The translation market: Speech to text

The ability to translate not only words but their context is a highly skilled and stressful job. AV technology is helping interpreters do their job in new ways but the threat of AI looms large over the industry.

“As the delegate spoke, (translator Marisa) Pinkney had to make sense of a message composed in one language while simultaneously constructing and articulating the same message in another tongue. The process required an extraordinary blend of sensory, motor and cognitive skills, all of which had to operate in unison. She did so continuously and in real time, without asking the speaker to slow down or clarify anything. She didn’t stammer or pause. Executing it required versatility and nuance beyond the reach of the most powerful computers. It is a wonder that her brain, indeed any human brain, can do it at all.”

That quote is from a BBC article entitled ‘The amazing brains of the real-time interpreters’ and sums up perfectly the fraught job of an interpreter. If you throw in brand new technology you’ve never used before or is tricky to master, then users are faced with a highly unwelcome added degree of difficulty.  And that is the dilemma facing those tasked with providing systems for translation and interpretation, you’ve may be innovating for end users who don’t want change.

televic confidea flex

When we think of the translation/interpretation market, we automatically picture giant rooms at the UN or NATO with rows and rows of seats or booths, however there is also a rental market at work here for one-off events, and demand for translation services in boardrooms in small-to-medium businesses is growing too. Unsurprisingly, given the Covid-19 global pandemic, the biggest trend in the interpretation market is the ability to translate remotely, but alongside this, what else are users looking for made for this sector? “We have seen the growing importance of inclusivity,” says Annabelle Zabetian, co-founder of the Kudo multilingual conferencing and translation platform.  “The importance of being able to deliver live messages into the broadest audience for actions to be happening at the same time for all.  What we've been seeing right now is anyone from corporate leaders to city officials want to be able to advise their audiences live and without a gap in the action for all the various parties, communicating at the same speed with all audiences.”

Much like the rest of the AV world, the growing rise of AV over IP systems is having an impact here. “Things are moving to IP and integration is becoming a more and more important topic, in the larger environments we hear more and more requests for integrated solutions,” says Lars van den Heuvel, director global product management, Bosch Security Systems. The worldwide lockdown caused by Covid-19 has forced the translation market (like all markets) to seek ways to work remotely, but was this demand there previously? “Before there was a very slow growth, it might have been fear that the technologies would not be up to the challenge, or the fear of adopting new technologies. But in a world where you just physically can't travel, you have to adapt a lot quicker to the tools that are available now,” says Zabetian.

Bosch DICENTIS iDesk

There was some demand for remote systems before Covid-19 says van den Heuvel, but now it might change how translation is performed in the future; “If you do Spanish interpretation then you can imagine that someone in Argentina can do that, so it can drive costs down, so let’s see how the interpreter organisations react to this trend.” Televic is one of the biggest names in this sector and has seen a big recent push into more remote applications for interpretation. “There are two ways of dealing with remote, either you have interpreters working from home, or there is a specific hub where interpreters go to,” says Didier Rosez, product manager.

When you think of where the big translation and interpretation venues are likely to be – New York, Geneva, Brussels - i.e. very built-up cities with high rental costs and low availability, it’s not surprising that remote hubs become attractive. “If you look to Brussels for example, they had a requirement in the past for 32 languages which has now increased to 40 languages,” says van den Heuvel.  “If you look at recent tenders it's not easy to expand an existing room with eight traditional booths.”

When real estate is at a costly premium AV tech can really help says Karel Vanheule, team leader, product management, Televic; “We are seeing a lot of interpreters being grouped together in large institutions in a room in the same building, and centralised in one big area, and that interpretation can be assigned to different rooms at the same time so that they save on space. The technology really important to them is having good video distribution available so they can feel close to the meeting room.  Remote can be cloud-based so you don't have the interpreter in the room itself but centralised in the same building, and in that sense the technology that’s important is good audio and video distribution towards the interpreter who doesn't have the clear view on the speaker (in the room) anymore but needs to have a very good view of the speaker on the screen.”

televic confidea flex
Sennheiser is also looking at helping end users when the space to build new booths just isn’t possible, “Our tourguide systems are also used for something called ‘whisper’ translation. The interpreter sits in the same room, no booth and whispers the translation into a wireless mic. The recipient(s) wear a receiver and listen to the translation e.g. via in-ears,” says Stephanie Schmidt, communications manager, professional audio, Sennheiser.

