Let’s get into the more technical side of things. I am sure you have heard of this big phenomenon that is supposedly going to steal translators’ jobs: machine translation. After the introduction of Google Translate in 2006, this has been a hot topic in each translation classroom. Will I lose my job as a translator because of machine translation? And the answer never really changed: not really.
Machine translation indeed affects those in the translation industry, but what didn’t artificial intelligence affect? In some ways, it made our lives easier; in other ways, it made our lives harder. Today, we will focus on machine translation and MTPE (machine translation post-editing). This is going to be a two-part series: we’re going to explain MTPE, and we’re going to talk about why post-editing is necessary and machine translation in game localization in the first part. In the second part, we’ll talk more about the advantages and disadvantages of using machine translation. Is machine translation advantageous or disadvantageous? What is MTPE? If these are the questions you have in your mind, keep on reading!
So, what is machine translation? Basically, machine translation is translating texts from one language to another by using artificial intelligence. This technology has gotten so much better since the 1950s, but it is still not as accurate and creative as a human translator. There are different kinds of machine translation systems. Let’s not go into detail, but the most popular one out there is the hybrid machine translation system where statistical machine translation and neural machine translation are used together to deliver the best MT output.
As with all AI applications, translators learned to adapt to machine translation (MT) by editing the MT outcome. We call this process MTPE, i.e. machine translation post-editing.
1- Machine translation engines “pre-translate” the text.
2- Human translators edit the MT output to deliver a better translation.
Is post-editing essential after machine translation? We believe that in order to capture the essence of the source text, a human translator’s touch is necessary. We’ll dive deeper into why this is needed.
Trusting artificial intelligence completely, and not checking afterward can be quite detrimental to whatever business you have. You might be a content marketer who is getting help from ChatGPT or you might be a translator working with machine translation. There is absolutely nothing wrong with getting help from these miracles of technology, but the ideal scenario is always AI combined with human touch. So, using machine translation does not mean that you are not a good translator, but if you just use whatever comes out and don’t check the output, then we might start to question whether you are doing a good job or not.
Machine translation in game localization is a bit tricky. We’ll explain this with examples. Let’s say your first source text is a manual explaining how to use a fridge, and your second text is a script from a video game. You’re using machine translation for both in the translation process. And let’s think about the scenario where you don’t post-edit, but the MT output is linguistically correct. The problem here is that even if the MT output is correct, it’s probably not going to make any sense in the second text.
The first one works alright, it’s a manual, and you don’t need to get creative for the text to serve its purpose; even if you translate word by word, you can get the general idea about how to use the fridge. However, game localization necessitates creativity. Therefore, using machine translation in game localization and not getting any help from a human translator or a game localization specialist will result in a bad translation.
There are studios out there that want to include other languages in their games, but they don’t want to spend any money on localization, so they use machine translation. They might think that they don’t need professional game localization because maybe they are just translating the user interface, so it’s not “that important”. Well, I have news: it is that important. When you download a game after seeing that there is language support for your native language, but then see that it’s kind of gibberish, it’s natural to not want to play the game.
To sum up, we’d like to emphasize again that machine translation plus a human translator is the way to go, especially if your text needs creativity just like in the case of a game localization project. Don’t forget to check out our blog page to see our next post where we’ll move on to the advantages and disadvantages of MTPE.
Written by
Ecenaz Batur
Marketing Specialist
Looking for getting tailored solutions for your game localization projects? Let’s open up a whole new world together!