The Human Voice Trumps All, at least for the foreseeable future, according to a recent article I read in ‘Wired‘.  And no, the title of this post has nothing to do with a certain political candidate who is running for President.

The article was about brain researchers who have repeatedly attempted to perfect brain to computer interfaces by using various electrode implants and prosthetics.

Researchers at Brown University initially had some success with subjects that could control a robotic arm with the mind using neural brain implants.

As fascinating as the progress was, one researcher named Kennedy was more intent on pursuing the more challenging concept of brain to computer speech cognition.

 

The Complexity of Human Speech

Kennedy knew that human  speech is vastly more complex than robotic arms. It involves the orchestrations and contractions and releases of more than 100 different muscles, from the diaphragm, face, throat, vocal folds, tongue and lips to be intelligible. And to fabricate a minimally invasive and compact brain implant electrode that could accurately orchestrate the complexity of vocal language  with only a few wires was a huge hurdle.

Kennedy’s research project  eventually ran out of money and the FDA revoked approval of the use of implants.

 

Synthetic Voices

There has been some progress on brain to speech. Algorithms have been created for prostheses which assist patients with ALS and others who can no longer speak. But the outcomes are not always 100% accurate.

 

Voice Recognition , TTS and Siri

Speech recognition software has been around for decades. Kennedy says it helps superficially but isn’t very useful in the real world. “Siri’s not good enough “, he says.

I’ve heard script to speech synthesizer recordings and agree with Kennedy. We’re a long ways away from perfecting artificial intelligence.

 

So, what does it all mean to the voiceover world?

Considering the complexity of human emotions, passion, expressiveness and the myriad of ways one can read a single line, it may take a lot more time to even come close to human vocalization that any app will offer.