Why document type matters in language translations

Many of our larger clients use us for a variety of reasons.  The Legal department needs help with contracts, patents and investigations.  The HR department needs help with eLearning and immigration documents and Marketing/Communications would like assistance with brochures and their website. Note, “document” will be defined as what is inside a file.  For example, a contract is a document, that may be a DOC or PDF.  This therefore also includes videos and websites.  Different kinds of documents require different skills in translation, formatting and software.  Furthermore, different documents have different risks related to pricing and quality.  In this article, we break documents down into basic elements from a translation point of view:

• Text
• Images
• Layout
• Audio
• Controls

Each element brings its own challenges to the translation project and each document comes with its own combination of elements.

Text

When most people think of translations, they rightly think of text. Text is the most common element in all of translations.  This text can have very technical, plain or stylistic language.  In fact, at BURG, all text sits on a denotation – connotation continuum:

The types of documents listed for each segment of the continuum are only examples and are not exclusive to that particular segment in the continuum.

Translators with different kinds of training and styles are best suited for different kinds of text.  For example, a translator that is knowledgeable and proficient in translating patents using the right patent office rules may not find translating marketing material or poetry very easy or vica versa. In both cases, the number of translators with the skills to translate these kinds of documents might be much fewer than those capable of translating plane language text found on other documents and websites.

Beyond technical terms, there are some words that cannot or should not be translated.  The common examples are numbers and proper nouns.  When working on financial documents, such as a general ledger, numbers can make up a significant proportion of text that should not be factored into the cost of a translation.  Another example is a business card.  Almost every segment of text on a business card carries significant risk of translation error.  How do you translate “Bob” into Chinese? Which job title is the most relatable in the target language? Should we keep the local address or use the branch office? What about the name of the company and its slogan, does that get translated?

Internal vs external communications is another dichotomy that significantly affects translations.  Typically internal communications carry less business risk than external communications, meaning quality can vary without impacting business.  Some clients take advantage by adjust the translation process accordingly to reduce costs and accelerate turnaround time.

Images

The second most common type of elements are images.  We will include videos (but not audio) and animations in this category as the same kinds of factors need to be addressed. While most videos include audio, audio is deliberately kept separate because of the unique challenges it has.

Images can contain text, but they can’t be translated directly.  The text must typically be rekeyed as proper text in order to be translated efficiently along with the rest of the text in the file. For example, it’s common for charts to have labelled axes and a legend or for marketing material to have stylized text.  Any extra work that it takes to find, prepare and present translated images adds to the cost of the project.  Moreover, an inexperienced team may not bother to check for such things and will go untranslated in the final deliverable, which may cause delays and unexpected additional charges.

Layout

Layout refers to the formation of elements on a page (or slide, etc.).  In a contract, the layout is typically one large text block per page.  In marketing material, it can be a combination of various text boxes and images.  On a website, layout can include videos, buttons and more.  For the purpose of this article, tables will be included in this section.   Tables (not a spreadsheet) have a tabular layout of text and sometimes numbers.  Spreadsheets are typically used by the client when layout is not important.

When translating content, the amount of physical space that the translated text uses up on a page is typically not the same as the source text.  In fact, in the vast majority of cases, translation from English into a foreign language expands.  As a result, the size of cells in a table, text boxes and the arrangement of graphics need to be addressed.  In bidirectional languages (Arabic, Farsi, Hebrew, etc.) the layout of the entire page needs to be flipped.  Scientific journal articles have a particularly complex layout including an entire section (the references) that are not typically translated.  The more work the LSP has to do to address layout, the more costly and time consuming the project will be.

It is also possible for a source document to contain no layout.  For example, clients may submit text strings and bilingual tables for translation.  These documents are temporarily created to contain text extracts from another file for the purpose of translation.  The challenge with these files is that the lack of layout leads to a lack of context.  For example, text meant for buttons on an app can be wrongly translated.  When requesting translations of software, seek advice from an LSP on how exactly to send the text over to avoid problems.

Audio

Audio can refer to the audio portion of a video, a podcast or a simple audio recording.  Alone, they cannot be translated.  A transcription of the audio is required first, which may involve an additional specialist – leading to more time and money.  Furthermore, audio have a special linguistic feature about it that makes it unique to all other elements.  Since the pronunciation of spoken words is related to the pronunciation of the rest of the words in the phrase via intonation, one change in any word at any moment in the recording typically requires a re-recording of the entire audio file (We advise starting with a synthetic voice until you’re happy with the non-pronunciation portion of the audio).  This makes changes to audio files significantly more expensive than changes to any other kind of element.  In contrast, a term in a sentence can be changed in an instant.  Moreover, how words are written and spoken are not always the same.  For example, for an abbreviation, should it be spoken as it is spelled or as a word? Should the native source language proper nouns be pronounced natively or in the foreign language way? Knowing the answers to these questions can save a lot of money on a project with audio.

eLearning courses are particular in that they tend to have a very large amount of audio files that accompany visual aids such as slides.  As a result, a significant cost of translating eLearning files is in the audio.  It is strongly advised to review the interim outputs from the LSP prior to recording to ensure all preferred terms are set and that key pronunciation ambiguities have been addressed.

Controls

Controls are items that the user interacts with, like buttons, checkboxes, pulldown menus, etc. Controls are most commonly found in eLearning courses, software and websites and can act as containers for text.  This means that after translations, they may need to be adjusted to make the target language fit properly.  In fact, due to the size of controls, it’s typical that they need to be adjusted significantly, or the translations need to be rewritten to be shorter. The reason they are rewritten, rather than being shorter in the first place is that often times the translators do not have enough information in the first place in order to translate the controls properly – the client does not always know the character limits of their controls.  Also, the text typically arrives in context-free strings or tables without any context or notes to help explain to the translator what the controls the words are on are meant to do.  For example, a good practice is to supply LSPs with a table in which one of the columns contains notes like “to submit the form to the company” to describe a text string that is on a button that submits a form to the company.  It is also helpful to provide an emulator or screenshots of the entire screen to provide context.

Since translation files are usually copies of the source files, all the functionality of the translation files need to be tested to make sure the text appears properly and that they work in the same way as the source files worked.

If you’d like to learn more about how BURG Translations helps you ensure high-quality translations, contact us today.