In translations, everything would be so simple if it were just continuous text to be translated. But in our modern media-centric world, the types of information blend together more and more, so there is now text in illustrations, screenshots with text, videos with accompanying speech, and many other combinations.
With these combined media sources, it is often the case that the source information text isn't directly accessible for translation memory systems (i.e. CAT tools), which means that the semi-automated, highly-efficient processes of modern translation don't work here.
We’ve found solutions to meet these challenges.
Text Extraction
Often it is possible to read texts from graphics or video software via scripting and thus feed them into the standard translation process. After translation, the translated texts can then be read-in and, from there, they replace the source texts.
Such scripts can either be integrated into the software interface or kicked-off in a batch process. It should be noted, however, that some post-processing of the media may become necessary, due to incompatible run lengths.
The effort for the development of such scripts is especially worthwhile if the same tools with embedded texts are used multiple times within the information process.
XML Exchange Formats
Some media (e.g. graphics, 3D models) can be exported into XML exchange formats. In addition to the actual image information, the XML also contains the texts in readable form. By using appropriate filters, such formats can also be translated with translation memory-based systems (i.e. CAT tools).
Preparation and Post-Processing of Media
If certain tasks can only be solved once, a technical solution is often not worthwhile. Sometimes, due to the quality of the data, it's simply not possible to extract text from media. In these cases our Media Designers are brought in.
They extract the text from the media and then manually reinsert them after the translation process has been completed. They also take care of the appropriate formatting of the documents in the target language, taking into account the typographical peculiarities of the respective languages.