Machine translation originally came about as early as the 1940s and has undergone significant changes and developments since to become a valuable tool in many industries. Put very simply these engines take previously established linguistic rules, add algorithms and appies them to new phrases to produce a translation.
Translation engines can work in variety of ways. Some simply take these linguistic rules and apply them to the languages to be translated, whilst others, like Google, take a more statistical approach, looking at patterns in previously (human) translated documents and making assumptions on language based on this analysis. This kind of approach is generally a lot more successful, but does rely on a large body of texts in both languages in order to create its engine. Google overcame this problem by using bilingual texts from the UN, providing it with a huge source of information.
However, whilst machine translation has its benefits, and in some technical areas can prove to be a real time and money saver, particularly when combined with the post-editing skill of a real-life human translator, for audiovisual translation, it’s a different picture.
This is particularly the case here at GoLocalise, where the translation we do is often tailored specifically for voiceover and subtitling. Our translators must ensure that voiceover translations match the original script in length and style so that they will be able to sync to the original video, whilst translation for subtitling requires our suppliers to stick to strict parameters of line length and reading speed. This kind of translation demands more than a literal translation of words and requires translators that can think outside of the box to overcome all the linguistic, spatial and temporal challenges – certainly something that no machine is able to achieve (at least, for now…!)
So, coming back to my original question, and the title of this blog, when it comes to accurately communicating your audiovisual message around the world, and ensuring that your content is understood and enjoyed globally – there can only be one answer – GoLocalise.