A Project manager or a Linguist can pre-translate job(s) automatically using Project > Pre-translate function.
Following submenu is available (depending on your Memsource Edition):
- Pre-translate where Empty... (see this chapter bellow)
- Human Translate where Empty... (see the Human Translation Engines page for more details)
- Copy Source to Target where Empty will copy source segments into target segments as long as target segments are empty
- Pseudo-translate where Empty... (see this chapter bellow)
- Delete All Translations will clear all target segments of the selected files, this action cannot be undone
Select the job(s) and menu Pre-translate. Please note that only segments with empty target will be pre-translated.
Pre-translate where Empty
Pre-translate where Empty can be set on 3 different pages in Memsource Cloud and can later be used for automatic pre-translation when job is created:
- In the general Setup of Memsource Cloud (for Admins and Project Managers). The screenshot above shows the default settings. The general Setup also includes an option to disable/enable the auto-propagation of repetitions. Admins and Project Managers can decide if translations of repeated source text segments are automatically inputted into the corresponding target text segments. Admins and Project Managers can also choose whether Linguist users can control this setting. See the article on access and security settings. The pre-translate section of the general Setup also includes pseudo-translation options (for Team, Ultimate and Enterprise editions).
- In Project settings (when creating a new project or editing an existing one). These settings also include an option to disable/enable the auto-propagation of repetitions. See bullet point 1 for more information.
- Via the Pre-translate button in Project overview when a job is selected.
Following options are available:
- Pre-translate from translation memory will insert matches from the translation memory into empty target segments
- Pre-translation threshold specifies the minimum score that needs to be reached for translation memory matches to be inserted
- Pre-translate from machine translation will pre-translate from the machine translation engine selected for your project. MT results are inserted only if there are no TM results (above the defined threshold) available.
- Pre-translate non-translatables will copy source to target for segments consisting of only non-letters
- Display non-translatables in Memsource Editors: This option allows you to decide whether non-translatable (NT) matches will be displayed in the CAT pane of the Memsource Desktop Editor and Web Editor. The non-translatable matches have a dashed line beneath them. See the CAT Pane article for more information.
Set segment status to confirmed for:
- 101% translation memory matches
- 100% translation memory matches
- 100% non-translatable matches
- 100% machine translation matches (Beta) when using the MTQE feature.
This feature makes it possible to select which pre-translated segments are to be marked as confirmed during pre-translation (without saving them into the translation memory). This feature can be useful for pre-translating 101% matches, marking them as confirmed and excluding them from an analysis with the exclude confirmed segments option.
Pre-translate & set job to completed
- Pre-translate on job creation (only available in general Setup and Project settings)
- Set job to completed once pre-translated
- Set project to completed once all jobs are pre-translated
For Machine Translation only. This feature can be used for projects created with the Automation widget. Your client submits files which will be automatically pre-translated on your MT Engine (including NT results if enabled) and sent back to your client.
- Lock 101% translation memory matches
- Lock 100% translation memory matches
- Lock 100% non-translatable matches
- Lock 100% machine translation matches (Beta) when using the MTQE feature.
Hint: Locking segments during pre-translation can be only done from PM's account.
Project managers have a few more options in the pre-translation settings than linguists.
In case the linguist runs pre-translate from his/her account, the settings of the project manager are applied when it comes to locking.
In order to unlock segments, the linguist needs to contact the project manager and ask him/her to unlock the locked segments. It is possible that the segments are locked on purpose and should not be translated by the linguist.
Pre-translation Best Practice
It may not be necessary or even advisable to always pre-translate files before translation. Normally, the Memsource Editor will take care of the pre-translation in a way that can be more efficient.
This feature makes it possible to pseudo-translate a file using the following options:
- Pseudo-translate using - selected character wil be added into original text ('Test' will be 'T$e$s$t$')
- Add to target segment start - for example 'aaa' ('Test' will be 'aaaT$e$s$t$')
- Add to target segment end - for example 'bbb' ('Test' will be 'T$e$s$t$bbb')
- Target segment length (%) - length of the target including the pseudo-translate character (at 100% 'Test' will be 'T$e$'), excluding the characters added at start and end ('aaaT$e$bbb').
Pseudo-translate is available from the Team edition up.