to be
subtitled.
The
chemical and optical processes described above are still used in many
countries, more or less as before, except that the plate making process has
been modernized. In the early days the titles were typeset (usually with a
Linotype machine), printed on paper, photographed and then plates were made
for each set. Later, with the adoption of new techniques in the printing
industry, came phototype setting (e.g. Cinétype) and still later computerized typesetting. Nowadays computers are used for the
production
of the titles themselves, and they can be time coded and “simulated” on
a videocassette for proofreading purposes.
The
Norwegian-Swedish film laboratories Filmtekst in Oslo, Ideal Film in Stockholm
and the Kagansky brothers’ Titra-Film in Paris held the most important
patents, as a result of which they dominated the European subtitling market
from 1933 right up to the mid-50s and were very important also on other
continents. (According to interviews given by O. Ertnaes’s daughter and Mme
Nina Kagansky, Paris. See also Brant, pp. 53-63. Her thesis contains a
detailed description of all the stages in both the optical and chemical
processes.)
The
latest development in this field is the use of lasers to burn away or vaporize
the emulsion. This makes both typesetting and plates unnecessary. The
technique has been developed by Denis Auboyer in Paris and by Titra-Film in
Paris and Brussels and has, with great success, been in commercial use since
1988.
In
this process a computer controls a very narrow laser beam, in the same way as
in a modern typesetting machine, i.e. the beam virtually writes the text in
such a way as to result in vaporization of the emulsion without damage to the
acetate film underneath. It takes the beam less than a second to write a
subtitle consisting of two lines, after which the next frame is fed forward.
Where no subtitles are to appear the film is fast-wound to the next operative frame. The sharpness of the letters is excellent, the contours
being enhanced by a slight shading caused by the darkening of the edges due to
the heat.
The
titles themselves are computer typeset and can be cued on the video display by
means of time coding or frame counting.
Laser
subtitling is cheaper than the chemical process, but requires costly
investment in equipment. However, the method is highly automated and needs
very little personnel.
Films
for the cinema were soon shown on television. On August 14, 1938, the BBC
broadcast Arthur Robison’s Der
Student von Prag in a subtitled version. (This was probably also the
first scheduled showing of a film in the history of television.)
But
it was soon discovered that the prints with subtitles intended for the
cinema caused a number of problems. The titles, legible enough in the cinema,
were very difficult to read on the television screen. One reason for this is
the difference in the speed at which the audience can read subtitles on
television as compared with the cinema, but the main reason is that the
picture on a TV set has a narrower contrast range than that on a cinema screen.
What was needed, therefore, was a method for incorporating subtitles
produced for television into untitled film copies or video tapes.
In
countries where the optical process was used for subtitling films, attempts
were made to use the existing subtitle film strip and run it in parallel with the
original untitled film in a second film scanner. The title images were mixed
electronically into the film images so that it looked to the viewers as if
the titles were on the film, except that it was now possible to control the
whiteness of the letters. If a roll with subtitles was not available, one
could be ordered from a company that made subtitles for films. This method is
still used occasionally today.
At
about the same time work started on the development of a rather crude, but
cheap and reliable, optical subtitling process for television: The
titles were typed on paper and then one-frame stills of each title were made with a film
camera. The resulting film negative was put in a scanner and then either the translator
fed in the titles manually, one at a time, synchronizing them with the
programme, or an automatic system was used to feed in the titles, more or (usually)
less reliably, with the help of punched-out marks on the edge of the film.
The
title images (usually with white letters against a black background, a “letter
box”, the whiteness and blackness being controlled to ensure optimum
readability) were mixed into the programme images and transmitted or taped.
Where no subtitles were to appear, exposed frames—blank frames—were placed
between the subtitle frames.
Quite
soon, some improvements were made on this method. For example, the titles were
printed with more attractive proportional typefaces on offset composers,
i.e. simple typographical setting machines which also allowed the use of
italics and kerning, squeezing the letters together. The titles could be
written on punch cards, inserted in a feed mechanism and either photographed
onto a roll of film or displayed live using a TV camera with image
inversion (black shown as white and vice versa). This “rapid subtitling”
method was used mainly for news items. Thus, photographing the subtitles and
developing the film were no longer necessary, but the feeding system was unreliable: sometimes the machine supplied several cards at a time or none at
all.
Both
these techniques allowed manual feeding of the subtitles during recording or
transmission or, as with film subtitling, automatic feeding by means of a
frame counter.
When
caption generators of various types (such as Aston, Capgen, Logica, Vidifont)
started to be used to insert captions in the television image by electronic
means, this made it possible to generate subtitles directly in the
transmitted picture itself.
However,
caption generators, which were intended for various kinds of captions or
titles and offered a wide range of typographical variation, proved
impractical for subtitling in large quantities. They were difficult to operate,
their word processing functions were very rudimentary, and above all they were
extremely expensive.
It
was therefore only natural that efforts should be made to produce dedicated
subtitling equipment, and this was achieved in the 1970s, more or less
simultaneously in several places. Two main systems were developed, both
based on the use of a word processor with a special subtitling program which
made it possible to write the subtitles in a form identical to that shown on
the television screen.
The
first system is based on the teletext principle (Oracle, from the U.K., and
Antiope, from France, are two examples). A computer generates concealed
signals in the image data, in response to which a simple character generator
in the receiver creates the characters and mixes them into the television
picture when a specified teletext page is selected.
The
second system uses a computer-controlled character generator in the
transmitter—much less sophisticated and much cheaper than a caption
character generator—and when the subtitler cues in a new subtitle, the
characters are generated by electronic means and mixed into the transmitted
image. Such systems are e.g. the BBC TV’s Television Electronic Characters (TEC) system from 1976, the SVT–TeleEkonomi’s system,
which came into operation in 1981, and the Screen Electronic system, from
about the same time.
But why feed the titles manually when time codes, which have so many other functions in television, could do the job?
When
a video tape is time coded, a “clock” is recorded on the tape which tells
you to the nearest 1/25 of a second when a particular frame will appear on the
screen. This time code can be read while the video tape is running and used to
start or stop some process, e.g. to show a selected subtitle, as desired.
Soon
further advances were made, which made it possible to install the complete
subtitling system on a personal computer, thus allowing the subtitler to
carry out the whole job, including the cueing of the subtitles in the right
place in the programme, in a continuous operation and in his own home or
office.
Brant,
Rosemary (1984), The History and
Practice of French Subtitling,
Paper.
Univ. of Texas, Austin (United
Microfilms International Dissertation Information Service, Ann Arbor 1989)
Gottlieb, Henrik (2002), "Titles on Subtitling 1929-1999", in Caimi, Annamaria (ed.) (2002), Cinema: Paradiso delle lingue. I sottotitoli nell'apprendimento linguistico, (Rassegna Italiana di Linguistica Applicata, Anno XXXIV, 1/2-2002), Bulzoni Editore, Roma, 436 p.
Ivarsson,
Jan. (1992). Subtitling for the Media.
A Handbook of an Art. Stockholm, TransEdit, 199 p.
Ivarsson,
Jan & Carroll, Mary. (1998).
Subtitling. Simrishamn, TransEdit
Kagansky,
Nina. (1995). TITRA FILM. Une
chronique cinématographique et familiale.
Paris.
Machado,
José. (1993). La traduction au cinéma
et le processus de sous-titrage de films. PhD thesis, Univ. de Paris III.
Testa, Bart (2002), "Screen Words: Early Film and Avant-Garde Film in the House of the Word", at Symposion Das frühe Kino und die Avantgarde, Vienna
Letters
patent at the Swedish Patent and Registration Office, Stockholm.