Google has unveiled Lyra, a new technology that allows Duo, and soon other apps, to offer natural-sounding voice chat with as little as 3 kilobytes per second of network bandwidth.
Update: Google has now made Lyra open source for other apps to use.
With so many of us still unable to visit with loved ones, video calling apps have played a key role in keeping us all connected to one another, with Google Duo and Meet hosting over 1 trillion minutes of video last year alone. However, this has also put a major strain on internet infrastructure around the world, and most video calling methods outright exclude anyone with a low-grade internet connection.
To help with this, Google has developed a new audio codec called Lyra that is specifically optimized to offer recognizable, understandable, and natural-sounding human speech in as small of a space as possible. As explained on the Google AI Blog, this was done with the assistance of a machine learning model, which was trained on “thousands of hours of audio with speakers in over 70 languages” to ensure Lyra could be usable by as many people as possible. More importantly, Lyra is efficient enough to be able to run on anything from a high-end cloud server to a mid-range smartphone with only 90ms of latency.
As can be seen — or rather heard — from the video above, Lyra offers audio that, while noticeably lower quality than a normally encoded recording, is distinctly recognizable as the speaker’s voice while managing to use incredibly little data. There are some more samples of speech in Lyra compared to other low-bandwidth audio codecs, over on the Google AI Blog.
Lyra will soon be getting its first real-world usage as it is rolling out now to Google Duo for Android, where the codec will be used for calls made on low-speed connections, with Google making particular note of dialup connections and rural areas in India and Brazil only offering a 2G network connection. From there, Google plans to release Lyra as open source, allowing other companies to offer low-bandwidth Lyra audio in their own apps.
Update 4/6: Just over a month after being first announced, Google has kept its promise and made Lyra fully open source. For now, this first beta release of the Lyra codec is only optimized for use by Android developers on Linux machines, but should be enough to give developers something to start working with, and later bring to all of their intended platforms.
We are releasing Lyra as a beta version today because we wanted to enable developers and get feedback as soon as possible. As a result, we expect the API and bitstream to change as it is developed. All of the code for running Lyra is open sourced under the Apache license, except for a math kernel, for which a shared library is provided until we can implement a fully open solution over more platforms. We look forward to seeing what people do with Lyra now that it is open sourced. Check out the code and demo on GitHub, let us know what you think, and how you plan to use it!
More on Google Duo:
FTC: We use income earning auto affiliate links. More.
Check out 9to5Google on YouTube for more news: