Machine translation has historically suffered a poor reputation. That was then. Having made huge leaps forwards, translation technology is now capable of producing texts that are barely distinguishable from human translations.
All thanks to statistical calculations and neural networks.
Texts are composed based on estimations. Given our translation machines have clocked up an impressive number of words, the quality of texts is constantly improving.
Where speed and price take priority over quality, machine translation offers a solution. Need to get that mammoth of a tender over to your global teams like, yesterday? Machine translation gets the job done in no-time.
Not all texts are unique. Particularly in the case of technical texts, much of the material has been written before: brochures and manuals are updated, products in the same category are very similar. And all that repetition saves on translation work!
Each text we translate is stored in smart databases known as translation memories, which include both the original text and its translation. Every phrase that is translated is subsequently checked against the translation memory.
If the phrases are 100% identical, then they’re a complete match and we can move on to the next. If the phrases are similar but not identical, then they are referred to as a ‘fuzzy match’, in which case we’ll use the translation stored in the database as a starting point.
All of this saves time and money. Plus, it allows us to produce more consistent translations.
As you may have gathered, the translation industry is becoming increasingly technical in nature. Machines are performing the work of translators, and you’d never know it. But the time is right, given that volumes of translation work are growing as deadlines are becoming ever more narrow. That said, a human touch remains indispensable.
Literal translations can fall (way) short when it comes to humour, proverbs and cultural nuances.
It also takes a human touch to pick the right translators and translation technology tools for a given project. Databases and machines aren’t the only smart things we work with: we have smart people, too. These people are language experts who ensure that content, register and emotion are all appropriately conveyed in your target language.