Idea Delivered Partially

See comments and RWS Documentation.

Enable Trados to automatically populate all tags and formatting when pre-translating with a translation memory imported via TMX

Currently, when using Google Machine Translation in Trados, all tags and formatting of the original text (including font color, size, italics, various colors of highlighting, underlining, etc.) are well preserved.

For example:

Screenshot showing Trados Studio interface with highlighted text and hyperlinks preserved in Google Machine Translation.

Screenshot displaying a warning in Trados Studio about the inability to retain original formatting when using translation memory from a TMX file.

However, when using the translation memory（sdltm） imported from a TMX file that primarily consists of plain text without tags or formatting, after pre-translations in trados with this translation memory（sdltm）, we find that trados either cannot automatically match the translation due to less than 100% match or, even if a match is found, trados cannot automatically populate tags and retain the various formatting of the original text (including font color, font size, italics, various colors of highlighting, underlining, etc.).

For example:

Screenshot of Trados Studio's translation memory search interface with a note on importing from a TMX file.

Screenshot illustrating Trados Studio's failure to match translation and retain formatting with less than 100% match from TMX file.

Screenshot of Trados Studio showing mismatched formatting and tags in the translated text compared to the original.

Please add the function in Trados to allows Trados to automatically batch pre-translate with the translation memory（sdltm） imported from a TMX file as long as the content of words and numbers of the sentence in the source text is the same as the sentence in the translation memory（sdltm） imported from a TMX file. Trados should accurately autofill all corresponding tags in the targeted text and adjust the formatting to perfectly match the original sentence (including font color, size, italics, various colors of highlighting, underlining, etc.).

jack zhong
  • in reply to Daniel Brockmann

    Dear Mr. Daniel Brockmann,

    Thank you for your reply; you are a renowned expert in Trados.

    I have just tested the method you mentioned. Although I have enabled the "Fragment Alignment Status" to "on" and switched to a translation memory (TM) with over 1000 units (though with low relevance to the current source text), I tried clicking "auto-insert tags" on several different interfaces, but none of them worked.

    I have looked into the feedback from other users on the Trados community, including a reply from Mr. Paul: I suspect that the issue might be due to the top-loaded TM being large but having low relevance to the sentences that need to be translated in the current document.

    However, all of these do not fundamentally address the actual needs of translators.

    The current situation is: even when there is a highly matched translation memory (sdltm ) imported from TMX with fewer than 1000 units, can Trados achieve automatic and rapid filling of various tags of source text and formatting (including font color, size, italics, various colors of highlighting, underlining, etc.)? In practical translation work, we expect to improve efficiency through Trados. Many small translation projects will not accumulate 1000 units of TM from start to finish. If all these small projects rely on manually filling in these tags and modifying various formats in the target language(including font color, size, italics, various colors of highlighting, underlining, etc.), then the efficiency of translation work is far from improved.

    Is it possible to meet the needs of translation work through the following method: Current artificial intelligence (such as ChatGPT, Google machine translation, or Trados's own machine translation capabilities) can understand the meaning of source and target texts quite well. Could AI, understanding the meaning of the source language and target language in each trados segments, automatically insert tags at the correct positions and adjust various formatting in the target language (including font color, size, italics, various colors of highlighting, underlining, etc.) at the precise word or phrase positions to match the original language?

    For example, achieving the following effect:

    Screenshot showing Trados Studio segment with source text in Chinese and target text in English, highlighting the need to keep original formatting.

    Screenshot of Trados Studio with an English source text segment and its Chinese translation, showing hyperlinks and formatting retention.

  • Once the translation memory is large enough, Trados Studio will attempt to automatically add the tags as part of the fuzzy match repair functionality. For that to happen, you would need to enable fuzzy match repair for batch pretranslation. This will also work during interactive translation. See RWS Documentation for more details. Setting status to "Delivered Partially".