We have already spoken a lot in other posts on the progress of technology in the field of translation: translation memories, automatic translators, tools for quality control, the list goes on. There are even software programs that allow us to enter text by speech recognition: just install a small program or plug-in to have them then work with a translation tool (Trados, SDLX, etc.). Thus, we can add text orally and avoid cumbersome process of spending hours and hours behind a machine typing thousands of words.
But technological advances never stop. There is always time (or creativity) for something else. Google, for example, is an example of this.
With the release of Nexus One in January 2010, Google let the world know once again that they are never satisfied with what they believe and are always looking to bring something beyond the limits of what the company is able to achieve. This device (although old, now a year after its launch) offers a voice translator that allows users to have discussions with any individual regardless of the languages spoken.
We have seen in the past the big G develop systems that do things with your voice instead of text, like the voice search. Using technologies such as those named above, Google expects to have the basic system in a few years. The operation is very complex, but the idea is to have the ability to recognize someone’s language and translate its synthetic equivalent in a different language. The system will analyze packets of dialogue, listening to the user the message you want to send is understood, then translating it.
According to experts at Google, it is clear that in order to work properly, it requires a combination of accurate translation and a voice recognition system that is perfect as the company continues working on it to this day. But also supporting voice recognition is not easy. Each person has a different voice, accent and tone, but recognition should be effective with mobile phones, which by nature are personal to each. The telephone, for example, should be able to recognize your particular voice.
The only problem I found is that the system should work seamlessly to function. While the text translators could have the same user problems and fix them, here the listener receives the translation of the system and nothing of what the other person actually said. That is why we cannot believe that we will be able to seamlessly translate what was said by one person. Even so, it must be recognized, Google tends to surprise us. Anyway, it is always going to be much more productive to place our trust in services offered by humans, at least until the technology is tested over a longer period of time. One example of this is the Phone Interpretation service offered by Trusted Translations.