The Times recently reported that Google is looking to progress in its voice technology for cell phones. The company is developing a software that will be able to translate almost instantly on cell phones. This technology would incorporate Google’s ASR (Automatic Speech Recognition) to convert the spoken words into text, translate it, and then reconvert that translated text into a computerized voice to transmit the message in a second language.
According to its director of translation services, Franz Onch, Google currently has the three services working separately and, therefore, the goal right now is to better combine them, which requires an automatic translation system that is extremely accurate as well as a voice recognition system that is its equal in efficiency and performance.
Nevertheless, Och emphasizes that, even though significant progress has been made in the performance of Google’s translation and voice recognition tools, each person has a unique tone and accent and that must be a focus. In fact, the idea would be that the system “learn” to recognize the user’s voice with greater ease as he or she uses the voice recognition tools, such as voice search.
Given the impressive advances in these technologies, Google is very optimistic and believes that voice translation could work reasonably well. As such, they are considering launching this new technology in a few years. Also, Google has a clear advantage over its competitors in that its online services give access to a large number of translation services.
What is true is that if what I describe above is possible, this would be a true revolution throughout the world. Being able to speak with a person in your own language and for them to have the same ability, with a fluid communication in the middle, is amazing and, indeed, I have my doubts about it.
Skepticism on this subject is apparent. We can’t forget that language is alive and always changing, in addition to the myriad accents or pronunciations, possible linguistic choices that are not “official”, irony, etc. Interpretation is more than a simple automated action. It’s not limited to providing linguistic equivalents, but “interpreting” exactly what different people, at times with different mother tongues, mean; needless to say, the task is not an easy one.
I simply cannot be as optimistic as Google and think that in the coming years they will release a technology that replaces a profession that is so specialized and complex as interpretation.