Here’s How Voice Assisted AI Technology Can Give People A Voice Again

clearly with friends.

Whispp

Johns Hopkins defines stuttering as a voice speech disorder. Stuttering affects more than 80 million people worldwide, and in the United States, more than one million Americans stutter.

A voice disorder is a problem with pitch, volume, tone, and other qualities of your voice that occurs when vocal cords don’t vibrate normally. There are several types of voice disorders classified as organic, which include structural and neurological (caused by a neurological disorder like Parkinson’s or Alzheimer’s), functional (muscle dysfunction), and psychogenic disorders.

A Dutch start-up has created an app designed to give a voice to people with voice disorders or speech disorders, such as stuttering.

Using artificial intelligence (AI), the Whispp app enables them to make understandable and relaxed phone and video calls.

“The app has a real-time assistive voice technology that converts voiceless/vocal cord-impaired speech or whispered speech (speech that does not have a clear pitch) into natural and voiced speech,” said Joris Castermans, CEO of Whispp. “People who stutter severely, for example, can reduce their stuttering frequency by an average of 85% while whispering. Additionally, people who suffer from spasmodic dysphonia or recurrent respiratory papillomatosis speak much more relaxed and fluently when they whisper.”

In an email interview, Castermans said the Whispp enables users to express themselves better and easier, enhancing their quality of life and allowing them to participate more fully in society.

“Communication is a fundamental aspect of human existence that presents a daily challenge for people who suffer from a voice disability or stutter severely,” said Castermans. “The inability to communicate can lead to social isolation and, in many cases, feelings of inadequacy and depression.”

With their own AI models for Whispp, the AI is audio-to-audio based with no textual intermediate, so the company doesn’t use language models.

“With this, Whisper converts non-voiced speech with a very low latency,” said Castermans. “Whispp’s AI converts every 20 milliseconds of audio into a real-time stream.

Whispp uses real-time, audio-to-audio-based assistive voice AI to create real-time speech conversion and accommodates a range of voice types — from whispers to rough esophageal speech. This allows the app to create a tailored solution for several voice conditions.

“Communication is a fundamental aspect of human existence that presents a daily challenge for people who suffer from a voice disability or stutter severely,” said Castermans. “The inability to communicate can lead to social isolation and, in many cases, feelings of inadequacy and depression.”

For example, Castermans says people who stutter severely speak fluently and are relaxed when they whisper. “This is because of a neurological change that occurs while they are speaking; aside from this, people who stutter severely didn’t ‘learn’ to be anxious while whispering.”

Castermans says that big tech and assistive speech tech companies predominantly focus on Automatic Speech Recognition, known as speech-to-text (STT), for non-standard speech. “This is very helpful for patients with reduced articulation (ALS, MS, stroke and Parkinson’s Disease) who can use text-to-speech to synthesize their speech.”

“The disadvantage of this approach, however, is the high latency of two to three seconds, which creates barriers to natural conversation,” said Castermans. “As a result, current AI speech technology solutions do not provide an adequate solution for people with voice disorders who have lost their voice but still have good articulation.”

The Whispp app is available on Android and IoS.

Follow me on Twitter.