New ‘Sound Bubble’ Tech Lets You Hear Conversations Clearer Even in Crowded Spaces

It’s like having a superpower to mute the world but hear what matters.

The frustration of straining to hear conversations in noisy spaces might soon be over. Researchers at the University of Washington have created headphones that build a “sound bubble” around users.

This technology amplifies voices within a 3-6 foot radius around you while dropping outside noise by up to 49 decibels. So, conversations in restaurants, offices, and airplanes can soon be clearer and more bearable.

How the ‘Sound Bubble’ Works

The ‘sound bubble’ headphones use six microphones spread across the headband and earcups on regular noise-canceling headphones. These microphones feed sounds to a small computer attached to the headphones, where the AI software processes what you hear.

Inside this AI software, there are systems in place to accurately track sound locations using different methods.

First, it studies how different sound frequencies reach each microphone. Then, by measuring tiny time differences, it calculates their distance.

Next, it analyzes how the sound bounces off your head.

As it turns out, your head reflects incoming sounds and creates subtle differences in how each sound reaches the microphones. Such reflections create patterns that help the AI accurately tell if a noise is near or far.

The whole process is lightning-fast as it takes just around 8 milliseconds (which is faster than you can blink). And, when it detects a sound outside your bubble, it drops its volume dramatically.

For example, it can take the sound of a vacuum cleaner down to the soft rustle of leaves. Meanwhile, voices inside your bubble stay clear and easy to hear.

Applications and Comparison to Existing Solutions

Traditional noise-canceling headphones aren’t great for socializing. They either block everything out or let in too much noise.

In contrast, these new AI headphones strike a better balance by keeping conversations crystal clear while reducing the chaos around you.

Aside from that, the tech works differently from what’s out there now.

For example, it improves upon previous innovations like the “Target Speech Hearing” (TSH) system from the University of Washington.

TSH could pick out one person’s voice by following your gaze. But, it could only focus on one speaker and got confused when other nearby conversations were just as loud.

This new system is more practical. It can handle multiple speakers within a specific radius at once and can filter out background noise even when they’re louder than the speakers.

What’s really cool is how well it works in different places.

Thanks to all the real-world sound recordings used to train the AI, these headphones adapt smoothly whether you’re in a quiet office or a packed café. And they work just as well for someone trying them for the first time as they do for the people who helped develop them.

Challenges and Potential

Like any new technology, some challenges remain.

For one, the system works best indoors where sound behaves more predictably right now.

Getting it to work outdoors is trickier because there are so many different kinds of background noise to deal with. But, the team is tackling this by recording more sound samples and testing in places like stadiums, which are kind of a middle ground between indoor and outdoor spaces.

Making the bubble size just right is another challenge they’re working on.

A bigger bubble lets you hear more people around you, but it makes it harder for the system to tell exactly where sounds are coming from.

The solution? They’re collecting more detailed sound samples from the edges of these bubbles to help the AI make better decisions.

Moreover, the current version runs for about four hours on a 5,000 mAh battery. That’s enough for a long lunch meeting or your daily commute but not quite ready for all-day use.

The team is also looking at ways to shrink the technology to fit into hearing aids and earbuds. However, effective microphone placement and maintaining computational efficiency in smaller devices require significant design adjustments.

Despite these challenges, the potential is huge. In fact, the team is already working on bringing the system to market through a startup. And, by using existing hardware, like off-the-shelf noise-canceling headphones, they hope to make the technology widely available.