As companies are faced with higher volumes of content that require translation, and as the time allotted for these projects shrinks, more and more organizations are weighing the pros and cons of machine translation – MT – as a viable solution to tackle these time-critical projects.
TransPerfect’s proprietary methodology combines machine translation with complementary technologies and human translators to quickly and reliably produce usable translations for review of large volumes of text, all completed in a fraction of the time that would be required for a standard translation process. While machine translation quality falls far short of human translation, if used properly in conjunction with human reviewers, the utility of machine translation can be stretched to include both non-distribution applications including document review, legal discovery, and internal correspondence, as well as distribution-level materials for which extremely fast turnaround times are a requirement.
TransPerfect’s Machine Translation Methodology
- Documents are collected in electronic and hard copy formats
- Front-end Source Material Analysis and Source Sampling help determine the ideal project strategy
- Optical Character Recognition – OCR – software is used on hard copy documents to create electronic files where possible
- All documents are run against existing translation memory – TM – for matching segments
- Remaining text is translated through advanced machine translation – MT – software
- Reviewers comb through machine translated text, correcting basic spelling and grammar issues
- Final files are delivered to the client in electronic format
Through TransPerfect’s scalable MT methodology, human involvement can be increased, decreased, or removed entirely based on the specific objectives of each project. To learn more about this valuable system, please view our MT white paper here.
Please contact us to learn more about our methodology and to talk with our client service team to determine if MT is a potential solution for your needs.