Turbo tool for quality and cost/time savings

Translation memory systems

Translation memories pay off

Efficiency is of great importance when translating more and more documents and content. This makes translation memory systems (TMS) an indispensable part of everyday translation work. A TMS is a program that segments the texts to be translated according to certain rules (e.g. sentence by sentence) and stores these segments with source and target text in a language-specific translation memory (TM). This allows us to use your previously translated texts for follow-up jobs, reducing turnaround time, lowering costs and ensuring consistency.

What are the benefits of a translation memory system?

There are three main advantages to using a TMS:

1) Faster turnaround times
Translators can use existing translations from the TM and simply check them in the new context, speeding up the translation process as a whole. This is particularly noticeable in technical documentation and large operating manuals, in which entire chapters are often reused.

An interface between the TMS and order management system also streamlines project management meaning we can process your orders even more quickly.

2) Lower costs
As soon as we receive new documents for translation, they are analysed using the TMS. Comparison with the TM detects matches from previous translations and identifies repetitions between the new documents or within a document. Since repeated text is not always retranslated which, in turn, reduces the effort required, all repetitions and matches are prorated.

3) Higher quality
The TM is integrated into the TMS throughout the translation process and is available to the translators. The TM provides suggestions to the translator during translation or they can actively search for specific phrases, expressions or technical terms to ensure consistency within the document and with previous translations.

The TMS also has various review tools, such as a QA checker, which verifies that all segments have been translated, ensures that numbers and formatting are correct, confirms whether tags have been placed correctly and more once the translation is complete. If a terminology database is available, the translation is also checked for the correct use of technical terms.

How does the translation process work in the translation memory system?

After we receive your request, we put the documents to be translated into the TMS. For this, the source file must be editable and well formatted. PDF files can now also be imported, but segmentation and formatting are often not ideal, which can lead to errors in the TM and thus cause difficulties with subsequent jobs.

An analysis is carried out in the TMS to determine the number of words and compare them with the TM. We use this analysis to put together your quote.

Once you have accepted the quote and given us the go-ahead, the files are sent to the translator. We use our GroupShare server solution for this, meaning it takes just a few clicks to send the translation to the relevant translator. The file can also be sent directly from the translator to the reviser via the server, meaning there is no need for e-mails.

Once the revision is complete, we carry out a final check in the TMS using the QA Checker. The target file is then created and the translation memory is updated.

As a certified partner of RWS, we work with the TMS Trados Studio and the associated terminology management solution MultiTerm. The terminology database (TB) is directly integrated into Trados. If a new technical term appears in the source text, it can be marked in the source and target language and inserted ad hoc as a new entry in the TB. This means the terms stored in the TB are displayed as a suggestion during future translations and can be used directly by the translator – this ensures your translations are consistent.

Machine translation (MT) engines can also be integrated directly into the TMS. When creating the project, the text is pre-translated by machine and then edited by our experienced post-editors. The post-edited translations are also stored in a TM so that subsequent MT jobs can first use the already post-edited texts. All new texts are then populated by the machine translation and post-edited.