Translation Memory Creation: How It Works

Sep 17, 2015

In our previous blogs, we emphasized the advantages of having well-maintained translation resources, which reduce your translation costs, mainly in connection with continuous translation undertakings. We have also touched on the process of "cleaning" and maintaining your translation memories, but what if you don't have any? If that's the case, you should definitely create one if you publish documents with similar content. Here we explain how to go about this step by step.

5 Steps to creating a Translation Memory

The process begins with two similar files – one with source text, the other with corresponding target text in a different language. To create a TM for you, we first need to pool as many of your original files with the translated equivalents. The file format doesn't matter too much. The documents can be in Word, Excel, FrameMaker, InDesign, Acrobat PDF, or whatever other format you may have.

Then, we apply these 5 steps in order to create your translation resources.

  1. SEGMENT EXTRACTION. We extract all text segments (basically sentences) from the source and target files to create a kind of bilingual database with original text and the corresponding translated text.
  2. SEGMENT ALIGNMENT. We then confirm that all segments are correctly aligned using our unique, in-house developed iSync solution. It pairs segments based on the placement markings and content. The process is highly automated and enables very fast processing, much faster than humans and with high precision.
  3. HUMAN EDITING. A human review of the result is necessary, though, in which native translators browse the paired segments to ensure they match.
  4. TM CREATION. At the end of the process we delete redundant segments that have no matches in neither source nor target, and we then export the bilingual text segments to the universal .tmx format or any other format you may require.
  5. TM QA AND EXPORT. The last step involves applying our QA tools on the final result to ensure segments are consistent, numbers, tags and symbols match and there is no text in the wrong language.

TMs belong to you

Once a translation memory has been created and delivered to you, it becomes your legal property, and you can use it in all your internal processes or whenever you outsource translation projects. This is important to remember, and you can always ask your existing or previous suppliers to deliver TMs they have created in projects ordered by to – we then hope you will relocate your projects to idioma :)

There's no time to waste when it comes to ensuring you have TMs and also keeping them up to date.

Contact us TODAY to get your TM created!