3 key considerations to better understand Machine translation post-editing (MTPE)
In recent years, a new trend in the field of translation has emerged that leaves no one indifferent: MTPE. This method involves a linguistic professional revising or correcting a text that has been translated automatically (as opposed to traditional proofreading, in which the text has also been translated by a person and not by a machine). Assuming that in many cases, MTPE is not an adequate solution for the translation needs of some types of texts due to the lack of context and the fact that it cannot faithfully adapt the message to the target audience, I would like to clarify some issues related to this methodology that are not known to everyone.
- Which are the most common tools for automatic translations?
In recent years the number of automatic translators has grown exponentially due to the emergence of new ICT, although the most common ones are Google Translator Toolkit, Bing Microsoft Translator and DeepL. - What is the main difference between MTPE and proofreading?
A revision of a translation done by one person may undergo few corrections (or none at all) if the quality is good. However, a post-editing job almost always involves changes because the quality of the machine translation often falls short of our expectations. As mentioned above, lack of context, inability to distinguish whether or not to translate a name, incorrect use of capitalization, etc. means that one should never rely on MT alone. - What is the reason for opting for post-editing?
The purpose of post-editing is to improve productivity. However, if the resulting machine translation is disastrous, the professional's intervention will have to be greater, in which case it may be more profitable to opt for human translation from scratch.
