NTT Docomo is developing a new translation system that provides real-time automatic interpretation. The Japanese mobile operator demonstrated the technology at Wireless Japan 2011. It works using a range of existing cloud-based services that recognize words, translate them and synthesize speech in the translated language. This provides a quicker conversation, with no awkward pausing while you wait ages for your words to be understood. However, as a result of the gain in speed, it has lost a little in the accuracy of the translation.
“Voice recognition is not at 100% yet. In order for customers to use it we have to improve recognition accuracy a little more, and we also need to improve translation accuracy a little more as well. However, we really don’t know how much time it will take to bring accuracy to 100%. Conversely, we are considering scenarios in which customers who could accept a certain level of inaccuracy could use this. Possibly even at this level of accuracy there are ways to use this so that some customers would be able to appreciate its convenience, so we plan to recruit people to monitor the service and offer it on a trial basis this year.”