Tired of bad internet and disrupted calls? Relax! Google has a solution for you. Yes, you read it right Google has a solution for your problem. “Google Lyra” is their new product.
As we know, the past year has been difficult for all of us. The pandemic has made it clear just how important an online communication system is for our life considering the lockdowns and social distancing measures. That’s why Google decided to accelerate digital communication and launched Lyra in February 2021.
Lyra is a revolutionary audio code that makes use of a machine to allow high-quality voice calls. It has aided in the mitigation of network access issues. However, in some areas infrastructure must be modified to give access to a large number of users.
Google never misses a chance to impress us through its innovations. That’s for sure. Google has made the product open supply, inviting different builders to function their communication apps and equip Lyra with highly effective new capabilities. Using Lyra, developers can easily encode and decode audio because the software runs on an android platform of 64-bit ARM together with growth on Linux. The growth of telework attributable to COVID-19 has additionally elevated the necessity for reliable digital platforms.
That’s the reason Lyra intends to increase accessibility for its customers. Lyra can also save battery by leveraging the computationally cheap Lyra encoder for alleviating the congestion of the network in emergency situations when several people are making calls at the same time. Also, Lyra can successfully archive large amounts of speech.
How will it work?
The Lyra structure consists of two items, an encoder, and a decoder. When you speak on the telephone, the encoder registers special aspects from your speech. These speech points are divided into chunks of 40 MS, processed, and sent over a network after compression. The developer then has the responsibility to revert the features to an audio waveform again which can be played out over the listener’s phone speaker. The features are encoded back into the waveform using a machine learning general model.
