Google announced its new Lyra audio codec in February that it will use in Google Duo. The codec enables higher quality audio at lower bandwidths, which means it is ideal for slower network speeds. On Friday, Google announced that it was expanding this new codec to include third-party developers, opening up a world of possibilities for Lyra.
As part of our efforts to make the best codecs universally available, we are offering open-sourced Lyra so that other developers can power their communication apps and take Lyra in new directions.
As a refresher, Lyra offers an encoder to capture audio, which it compresses and then decodes using a generative model, a type of machine learning model. The model enables the new generation of audio with particularly low bit rates of only 3 kbit / s with consistently high output quality. For a more in-depth explanation, check out Google’s AI blog, but Lyra is essentially ideal for use on the best cheap Android phones or slow networks. Google underscores the importance of technologies like Lyra, especially in emerging markets where fast internet speeds don’t exist in an age where teleworking has become the norm.
If you make Lyra open source, the codec can be used in more communication applications. However, Google hopes developers will find new ways to use the codec. This could include applications for non-voice applications or even music, which would be ideal for platforms like Spotify that are continually expanding into more global markets, but the possibilities are endless.
According to Google, Lyra is written in C ++ “for speed, efficiency and interoperability” and today’s beta version and demo on GitHub allows developers to provide quick feedback.