Please see our main article about Analysis for general information.
The Default Analysis is the standard analysis run before translation. It provides a breakdown of the segments/words/character counts, shows matches with the translation memory (if used in a project), displays non-translatable matches and Machine Translation Quality Estimation (if enabled), and can produce a quote for a client.
The Default Analysis is always working with the source text—comparing the original file with the content of the translation memory, calculating repetitions, etc.
Note: Launching this type of analysis after the job is translated can give you false results.
To run an analysis, open the project, select one or multiple jobs and click on the Analyze button. Once you do, you'll see the following options:
- Type (Select Default)
- Name - If left empty, it will be automatically named "Analysis #1", etc. Renaming macros can be used to generate names automatically.
- Include internal fuzzies - Segments in selected documents are compared with one another and when similar segments are identified, they are included in the analysis as fuzzy matches, depending on the similarity.
- Include non-translatables - Segments are analyzed for non-translatables. Non-translatables are anything that doesn’t necessarily need to be translated. Examples of these include names, numbers, email addresses, URLs, symbols, and codes. If non-translatables are identified, they are included in the analysis. The AI-powered non-translatables can have a score of 100% or 99%.
- Include machine translation matches (Beta) - With the MTQE feature enabled, you can run a default analysis that includes the MT scores.
- Exclude confirmed segments
- Exclude locked segments
- Exclude numbers
- Analyze by provider - When multiple jobs are selected, they will be grouped by the assigned provider and analyzed together. The Analysis will be assigned to the provider and will be visible to them.
- Analyze by language - When multiple jobs are selected, they will be grouped by language combinations and analyzed together.
- Count units of the: Source or Target - Generally, users should always use Source to get the right results since the target of some segments will most likely be empty.
The results can show:
- Net rate discount
- Repetitions (If the repetition is also a 100% or a 101% TM match, it will be listed in the respective TM match row.)
- Translation memory matches 101% - 0%
- Non-tranlatable matches 100% - 75%
- Machine Translation Quality estimation 100% - 75%