How Machine Translation Post-Editing can save you time and money

Join our mailing list

Sign up for our industry insights and free translation advice

Get in touch

First, what’s MPTE? 


Machine Translation Post-Editing is a fancy way of saying correcting the mistakes automatic/machine translators make. But wait, if we’re just running your text through an automatic translator, why are you paying us for the privilege? Because there’s far more to it, and we’d love to tell you what.


Every year the media is flooded with news items about how machine translation is taking over from humans, but we could not be happier to tell you: it certainly isn’t! 


If you’ve ever used a machine translator to translate into your own language, you’ll know they can be very hit and miss. Why? Context. Sure, machine translating tools have come a long way in the last few years, and sometimes they’re spot on, but other times what they produce can’t be deciphered no matter how hard you try. 


So why are we using them? Because they speed up our processes. If used in the right way, in fact, machine translation post-editing can cut costs and speed up workflows, which is a win for both you and us. Now keep reading to find out about types of machine translation proofreading services available and how the process works. 



Levels of MTPE


Now that we’ve got the basics covered, let’s get into the nitty-gritty. There are several levels of MTPE. What do we mean by that?


Raw Machine Translation (MT) is output that is not proofread by human translators at all. Used for low-risk content that needs to be translated for informal purposes only in a short space of time, it’s the lowest level.


Light MTPE makes the text understandable. Human translators comb the MT output to find the bits that make no sense and revise and edit them, making this a popular choice for documents with high volumes in various sectors. 


Full MTPE workflows try to match “human translation-from scratch” in terms of quality, outperforming the latter in terms of costs and timelines.


Each of the three options has its price and time considerations, and the 3rd option is far more human-reliant than the other two, but we still haven’t answered the question: couldn’t you just run your texts through google translate, yourself?



Post Editing Service Providers – What we do


You could. But that’s not what we’re doing by any stretch of the imagination. In fact, we’re often doing what google translate itself does – we’re producing the automated content by training an engine to do it ourselves. 


Don’t trust automatic translators? But how about one designed by us, fed data by us and nurtured so that it has more of that all-important thing most free automatic translators are lacking? Context. 


And one step further?



Integrating Translation Memories with Machine Translation Post-editing


A translation memory, you say? Yes. It’s a document that stores all the translations we’ve ever done for you in any language pair. The beauty of it? When the same string of words comes up again, we don’t need to translate it again from scratch, or have a machine translate it, because we have the translation already. 


When you combine what we’ve already translated and the output of the machine trained by us, it’s safe to say the process is far more streamlined – especially when you compare it to having someone type out every word. 


Looking to cut down on turnaround times and budgets for your large-scale translation projects? We can’t wait to get started.