Project
Communication powered by research
Euphonia
Project Euphonia
Communication powered by research
Project Euphonia is a Google Research initiative working to make speech recognition software more accessible.
Conversations are a daily occurrence, but talking can be a major challenge for millions of people around the world with atypical speech.
How it works
Record
Volunteers will record a set of 300 simple phrases to provide a variety of reference material.
Refine
Scientists analyze this audio data and use it to refine speech recognition software so it can effectively understand as many voices as possible.
Results
This research is integrated into apps and initiatives that provide assistance for people with non-standard speech.
We use the information you share to make Google's speech features easier to use for everyone, now and in the future.
Euphonia in action
Our work with Tim Shaw, a former NFL player living with ALS
How a former NFL player is using AI to get his voice back
Customized speech recognition for people with non-standard speech
“So very grateful for this project, and I am hopeful this will help many who struggle to communicate their thoughts to the world.”
“I’m with you for the long haul. I want to make this work. I used Google Assistant to gain my independence…This program means so much to me and other people like me.”
“It gave me motivation to help out with the research so that I can help not only myself but also a larger group of people with communication challenges that are often left out.”
Speak up for accessible speech
Before you get started
Am I eligible to take part in Project Euphonia?
Am I eligible to take part in Project Euphonia?
The following requirements are necessary to participate in this Google Research initiative:
-
You are at least 18 years of age
-
You have difficulty being understood by others due to your non-standard speech (not because of your accent or vernacular)
-
You have your own Google account
-
You have access to a mobile or desktop device (Android or iOS) with a microphone
-
You can speak and read one of our available language options: English, Spanish
-
You have the mental capacity to give consent
What will I do as a participant?
What will I do as a participant?
You will read 300 phrases aloud as you normally would using the provided recording tool online. These recordings are then studied and used to train speech recognition software.
What will I need to get started?
What will I need to get started?
You’ll need access to a phone, tablet, laptop, or desktop computer with a microphone. It does not matter if the mic is built-in or plugged-in. Just click on the blue “Record phrases” button to get started.
What do I do if I’m having trouble?
What do I do if I’m having trouble?
Try opening the recording tool in a private browser. You can right-click in Chrome and select “Open in an Incognito Window.” If you’re still having difficulties, contact us.
What will I get in exchange for participating?
What will I get in exchange for participating?
Participation is entirely voluntary, but we understand the time and energy it takes to record these phrases. We provide cash-equivalent gift cards as a thank you to all participants who complete an entire set of 300 phrases, once they have been reviewed and approved for Euphonia research by a team of speech and language specialists.
The gift card will be valued at $60 USD or currency equivalent. Please keep in mind that standard speech and accented speech samples are not considered suitable for Project Euphonia.
When will I have access to technology that can better understand my voice?
When will I have access to technology that can better understand my voice?
Project Euphonia is still in its early stages, but our goal is to improve the quality of everyday life with this research. That goal may take years to accomplish, but every audio sample we can analyze gets us closer to helping those with non-standard speech communicate.
Recording tips
Read every word. The recording tasks will have you read specific phrases. Please do your best to speak each and every word aloud while you record. If a particular phrase is too challenging, just skip it.
Limit the noise. It's important that your voice, and only your voice, is heard clearly. Find a quiet environment to record in, without other voices or devices playing in the background.
Just try again. You will always have the option to re-record a message as many times as you like. Play back your phrases to check that there aren’t any missing words, muffled microphones, or other voices in your recording.
Be yourself. Don’t aim for perfection or worry about how you sound. We need real, everyday voices. It’s important to capture natural examples of what our software struggles to understand so that we can improve it to be as accurate as possible.
Default to mobile. Please record on an Android device (i.e. a Pixel or Samsung phone) if possible, but you can still record from any mobile device, tablet, laptop, or desktop.
Track your phrases. The phrases you finish will get moved to the back of the queue when you log in. The progress bar will show how much is left, and you’ll notice anything you’ve already recorded will be highlighted in blue.
In partnership with
Join the conversation
Contribute your own voice to help scientists refine the algorithms that understand what we say.