Google has been trying for quite some time to push voice based programs as a viable means of interaction with electronics, and this newest development seems to have finally found a way to do so.
What we are referring to is the new voice synthesis program called WaveNet, which uses a deep neural AI developed by Google in order to create a very human like voice.
While this may not feel like such a big deal, it still is quite valuable for various interactive applications that suffer at the hands of the stilted robotic voice that past technology had forced us to live with.
How big an issue can an impersonal robotic voice be, can only be understood if you’ve spend a lot of time with voice based gadgets like the Pilot Language Translator.
Thankfully you won’t have to worry about this any longer, since this new system actually understands the nuances of human language and can generate it’s own audio samples.
This is a big achievement, as all the programs up until now could only create audio samples that were stitched together from existing fragments of sound.
In order to make this very important distinction possible, the system goes through atleast 16,000 audio samples per second, so it can actually understand what makes human language so unique, in both spoken and unspoken terms.
While it will still take a long time, at this speed, to actually achieve constant language fidelity, the fact that we have already seen so much progress in such a short time period, makes us believe that the future is not so far away.