How to Economize When Localizing Graphics
by Maxwell Hoffmann

Technical documentation, marketing materials, websites and software generally have graphical assets that need to be localized. As important as they are to the meaning of these communication portals, graphics often create additional challenges to the localization process, generating additional costs and further increasing time-to-market. Also, any problems with graphics in your source language files will be multiplied by the number of languages you have your project localized in. The obvious solution is to optimize graphics before localization is initiated.

To prepare your graphics for successful localization, separate graphic localization challenges into two categories: technical and cultural. Each category requires separate expert resources. A basic understanding of graphic preparation can help you reduce your localization costs substantially.

To save costs, try to exclude text in the graphics; whenever possible create the text as a separate component in the page-layout application used to create the document.   Example 1: Example of graphic with text embedded in the graphic source file. Text expansion after translation can cause text to straddle over other graphic elements, (like an arrow) and require manual repositioning of text lines, a costly process. Example 2: Example of same graphic with numbered callouts or captions which correspond to text in an adjacent table. Note that the table has sufficient horizontal room for text expansion. In extreme cases, the translated text can now automatically “wrap” to a second line in the table cell, eliminating manual repositioning of individual text lines.

Technical Challenges and Solutions
Technical challenges of localizing graphics typically involve replacing the source text with the translated target text. This is often done manually but is ideally accomplished with the use of computer assisted translation tools (CAT). The task becomes more expensive when the client only provides the localization vendor with rasterized/ flattened graphic files. Such files require substantial content recreation and publishing as the text cannot be manipulated in a bitmapped form.

To reduce costs, try to exclude text from the graphics; whenever possible create the text as a separate component in the page-layout application used to create the document. This requires less work to localize, as the graphic text is part of the main document text "flow" and not a layer inside the graphic file. This ensures that the text will appear "in sequence" to the linguist, and also allows for this text to be incorporated more easily into translation memory. Note: if the text must be adjacent to graphic elements, try to position it in such a fashion that there is some horizontal "halo" of space for text expansion. In such an instance, ensure that source art files are available to the linguist for text replacement and manipulation.

Substantial cost reduction is achieved when the localization team has access to text layers in the original graphic file format. It is critical to save the original source files and provide them to your localization vendor, in addition to the final output files. For example, to localize a GIF or JPEG file, you should provide the original Photoshop (.psd) source files, or any other layer-based original format used to generate the final graphics, color palette information, preferred fonts, export or save settings and any design specifications that were used to create the original graphics.

Externally referenced graphics in documentation
Another common practice that increases project budget and delivery time is the use of "embedded" graphics. Some users of Microsoft Word and page layout programs like PageMaker are in the habit of “drag n’ drop” graphics moved from a folder into the document window. Although this seems like a quick way to get a graphic on the page, this practice leads to "embedded" graphics which (a) tremendously increase file size (b) increase the number of steps in extracting and reinserting translated graphics, (c) make the accuracy of graphics more difficult to track and update and (d) add unwanted, billable project time for lengthy uploading/downloading of large files during the review process.

Try This When Delivering Graphic Source Files When delivering source files to a localization provider, provide a document that lists all of the graphics, along with their respective formats, and information such as the graphics that do not have translatable text, graphics that include text, and the respective pages and files where they can be located. Also include any special dpi (dots per inch) values and whether graphics have been resized.

Import graphics "from file" into your original source documents. This creates a "referenced" graphic, wherein your document contains a link, not the entire graphic file. Use a "relative" pathname to the source graphic. In other words, have a logically placed subdirectory where all document graphics are present. Your localization vendor should "mirror" your directory structure to ensure that documents open properly with imported graphics at the vendor or customer work sites.

Cultural Challenges with Icons
Successful localization of graphics also involves cultural research as some images, such as icons, are interpreted differently around the world. Some graphics may be offensive, others sacred or simply inappropriate for the context, while others images may not mean what you think they mean in a target country. For instance, a “pointing finger” (frequently used in cursors for hyperlinks, etc.) can be considered a rude or insulting gesture in some cultures.

Text Expansion and Screen Captures
Graphics designed without considering text expansion are difficult to localize. Text expansion occurs during translation for a variety of reasons: (a) many languages have equivalent words with far more characters than English (b) some languages tend to avoid the use of abbreviations (German, for instance) and (c) hyphenation and automatic line break issues may be quite different from English. For example, when English text is translated into Italian or German, the resulting text often requires approximately 25 - 35% more physical space on a page than the English text.

Screen captures are graphics containing user interface components of an application, such as a dialog box or menu. Successful translation of screen capture text is accomplished through the localization of the software that was used to generate the source language screen captures, followed by recreation of the screen captures using the localized software. Text expansion can be particularly crucial to software strings for text labels that occur over buttons and fields in a software user interface. Early consultation from your localization vendor can avoid cost over runs and “headaches” later in the project cycle.

Conclusion
Although some of the tips presented here may seem obvious, you will likely still find examples of unplanned graphic/text combinations in older legacy documents. You may also find undesirable placement of text strings within illustrations created by contractors. Work closely with your localization partner to establish logical guidelines, even a graphics “style guide” for use by your entire team . This will help ensure consistent content creation and optimized (and economical) localization project delivery schedules and budgets.

ENLASO Localization Tools Freeware Extracts Text From Graphics...
ENLASO's Localization Tools Freeware designed by Yves Savourel has some wonderful applications that allow the user to directly extract text requiring translation from Adobe® PhotoShop® (PSD) and generic binary files and then merge the translated text back into the original files. Download the ENLASO Localization Tools Freeware