Welcome to our second post about Transit NXT and machine translation. This post is about Transit’s “Editor MT” and it is relevant for translators / MT editors, reviewers and project managers alike. Translators will see how they can manually or automatically retrieve suggestions from the machine translation engine from within Transit’s translation editor and how they can be edited. Reviewers and project managers will learn how they can track the use of MT.
We will use GoogleTranslate as our machine translation engine for the purposes of this tooltip, but the options we show apply to both GoogleTranslate and itranslate4.eu. Read our previous post about Transit NXT and MT to learn what steps are necessary to configure and use any of these two machine translation engines from within Transit NXT’s translation editor. Once you have followed these steps, you can start.
First, open the language pair(s) of the project in which you wish to use the configured MT.
Request machine translation proposals manually
To manually request a translation proposal from the machine translation engine, you can right-click in any segment you want to translate and select Request machine translation from the context menu as seen in the figure below:
Transit sends the segment to the configured MT system and the result is displayed in the Source Fuzzy window as seen in the figure below:
Request machine translation proposals automatically
You need to make the appropriate settings in the User preferences if you wish to send your segments to the MT system automatically instead of having to select the option from the context menu for each segment. For this, select User preferences | Machine Translation and select the Automatically request translation suggestions option as shown in the figure below:
Then you can establish the fuzzy match threshold that you want to use for the MT search. If, like in our example, you establish a threshold of 70 in Only for segments with fuzzy matches lower than (%), Transit will automatically send segments that have a lower fuzzy match percentage to the MT and then display the machine translated segment as a translation proposal in the source fuzzy window. In the following figure, MT indicates that the suggestion was received from the MT system:
Everything else is exactly the same as if you were working with proposals from your translation memory. You can insert terminology, edit the MT proposal exactly as if it were a fuzzy match and then paste it into your target language window with ALT-ENTER. Depending on the User preferences you have established for Dual Fuzzy you might receive fuzzy match proposals from the reference material in addition to the proposal from the MT engine so you can choose which one needs less editing and / or suits you best.
If the segment contains markups, the display is extended by a row. In the additional row, Transit NXT displays if the segment text has been transferred to the MT system with or without markups. If the respective MT system does not support the processing of markups, Transit NXT automatically transfers the segment text without markups. In this case you will have to insert the markups after accepting the MT suggestion for your translation. For more information on how to insert markups manually, please check out our tooltip on markups Inserting markups on the fly. The following figure shows an example of the segment with markups sent to the MT system:
You can then insert the MT proposal into the target-language segment by using the ALT+ENTER shortcut.
As a project manager, you have have various means to control the use of MT:
- Check the Edited by field of the Notes window to find out whether the segment was translated by a user or by the MT system as seen in the figure below.
- Filter segments edited by the user or by the MT system by using Transit’s segment filter. This can be done through the Segment attributes tab from View | Segment filter | Create as seen in the following figure:
Please refer to our Filtering segments in Transit NXT blog post for more information on filtering segments.
This is how the Editor MT can be used in Transit. Once you have your API key, you are all set to use this feature offered by the new service pack.
Please feel free to leave your comments. We will soon be back with another tooltip on machine translation!