Translation Memory
In the Translation Memory tab, you can:
How to access translation memories
You can access translation memories by clicking on TM in the top black ribbon.
Right-click on the TM tab.
Select Open in a new window from the context menu.
Finding and editing segments in the TM
Note
The translation memory is assigned to the XTM Customer and a language pair.
You can search for a specific phrase in the translation memory or search the whole stored translation memory for a specific language pair assigned to a customer.
Select the TM > Manage tab.
Select Customer.
Select the Source language.
Select the Target language.
Enter the Search phrase if required.
Select whether to search in:
source.
target: to search for multiple target languages, select the Add/Edit target languages icon
Specify the match type by selecting one of the following options:
Phrase match.
Exact match.
Select Use stemming
Select Case sensitive to narrow down the search.
Select Search.
The number of segments found and the segments themselves are displayed in the lower half of the screen.
Once you have found the required segments you can edit either the source or target phrase manually, change the segment status or use the find and replace feature.
Approved – which are final and approved translations.
Not approved – translations that are still being worked on and have not been approved at any stage of the workflow.
Hover over the
icon to show any metadata stored with the segment.
Find the relevant Translation Memory segment.
Make the required changes in the segment. You can change the target and the source.
Click away from the segment. The edited segment is marked with an edit icon
.
Warning
You cannot reverse this action!
Enter the word or words to be changed in the Search phrase field and the word or words to replace them in the replace with field.
Use a combination of replace and next buttons or the replace/next button to work through the segments.
If you want to replace all the segments in one go without checking them further, click the “Replace all” button. XTM will show a message to say how many segments have been changed.
Warning
You cannot reverse this action!
In the TM > Manage, search for the relevant TM segment. Enter the required segment.
Select the Delete segment button
on the right-hand side of the list.
Warning
You cannot reverse this action!
In the TM > Manage, search for the relevant TM segments.
Select the Delete all the segments on this page button
on the left-hand side of the list.
Go to TM > Manage.
Search the TM segments. Select the TM segment.
In the right-hand toolbar, select the Check current segment icon
.Spelling errors are highlighted in yellow. Hover over the highlighted word to bring up spelling suggestions.
Importing translation memory
Translation memories can be imported into XTM Cloud from:
TMX.XLS.XLSX.bilingual
XLIFFfile (one XLIFF file per language pair).
Tip
For the best results, we recommend using the TMX v 4b format.
Note
A source segment with no matching text in the corresponding target cell will not be uploaded.
A target segment with no matching text in the corresponding source segment will not be uploaded.
Empty source and corresponding empty target cells will not be uploaded.
Select the TM > TM import tab.
Select Customer.
Specify the Import project name.
Select the source and target languages.
Specify the Import or set approved/not approved status. Select from the dropdown:
Set as approved.
Set as not approved.
Import statuses and set as approved if missing.
Import statuses and set as not approved if missing.
Select the preferred options for Whitespace in TM. Matching is affected by the differences in whitespace processing.
Optionally: Select the checkbox to Import alt-trans elements from the XLIFF file if you wish to import the matching data that is stored in your XLIFF file.
Optionally: Select checkbox to Import segments with the same source and target. Select if you want to import translation units that have the same source and target texts.
Upload the file you wish to import.
Select the Import button.
A new record in the import history section is created that displays the status of the import. The status changes from In progress to Done when the import is finished.
Exporting translation memory
Select the TM > TM export tab.
Select the source and target languages.
Optionally: select the project and the status.
Select the export file format type.
Select Export.
Note
Segment ID matching must be enabled for this XTM Cloud instance
Select the TM > TM export tab.
Select the source and target languages.
Optionally: select the project and the status.
In the Segment ID select the three dots ... to open a popup.
Type or paste multiple source file segment IDs to the Segment ID popup.
Important
Enter only one segment ID per line!
Select the export file format type.
Select Export.
XTM Align
With XTM Align you can quickly and easily create bitexts (parallel texts) from your previously translated documents. The bitext contains a source document and its translation aligned at the segment level. The bitext can then be uploaded as translation memory into XTM Cloud for use in future projects.
XTM Cloud supports Unicode emoticons in the bilingual terminology extraction and in the Aligner.
Aligning documents in XTM Align
Important
XTM Align and bilingual terminology extraction are not available on freelance XTM Cloud and group accounts. You need align access rights to use these features!
Note
The process of alignment may take some time when the glossaries are generated for the first time. It depends on the size of the documents. If no bilingual dictionaries exist in XTM Cloud, the process is slower.
Select TM > Align.
Type a Reference name for this alignment process.
Select the source and target languages.
Upload the source and target files
Optionally:
enter the Customer name.
Filter template.
Include inline elements.
Select the Align button.
When the alignment is finished the status will change to Done.
Select the Download button in the Status column. The Excel file containing the aligned text is downloaded.
XTM Align creates two Excel spreadsheets for each alignment:
One with 90+ added to the name of the file containing only segments with a 90%+ probability of accurate alignment. This spreadsheet can be used to quickly check and upload the memory, as it requires only visual confirmation of the alignment without any further work.
The main spreadsheet contains the full alignment details: individual cells can be deleted, and target text modified to reflect source text segmentation.
