Future

An announcement was made in August of 2001 of an European company that has developed automatic translation of mobile phone calls to a foreign language.  NPR interviewed the inventor and broadcast a demo.  Understandably, the receiving end must wait for the translation to take place.  Since there was a time delay, the NPR reporter didn't seem all that impressed and asked several sharp-stick questions pertaining to the time delay involved and even suggested that there might be problems with the service since there was a delay.  The inventor mentioned that the service was 1st generation and that the task is enormous.  He even pointed out that the service was able to decipher and translate correctly a phrase where the original speaker corrected himself (Ed: To me that is real time proofreading).

The time delay will be solved by increasing processing capabilities of the future.  To me the service as offered today wasn't as significant as the technology being applied.  That technology is made up of using the individual building blocks already made, speech recognition, speech to text, language translation, and then text to speech.  All performed in an impressively short period of time, albeit a recognizable delay for a conversation.  What is notable is that there was also a component of proofreading and automated correction added to the process.  To me, that was the most impressive component.

Language recognition was a manual process in the step but eventually will be a tweak of the speech recognition process.

When the NPR interview ended, the table was turned over to another reporter for the next story.  Small chat often blends the two stories.  The new reporter asked the interviewer if he would buy one, meaning not due to any price but if the service was meaningful enough, and the response was "No, I don't think so."  The comment was made as if the invention was not quite there yet (you could almost detect the smile between reporters).  Apparently, we have been spoiled by Sci-Fi.  I say the reporters being pessimists is due to the Star Trek Universal Translator model we all desire, where speech is spoken in whatever tongue but the recipient hears it in their own familiar language.  The only other explanation I can think of is that NPR News Service is worldwide and often translators are used.  The production team behind the scenes makes arrangements for everything to fall into place for the broadcast.  Maybe when you have constant access to knowledgeable translators, you see a mobile phone translation service with a time delay as a trivial diversion.  NPR, you missed the boat.  The rest of the world does not have translators standing by.  We still rely on ancient technologies such as the Internet and text translation utilities.

What are the alternatives?  Consider the amount of time it takes for translating text that must first be typed in as in a chat session.  Actual speech is much easier, and faster. Even a long delay would be faster than writing.

Imagine the mobile phone service evolves to encompass all languages.  This service would allow worldwide problems to be resolved easily without need to locate a human translator.  The service falls exactly in line with what could be expected as we make our way to the Star Trek model.  The same service will move to the Internet when it is used as our main path of voice conversations.  Soon the Internet will carry our conversations and communications complete with near instant translations, synched to our lips for the video viewer on the other end.  The proofreading will be performed real-time.  An archived (no-proofreading required) written version will be provided by the same software translation service.

I can imagine the next step would be to loop our voice output into a local version of the service, feed its corrected output back out a speaker and therefore have the ability to speak flawlessly.  Of course, typing wouldn't be necessary because an archive would be provided.  If we could tap into the brain part that translates the thought we hear into input for the local device, no one would need to hear the original.  Nerve triggers could wiggle the lips to match the output timing.  Perhaps distracting initially but I'm sure we could get used to it, both the weird sensation and learning to ignore it on others.

And when you have the local version, flawless output could be in any language.

Proofreading could be automatic.  An end-user could speak flawlessly.  Foul language could be removed automatically so that there are no slip-ups when visiting family or the elderly.  Of course, the reverse could also be true, the speech could have articulated adjustments for the situation.  Our spies could mingle inconspicuously in the ghetto or in stately dinner parties.

I'm also thinking of the advantage of being able to sing the most beautiful French love song in order to woe the dream girl I encounter while vacationing on the Riviera.

We're getting there.

I was just thinking.  My only concern is that if she is keeping up with the times too, how will I be able to know from her voice if she really is French, or even a girl?  And even if I could tell that she is a girl, say with cosmetic surgery and the like, what would the kids look like?  There are no guarantees.  I'm not so sure this future is looking so bright.

The age of global Internet connections is upon us...

Let us help you eliminate errors on your website or in your publications.