Amyotrophic
Lateral Sclerosis (ALS), also known as Motor Neurone Disease, is a progressive
neurodegenerative disease that affects nerve cells in the brain and the
spinal cord. Within a couple of years of diagnosis, most patients end up paralyzed
in a wheelchair and forced to communicate via text-to-speech devices, typically
with a default ‘computer’ voice, pre-recorded sentences or words mechanically
stitched together. Our goal was to change the way people live with ALS by
giving sufferers the full use of their own authentic voices.
Project
Revoice uses a unique deep learning algorithm to analyze the DNA of a person’s
voice and create a complete digital voice clone, from just 2-3 hours of
recorded audio. When integrated with text-to-speech devices, this gives people
with ALS the ability to speak freely and naturally in their own voice, even
after they lose it.
We
launched the program by recreating the voice of Pat Quinn, ALS sufferer and
co-founder of the Ice Bucket Challenge, using old interviews and speeches found
online. After sourcing over 100 individual files of varying quality, we
manually analyzed, cleaned up and transcribed every usable section to create a small
but unified data bank for the algorithm to work with. We then integrated Pat’s
Revoice with his eye-tracking computer and captured the moment when he finally
spoke again.
Since
launch, Project Revoice has had an earned reach of more than 900 million, with
41 million people joining the conversation on social media. Most importantly, over 500 patients joined the program in the first month
alone. Going forward, this program
will change how people live with ALS worldwide.