Important
You cannot edit the source segments in the spreadsheet: they cannot be merged or edited in any way as they represent the way in which the source text has been segmented.
You can edit the target segments to reflect the source segmentation if required. For example, you can merge multiple target segments into one cell.
Bilingual terminology extraction
Important
To extract bilingual terminology during alignment you need align access rights.
Bilingual terminology can be extracted in an Excel format when parallel texts are aligned. Terminology extraction is an AI-driven process during which translated term candidates are automatically extracted along with their respective translations. As a result, a highly specialized dictionary is compiled, which contains terms in source and target, along with context metainformation.
The bilingual terminology extraction feature excludes generic words and terms that are already in the terminology of a given Customer.
Check Language pairs available in terminology extraction.
Select TM > Align.
Type a Reference name for this alignment process.
Select the source and target languages.
Upload the source and target files
Optionally:
enter the Customer name.
Filter template.
Include inline elements.
Select the Extract bilingual terminology checkbox.
Select the Align button.
When the extraction process is finished the status will change to Done.
Select the Download button in the Status column.
The downloaded folder contains the alignment result along with the extracted bilingual terms in a separate Excel file.
Excel file description for alignment
The output of XTM Align is an Excel file.
The columns in the Excel file for alignment | ||
Column | Title | Contents |
|---|---|---|
A | Src. Seg | The source segment numbers |
B | Source language code | The source segment text |
C | Target language code | The target segment text |
D | Tgt. Seg | The target segment number |
E | Probability | The probability score for the match on a scale from 0 to 1, where 0 indicates zero probability and 1 indicates an absolutely accurate probability |
In addition, cells are given a colored background to reflect matching accuracy.
Color matching accuracy in the Excel alignment file | |
Color | Matching accuracy |
|---|---|
| Unmatched cells in either source or target |
| Uncertainty as to the quality of the match |
| Text from multiple segments has had to be merged into one cell |
| A very good match |
none | A good match |
Important
In all cases, the alignment of the segments must be checked by a Project Manager or linguist prior to uploading the spreadsheet into XTM
Correcting the alignment in Microsoft Excel
Important
Source text can never be modified!
When deleting cells in either the source or target columns, always choose to Shift cells up.
You can only split or merge text in target cells.
Target language cell text can be split into multiple cells if required to reflect the translation of the source segments, e.g., target cell text that represents two or more source segments may be ‘retranslated’ to reflect the source segmentation.
Text in multiple target cells can also be merged into one cell if required.
After you have completed the review and correction process in the spreadsheet, you can import the TM into XTM Cloud. Check Importing translation memory.
Language pairs available in terminology extraction
Languages | Arabic | Chinese | English | French | German | Japanese | Polish | Spanish | Turkish |
|---|---|---|---|---|---|---|---|---|---|
Afrikaans |
| ||||||||
Arabic |
|
|
|
| |||||
Assamese |
| ||||||||
Bengali |
| ||||||||
Bosnian |
|
| |||||||
Bulgarian |
| ||||||||
Chinese |
|
|
|
|
| ||||
Croatian |
|
| |||||||
Czech |
|
|
| ||||||
Danish |
|
|
|
| |||||
Dutch |
|
|
|
|
|
| |||
English |
|
|
|
| |||||
Estonian |
| ||||||||
Finnish |
|
| |||||||
French |
|
|
|
|
| ||||
German |
|
|
| ||||||
Goan Konkani |
| ||||||||
Greek |
| ||||||||
Gujarati |
| ||||||||
Hebrew |
|
| |||||||
Hindi |
|
| |||||||
Hungarian |
|
|
|
|
| ||||
Indonesian |
|
| |||||||
Irish |
| ||||||||
Italian |
|
|
|
|
|
|
| ||
Japanese |
|
| |||||||
Kannada |
| ||||||||
Kinyarwanda |
| ||||||||
Korean |
|
|
| ||||||
Latvian |
| ||||||||
Lithuanian |
| ||||||||
Macedonian |
| ||||||||
Maithili |
| ||||||||
Malay |
|
| |||||||
Malayalam |
| ||||||||
Marathi |
| ||||||||
Nepali |
| ||||||||
Norwegian |
|
|
| ||||||
Persian |
|
|
| ||||||
Polish |
|
|
|
|
|
| |||
Portuguese |
|
|
|
|
|
| |||
Romanian |
|
|
| ||||||
Russian |
|
|
|
|
|
| |||
Serbian |
| ||||||||
Sindhi |
| ||||||||
Slovak |
|
| |||||||
Slovenian |
| ||||||||
Spanish |
|
|
|
| |||||
Swedish |
|
|
| ||||||
Tagalog |
|
| |||||||
Tamil |
| ||||||||
Telugu |
| ||||||||
Thai |
|
| |||||||
Turkish |
|
|
|
| |||||
Ukrainian |
|
|
| ||||||
Urdu |
|
| |||||||
Vietnamese |
|
|
Language pairs supported by Systran Neural Fuzzy Augmented
Languages | Arabic | English | French | German | Turkish |
|---|---|---|---|---|---|
Arabic |
| ||||
German |
|
| |||
English |
|
|
|
| |
French |
|
| |||
Turkish |
|