How To Keep Your Audio and Video in Sync, In Any Language

Every year hundreds if not thousands of new terms enter the English
vocabulary. It’s such a practical language for communicating complex
ideas, in fact, that English has come to form the basis of a universal
verbal shorthand. But for those of us who create foreign language
versions of media that originates in English, the language adds another
layer of complexity to a typical production. If your media is headed
for international distribution, you not only have to know graphics,
video and audio, you have to know your lingo and how it affects
everything else.
Although it’s common while speaking another language to borrow terms
from English-particularly for high-tech modern inventions-and sprinkle
them throughout a conversation, it’s inappropriate to use foreign
vocabulary in technical presenta-tions. Even the Vatican has begun to
create words in Latin to replace modern technical words in English. If
your video will be translated into multiple languages, you need to
remember that word order, as well as the time it takes to say
something, will be very different for each version. In order to avoid
havoc during post, there are a number of things to keep in mind
throughout production, and especially when preparing your original
script.
So how do you produce a video that is friendly to foreign language
translation? When writing your script, always consider your
international audience first. If the verb or action described were to
appear at the end of the sentence or segment, would it confuse the
viewer? Most foreign languages, both spoken and written, run longer
than English does. Subject/predicate orders vary. If you are working
from a transcript, it pays to take the time to remove redundancies of
speech before going into translation. (If you do this kind of work
regularly, it also pays to have a competent translations department.)
But it’s also smart to think about your graphics at this early stage
and plan for extra time to clone graphics and still scenes that need to
be relayered with translated text. For example, when you add subtitles
on the screen, do you need to alter the way text, graphics and video
appear in that segment?
Learning the Lingo
We create multilingual training videos and DVDs for multinational
corporations and regularly work with these languages: Germanic (German,
Dutch, English and some Scandinavian languages), Romance (Italian,
Spanish, Portuguese, French) and Asian (Japanese and Chinese). Chinese
represents several languages itself (Cantonese, Mandarin and others)
and is unrelated to Japanese, though one of Japanese’s three writing
systems appears to have evolved from Chinese characters. Cantonese is
spoken in Hong Kong and uses traditional script; Mandarin is spoken in
mainland China and uses simplified script. Romance languages all
descended from Latin with varying degrees of influence from Celts,
Germans, Arabs and others. Germanic languages descended from a common
Germanic tongue with varying degrees of influence from Celts, speakers
of the Romance languages, Arabs and others. At Fudev International, we
currently have 50 independent translators in Spanish alone who are
contracted according to area of specialty.
Our clients typically provide us with video that has been previously
produced. Our job is to take that material and create a training DVD
with upwards of 12 different language tracks. Apart from navigating the
different languages, our biggest challenge is often fitting all that
extra audio, text and graphics material that results from the
translation process onto a standard DVD. We recently completed two
fairly large projects-one for JLG Industries (see project shot, below
left), makers of aerial work platforms, and one for DuPont-that
illustrate those challenges in detail.
JLG asked us to produce a series of multi-language DVDs in both PAL and
NTSC. The massive amount of technical and legal vocabulary was a major
issue from the start. Had we waited until we booked talent to verify
the legalese and technical jargon, our sessions would have been
disastrous-not to mention too long and over budget. Our translation
department verified and re-verified everything. This gave us great
confidence going into production.
Although the scripts were condensed as much as possible, as we tracked
through the piece, we soon realized we were going to have trouble with
sentence length and needed to find ways to stretch the video to match
our translated audio tracks. We don’t time-compress spoken audio more
than 2 percent-this was way beyond that. So even while we were still
tracking audio, we decided to let video post begin rendering an
internationalized version of the video.
Our video post studio immediately selected scenes we could edit. We
identified maybe twenty locations over the length of the 40-minute
video where we could grab a still and extend it a few frames. It was
all seamless and imperceptible. Using the newly edited video to help us
edit audio, we had no more serious timing issues during production.
The length of the video and number of languages-eight in all-proved to
be an even bigger challenge. It wasn’t immediately clear how we were
going to fit all the different versions onto one disc without
sacrificing the quality of the video or the full-motion menus, which we
knew we didn’t want to do. On top of that, we also wanted to keep all
the sub-menus, as well as the on-screen text, consistent with each
foreign language. Our client wanted to used DVD-5 for duplication so
DVD-9 was not an option. Our solution was simple: We recreated the
still scenes that featured localized graphics and text. Our versions of
the stills had no text. Once we reached the DVD authoring stage, we
recreated the text in subtitle format and matched it to what was on
screen. Then, in the menu set-up options, we tied the subtitle
selection to audio selection and disabled the user control of the
subtitle feature. This way, on-screen text stays visible at all times
and appears in the same language as the selected audio. This let us use
a single video stream featuring localized on-screen text for eight
languages and eight user-selectable audio tracks.
Dealing with Lip Sync and SD Quality
The DuPont project consisted of producing multilingual videos in PAL.
The biggest challenge for us was tracking for lip sync using scripts we
had no control over (they were supplied to us already translated).
Luckily, the language wasn’t technical, but it continued non-stop over
six videos, which were each 12 minutes long. The recording sessions
were quite long, and our talent certainly rose to the occasion.
We also had to recreate all of our graphics in the foreign languages
without having source video. This was a real challenge for segments
with a lot of motion graphics that included picture-in-picture video
and overlays. First, we identified those scenes that were reused in
picture-in-picture effects. We copied those files to a new archive in
the source media directory and as in the JLG project, we recreated all
scenes where graphics and text were to appear, and left them blank.
Finally, we archived the naked versions into a general "new graphics"
folder within our source media directory. Because we were working in SD
and had to conserve on space, we created just one project file in Adobe
Premiere Pro. This project file featured the original video track plus
four additional video tracks for the new graphics files and effects. We
also created eight audio tracks, one per language. We needed only a few
video/graphics tracks because we knew Premiere would always pull from
the folder called "new graphics" in the source media directory. We had
a separate directory for each language. Whenever we wanted to switch
from German to Spanish, for example, we would rename a directory to
"new graphics" and Premiere would automatically update itself to the
new language. This made the project much more manageable. In the end,
you couldn’t tell that the original video had been adapted.
Having in-house translators has been indispensable for us and often
heads off more serious problems during production or post. While
creating graphics for the DuPont project, we noticed that the German
text to appear on screen was different from the German audio. The
original English text was written, "Stop for Supervision." The
on-screen text was translated as Stehen fà¼r Vorgesetzte. The script for the audio, however, had been translated as Stop fà¼r Vorgesetzte. There’s a difference. But because we caught it in time, we were able to make the correction before going to mastering.
We never did run out of space for either of these two large projects,
which was no small miracle. In fact, both projects were produced
simultaneously alongside other SD with only 1.2 TB of total storage
space. The JLG project was much more fluid because we were involved
from the beginning and had an intimate knowledge of the project almost
from the ground up. The DuPont project, which we joined later in the
production process, was more challenging. But with some dedicated
producers, translators, great gear and plenty of advance planning, we
avoided a lot of lingual-and production-headaches.