By Dani Tietz
Picking up a smartphone and putting in a numeric password is something millions of people do every day.
More recently, people have been able to hold their smartphone up to their face only to have it open through face recognition technology.
But artificial intelligence, which is readily available in nearly any device produced today, often has trouble recognizing a human voice command.
“Voice security is different because based on your mood, energy, time of day, whether you’re sick or have a cold, your voice is going to sound different even though it’s still you,” Tim Sinclair said.
Sinclair, who the former WBGL radio host, current PA announcer for the Fighting Illini and Chicago Fire and CEO of Ringr, is currently working on a voice recognition project for a major technology organization, which has to remain nameless.
Since Thanksgiving of 2017, Sinclair has been tasked with collecting a variety of voice samples so that technology organizations, working with companies like Google, Alexa, Bixby and/or Siri, can develop artificial intelligence that distinguishes between variations in voice.
“They are trying to teach voice software that one person can sound different, but still be that person,” he said.
Within two days of reading information about the project, Sinclair was all-in, ordering the microphone and setting up a studio within his home that met the company’s precise specifications.
Sinclair hopes to collect 75 voice samples prior to submitting the voice database on February 23.
Currently, Sinclair is looking for participants who speak English as their first language and are between the ages of 18-60 years old. Each participant is recorded five times for about 20 minutes throughout a 14-20 day period.
As the work is completed, participants receive $50 compensation for their time.
Although the work is tedious on Sinclair’s end, having to spend an hour-and-a-half of editing time per 20-minutes of recorded time, he believes the effort is well worth his time.
“The idea that this is going to make a difference for a big company who is going to impact millions of people, and not just everyday people with their smartphones and tablets, but perhaps for more secure kinds of things in government organizations makes this fun,” he said.
Sinclair said participants will need to begin recording before the first of February at the latest.
To find out more information visit timjsinclair.com/single-post/2017/12/01/Voice-Recognition-Project.
Leave a Comment
Your email address will not be published.