Moving beyond the keyboard
A woman calls a hair salon to make an appointment for a client of hers. The receptionist answers, asks a few questions, and offers a convenient, 10 am appointment for a women’s haircut. Everything is perfectly normal about the call — except one thing. The woman calling isn’t human. She’s actually a computer, trained to recognize conversation and respond appropriately.
Say Hello to Voice Recognition Technology
This is just one example demonstrated by Google’s CEO, Sundar Pichai, at Google IO 2018. (1) Designed as a productivity tool, voice recognition technology works when a computer software program or hardware system interprets the human voice and responds accordingly. Voice recognition technology isn’t new — in fact, one might trace it all the way back to Wolfgang von Kempelen’s first acoustic-mechanical speech machine in 1784. (2) But one thing’s certain — its popularity and use is exponentially growing, especially in the translation industry.
Who in the Industry is Benefiting from this Unique Productivity Tool?
First and foremost, people who are translating natural flow text into their own native languages are benefitting. For example, if an American man living in Germany gets a contract to translate a novel into English, he can use voice recognition technology to quickly do a baseline translation. It’s also useful in situations where translators don’t need to do heavy post-editing work or manipulation of the text.
Finally, another important market niche is people with visual impairments or other disabilities. “For people where typing is difficult, or where visual aids are needed, voice recognition can greatly enhance their productivity. In fact, there are many translators with visual impairments,” says Zsolt Varga, Product Owner at memoQ.
“People with disabilities can now overcome difficulties we’ve never dreamt of.”
Voice Recognition will Eventually Replace Predictive Technology
Today’s global economy means more content needs to be translated quickly. As a result, companies are increasingly relying on machine translation (MT) technology, combined with post-editing workflows done by translators. Voice recognition technology can accelerate the post-editing process and boost productivity for translators, especially for big jobs.
“In 2019, machine translation is probably still going to be the order of the day, and therefore, it needs to be supported by human translators or reviewers,” says Varga.
As a result, we suspect more translation software companies will offer their own voice recognition products in 2019. For example, at memoQ, we have just launched our first voice recognition technology product. Called Hey memoQ, this dictation app allows people to speak directly into their iOS mobile devices. The Hey memoQ mobile app then sends the speech to Apple, where it is processed and sent back to memoQ, allowing the user to see written text on the translation grid.
Because a person can speak more quickly than type, we believe voice recognition technology could eventually replace predictive technology. In fact, our users have gone as far as saying,
“Predictive technology is old technology for people still typing.”
There are still Technical Limitations to Overcome
Even though voice recognition technology exists in the translation industry, its use still hasn’t peaked. That’s because the technology still is limited in several use cases. For example, it can have trouble separating voice commands from the words that actually needs to be recognized and processed.
“The more frequently you have to switch from speech recognition to manual entry and back, the more cumbersome and difficult it is to use the technology,” says Varga. It can also be challenging for voice recognition technology to understand local dialects and accents, like Indian English or Hungarian English. Finally, highly specific and technical content can be difficult to translate as well. For instance, voice recognition technology doesn’t seem to work that well for translating medical text.
But these are all challenges that the industry is ready to tackle head on. In many ways, the use of voice recognition technology in the translation industry is still new, so it’s only natural to face early technical limitations. Software engineers and product experts are collaborating now to brainstorm solutions and continue to innovate.
Imagine Your own Personal, Talking Robot
It might seem like a science fiction movie, but voice recognition technology is quickly making robots reality. Now people can speak into devices that respond like humans and perform convenient tasks, like making salon appointments and restaurant reservations. This type of technology is especially productive in the translation software industry. Combined with machine translation, voice recognition technology can speed up large translation projects in ways we never imagined.
Prepare yourself for the next generation of technology — and convenience. It’s already here.
1 “Google Assistant making a haircut appointment.” Google IO 2018. https://youtu.be/YCWJ0z6_z34
2 “The Past, Present, and Future of Speech Recognition Technology.” Medium. Clark Boyd. January 10, 2018.
Content from the Couch
The on-demand entertainment content market is skyrocketing, so does the demand for audiovisual localization. Translating audiovisual content requires a totally different mindset. For LSPs, it means learning new techniques and technologies. Is the industry ready for it?