The way technology has managed large meetings needing translation services hasn’t really changed much in the last 10 years. They work in large, highly organised teams using established technology from the same 4 or 5 vendors. The only real change is the one we have covered above, the growth of remote hubs. But what about smaller meetings, can this technology offer those meetings interpretation services or are the logistics involved/budget required just too far out of reach?  “For smaller meetings the issue is the budget.  There is a certain budget associated with being properly equipped with hardware systems for interpreting.  Being able to have smaller meetings means these services need to be made more accessible at a financial level too.  With online technology this is possible as there is a lot more flexibility in the number of hours from booking interpreting services, because interpreters don't need to come to the venue, so can you save on (the time and cost of) transportation. Online meetings are a lot shorter generally speaking, so you need flexibility for one-hour solution,” says Kudo’s Zabetian.

The opportunity to sell your products beyond giant rollouts to the UN and NATO etc has not escaped manufacturers, as this from Bosch’s van den Heuvel illustrates; “When we developed our Dicentis conference system, we switched to a fully IP-based system so all of our components for delegate units and the interpreter desk are connected directly on an IP network, which gives quite a good foundation for the number of languages that we can provide, but also if we look to smaller meetings we see an advantage there that is much easier to integrate with those types of data conferencing applications (such as Zoom, Teams etc).”

televic confidea flex

One interesting aspect and one that’s unique to this sector is its approach to touchscreen technology, at a time when every product has a touchscreen, this industry is rejecting them. “One thing we saw was interpreters really don't like them,” says Rosez.  “They want to concentrate and maybe adjust some things while doing the interpretation without even looking at the device.” The industry is aware of this however, and is adapting in clever ways says Zabetian; “Even online or remote interpreting solutions are providing a certain level of hardware accessory that if interpreters would feel more comfortable touching hard buttons they can have a volume control knob that's an USB accessory to their computer interface, online solutions have to adapt to those preferences.”

Because end users have such specific requirements in this sector no product development in this sector is done without involving them in the process, but as already mentioned, don’t expect a radical overhaul of what is already available. “When we started developing our new interpreter station we did a lot of work with interpreters on the institutional side and on the freelance side, and we had a lot of new concepts. But after all those discussions what we actually saw is that interpreters are looking for evolution not revolution, you have to understand that interpreting is cognitively a very stressful job. And everything that is on top of that is basically overload,” says Rosez.

AI is firmly on the horizon of all interpreters, whether they want it to be or not. With the growth of machine learning, are we nearing a time when AI will entirely replace interpreters? The answer is a firm no, not for at least a decade. “I think AI will help them in the future, but not between now and 10 years, especially when you look at the governmental meetings, I can imagine that this will happen faster in the corporate market because they have less sensitivity with their interpretations. If you do something wrong it can easily be misunderstood, in political arena you want to avoid that. I also think interpretation is not just making the right translation, there's a lot of emotion. If a person is speaking with anger in their voice, or saying something in a light-hearted way, they take that into their interpretation as well.  I think that's a very important aspect which cannot easily be replicated by AI at the moment,” says Bosch’s van den Heuvel.

The message from the industry is clear, AI doesn’t have to mean the end of translation and interpreting; “In terms of the translation market it's pretty clear that in the past 10-12 years translation has greatly evolved due to AI,” says Zabetian. “But it doesn't really mean that there has been less work for language service providers and professional translators, they are still thriving, and the market is expanding.”  In fact, for those smaller meetings, AI can help not hinder says Jakub Kolacz, manager product and commercialisation, Sennheiser Streaming Technologies, “AI delivers the chance to add translation into many situations where a human translator could not be afforded. Not all situations require a human grade translation.”

It may have another benefit too says Televic’s Dezyn, “One of the most difficult things for an interpreter is working with numbers, understanding numbers and replicating that, cognitively it seems to be a different link in your brains and that's one of the things that they would like to see.” Because of the sensitive nature of the content, and the highly skilled nature of interpreters, it would take a quantum leap in AI to replace what is in place now. Until than happens, and it could be 20 years or 30 years before it does, end users will rely on AV tech to get the job done.


Most Viewed