Smart Technologies, Smart Translations – Languages & The Media 2014

The  theme of this years 10th Language & The Media conference was Smart Technologies, Smart Translations with a panel and audience discussion on:

1) Machine Translation (MT)

Sketch of a robot reading a book and writing a translation

Machine Translation (MT)

I thought it was interesting to find out that two Spanish universities have rolled out auto machine translated subtitled recordings of lectures. (I assume with post-editing also). There are numerous companies, and academic institutions working in and researching in this area including SUMAT, EU Bridge, transLectures.

2) Speech Recognition

Sketch of someone using speech recognition with really bad recognition output

Speech Recognition (SOURCE: toothpastefordinner.com)

Speech-to-text technology is already playing an important role in producing real-time or live subtitling and captioning. However there was also some interesting discussion about text-to-speech technologies and access to audio-visual media via audio description.  Particularly with regards to speech synthesis. Many expressed a concern that the quality of the machine voice is not, or would not be a pleasant experience for the end user. However  it was also pointed out that this would not replace current human based audio description but allow for bigger volumes of written content such as news to be made available to audiences that currently have no access such as illiterate audiences.

3) Cloud Technologies

A graphic illustrating benefits of cloud technology: infrastructure, platform as a service, software as a service

The benefits of cloud technology.

There is a lot of talk about this everywhere at the moment. Within the context of subtitling and translation workflows I can see a lot of benefits. It is already being done and where as previously (or currently) much of the business process management has been based on trust, a cloud based system can allow for greater transparency and even force checks. For example a translation manager or coordinator might currently communicate with a freelance translator via email but the actual work is done by someone else on their computer in a different location. The status of a project or job would be unknown and the co-ordinator has to trust that it is being worked on and will meet the deadline set. With a cloud-based application – a translation coordinator (or even the client themselves) could potentially log into the application and see the progress or status of a job being completed by anyone anywhere in the world. It also makes it easy for multiple access for multiple translators on a single project if required. And depending on how you build the application it could also force a task or process (for example only allowing a final subtitle file to be sent after a QC check has been made).

Advertisements