Translation Memory
The core of MultiTrans Prism is our revolutionary TextBase Translation Memory. TextBase is unique among translation memory tools in several ways:
- A corpus-based approach for a contextual, multilingual and multidirectional TextBase TM
- Advanced multilingual search capabilities, transforming a TM into a reference database
- Fully automated alignment of parallel texts with amazing accuracy thus eliminating tedious human effort
- On-the-fly alignment of documents/segments, making a TM into a dynamic document
- Can Identify and leverage content on multiple levels from full paragraph matches, to sentence or traditional segment level matches, to sub-segment matches - an approach called "Advanced Leveraging"
Corpus Approach
MultiTrans Prism TextBase corpus-based approach means that when a translation memory is built in MultiTrans Prism it incorporates the entire text of your document and indexes these as a corpus. This is opposed to the traditional translation memory approach which is to cut up a document into segmented chunks (text separated by either punctuation or paragraph returns). The advantage of the corpus approach is that all of your segments remain in the document order and in their natural context within the actual document. This is immensely helpful to the translation process as translators and editors can now see how text was used within its natural paragraph which can influence style and tone.
With this corpus approach you also have the option of making your TextBases multilingual and all TextBases are multidirectional. TextBase TM allows you to manage large volumes of translated content into an unlimited number of languages in a single, multi-directional structure. This dramatically reduces your server storage requirements and administrative burden. For example, conventional translation memories require that a separate database be constructed for each language pair. For a global organization that operates in 20 languages, for example, this could mean up to 190 separate databases to cover all language combinations. MultiTrans Prism therefore simplifies and reduces your administrative burden. Multidirectional means any TextBase TM can be "flipped" in language direction and remain equally valid, retiring the notion of TM source and target language. This eliminates the need to maintain separate translation memories for each language direction (for example, English>French and French>English).
Search Capabilities
Since texts are stored in a TextBase TM as a whole document corpus it is ideal to use as a searchable reference database. MultiTrans Prism offers sophisticated search options to retrieve text strings of any length, such as terms, sub-segments, full and fuzzy segments and paragraphs. The concordance searches not only return the translation results of all occurrences but users can also easily see where in the documents each occurrence originates. The context of each individual occurrence is available along with all relevant metadata. These features make your translation memory into a bilingual research document where you look up words, phrases, expressions and sentences and see how they have been used in different contexts.
Automated Alignment
MultiTrans Prism TextBase TM uses sophisticated algorithms to align the translated text with the source text. Since this alignment process is automated to a highly accurate degree you do not need to do alignment and manual verification as a separate step before you can use the translation memory. With MultiTrans Prism you upload your documents for alignment and it creates the TM directly. Not only do you get a TM directly, but the alignment agent is incredibly fast, you can create new translation memories of over 10,000 segments in under 5 minutes. A TextBase TM with millions of words can be built in less than an hour and be ready for immediate use. A conventional TM database of comparable size would take years to build!
On-the-fly alignment
Since the alignment agent creates your translation memory directly in the TextBase TM, alignment functions to change or modify segments always remain available. In the TextBase TM, alignment is not a one-time process but something that can change throughout the life cycle of the TM. Although the automated alignment agent is highly accurate, users have the option of modifying these by expanding or shrinking the source or target segments to accommodate multiple segments to one correspondence. Users can also confirm alignments they have reviewed and even assign segments penalties or bonuses, based on their confirmation status, to allow confirmed segments to rank higher in analyses.
Advanced Leveraging
MultiTrans Prism has been described by the Translation Automation User Society (TAUS) as one of the new generation Advanced Leveraging TM technologies. Advanced Leveraging offers to unlock additional value from your translation memories by allowing you to mine data on the sub-segment level which can greatly increase the amount of leveragable content.
How Advanced Leveraging TM compares to Conventional TM
The Advanced Leveraging TextBase TM indexes your integral documents in a parallel manner. Since it does not divide your document into individual sentences, it can easily align paragraphs and sentences nearly perfectly without any human intervention. The context of the entire source and previously translated document is preserved so that in the rare cases of misalignment, you can easily realign a sentence on-the-fly. The Advanced Leveraging TextBase TM can be multilingual and multidirectional, eliminating the need for duplicating bilingual memories.
.png)
A conventional TM divides your document in out-of-context sentences, and then tries to align them. It can be compared to an Excel worksheet, where, for instance, in column A, you have the English, and in column B the French sentences. It is very tedious to check that each sentence or sentence group is perfectly aligned because the context has been lost, and correcting the alignments manually is almost impossible.
For more information on Advanced Leveraging TM (ALTM) please click here.
The TextBase works in conjunction with the TermBase in the Analysis Agent to identify leveraging and prepare documents for translation. Likewise it works with our Translation Agent and XLIFF Editor to interactively process documents for translation.
To see a complete list of all our products, please click here.
