Translation technology: our top three trends for 2019
Want to know what’s on the horizon for translation technology in 2019? Well, look no further. The Brightlines team have been looking at what we believe will be the three most influential developments in translation technology throughout 2019.
1. Audio-visual (AV) Localisation
The on-demand entertainment industry is booming. Streaming platforms such as Netflix, Amazon Prime and Now TV provide users with what seems like an endless pool of entertainment at the click of a button. But how do these platforms adapt their content to global audiences? That’s where AV localisation comes in!
The process of AV localisation involves adapting the speech and text elements of a tv show or film to appeal to a specific audience. Localising in this way involves some of the following elements:
- Translation of the script
- Awareness of cultural nuances
- Matching the speed of the target language to the source
And these are just to name a few!
Project Voco: A new technological advancement was brought to light in November 2016 at the Adobe MAX conference: Adobe presented Project Voco, a tool that allows the user to imitate a specific human voice speaking words and phrases from another language. For example, the software could analyse the voice of someone saying ‘hello’ and recreate the sound of their voice to say ‘bonjour’.
At the moment, Project Voco is still in the development stages and no release date has been announced – perhaps 2019 could be the year! It’s still not clear whether the platform will enable multilingual support, so keep your eyes peeled for details when it’s released.
2. Voice Recognition Technology (VRT)
You might be thinking voice recognition technology isn’t new, and you’d be right. However, the use of VRT is growing – as is its popularity in the translation industry. By integrating VRT with machine translation, the technology can be used to quickly dictate a base-layer translation.
This technology can be beneficial for those who are translating text into their native languages – the flow and speed at which the baseline translation can be created is definitely quicker than writing it yourself! Although the content will still need to be proofread by a professional translator or agency, it’ll reduce the overall translation time. It’s also a useful tool for translators with disabilities or visual impairments, making large jobs easier and faster.
So why exactly do we think VRT is one to watch in 2019? Well, one of the reasons VRT hasn’t fully caught on in the translation industry is because of the limitations that come with the technology – which we think could be overcome in 2019.
One of the main problems with VRT is its inability to distinguish the difference between commands and words that actually need to be translated. The technology can also struggle to interpret different dialects and accents, and large pieces of technical copy can be difficult to translate – much the same as the problems we see with machine translation.
You also might be familiar with translation errors that crop up when VRT is used on live news channels such as the BBC. The most apparent mishap that has hit our screens recently is the translation of Michael Gove claiming Batman simply cannot be the Prime Minister… well said, Michael!
While some of the translations are hilariously incorrect – as is evidenced above – it highlights the limitations that are apparent with VRT when it comes to recognising accents and dialects. We think 2019 could be the year that the industry overcomes these challenges, and VRT will definitely be on the radar soon enough.
3. Neural Machine Translation (NMT)
Much like VRT, NMT isn’t new for 2019, but something that we think will be hitting the spotlight soon. If you’ve not heard of NMT before, here’s everything you need to know:
NMT is a data-driven version of machine translation that uses a neural network of translation memories – and by this, we mean words, phrases and full sentences – to predict the most relevant translation. But why is this in our top three for 2019?
New advancements in NMT have been taking place this year through the University of Amsterdam with a project named DatAptor. Researchers have been training a translation machine with datasets that contain words specifically relevant to a chosen topic. For example, translating a specialised piece of technical text would require a certain vocabulary that wouldn’t be required when translating a piece of advertising copy. By training a machine to identify when certain words should be used, NMT is becoming a far more adaptable and intelligent piece of translation technology.
Technology is great – but will it beat human translation?
In short – not yet! It may in the year 2100, but for now human translation remains the most effective method of translation. A professional translation agency – like Brightlines – will be able to offer you the best method of translation to suit your business needs. Contact us today to find out how our translation services can benefit you and your business.