Back to the Future deadlines made possible…almost

As the presence of AI makes itself felt in many areas of our lives, one area that is being heavily impacted is audio-visual translation.

Here Dave Bryant, co-founder of start-up Fluen AI explores how technology is democratising of Audio-Visual Translation

The part of the solution that we hear most about is technology and there is no doubt that it is changing the way things are done. Many of the changes get the ‘disruptive’ tagline, which is shorthand for processes being greatly improved, but, in our field of expertise, market for translations of videos, audio files, podcasts and so on, its more than just improved. The advances we have been able to incorporate has changed the ground rules forever.

Our solution deploys several types of AI. The much-discussed Generative AI, ChatGPT, Natural Language Processing (NLP) and Machine Translation (MT). The integration of all these technologies has resulted in a dramatic shift that allows us to take a video file, caption it in its original language and, because of the capability of AI, the context is created, jargon and proper nouns are handled correctly and captions are timed, segmented and synchronised.

This process is repeated for the translations, where the entire text is presented to the MT engine for translation. Once again retaining context and ensuring that language structure differences are handled correctly. The translation is then run through the same process as the same-language captions, timing, segmenting it naturally and fluently. This is a major, disruptive step in creating captions and subtitles in a way that has not been possible before.

However, as impressive as our technology is, it also allows a huge change in the way people think about making their videos accessible.

All this would have cost an arm and a leg a couple of years ago. For example, if you wanted to caption and translate a 3-minute video into four languages, it would cost around US$180, depending on the languages. This same workflow using our AI-driven technology solution dramatically reduces the cost to around $1.

Nearly one hundred percent of our customers make no changes to the captions or subtitles demonstrating that the quality of the end product is deemed to be excellent.

Of course, every customer not only wants top quality, they also want everything the day before yesterday! While no one in the language industry will ever be able to provide clients with Back to the Future turn-around times, workflows using the technology we have at our disposal have been dramatically speeded up.

Historically, 3-minute video might take at least a day to caption; a day to translate and, perhaps, another to review. This adds up to 2-3 days between submitting the video for translation and getting it back. Using our automated workflow solution, it takes less than 5 minutes.

All of a sudden this opens up huge opportunities for clients for whom audio-visual translation and captioning has been beyond their budgets. Now for companies sitting on large back catalogues of videos, this solution will allow them to optimise existing assets and break into new markets cost-effectively.

An e-Learning site may have hundreds of hours of training videos that have been created, but up until now, it’s not been commercially viable to translate them. The fact that it is now a couple of orders of magnitude cheaper, means that creating a library of videos that are accessible to most of the world is no longer a dream but a reality.

There are so many other examples where this new technology-driven solution is going to prove a game-changer. Perhaps a pop-culture site that needs to get the videos out there for maximum impact, or a sports team that has a huge international following and their coach does a match day interview. Time is of the essence in getting the video translated and posted to websites and social media before the big game.

The democratisation of audio-visual translation means that non-commercial clients – ordinary members of the public – wanting to send video messages with translated captions to family and friends around the world can do so for less than $1 and get it done in 5 minutes.

There are many more market segments where the ability to create high-quality, fast-turnaround captions and translations at a very affordable price will allow the subject matter to be made accessible to audiences where it’s not been possible before, either for those with hearing difficulties or language differences.

To be at the forefront of this democratisation movement makes us very proud – especially when we get emails from our customers explaining how we have improved their business or their personal life by removing the barriers to making their videos and podcasts available to many more people throughout the world.

 

ELIS SURVEY 2021