The company says the new method can be used for real-time audio streaming, which will help significantly improve speech quality for low-bandwidth connections, such as phone calls in regions with unstable communication. The method works for music too, so one day we’ll get some really small music audio files with it.

what is known

EnCodec It can compress audio at 64 kilobits per second 10 times faster than mp3 without loss of quality. Therefore, the file will occupy 10 times less space on the hard disk with the same sound characteristics. The developers say that the neural network does its job very well and the human ear cannot detect the difference between the original and the encoded source.

It looks like the company will use the new technology in its meta version. In an October 25 report, the researchers highlighted the benefits of EnCodec for this large-scale project, saying the enhancement could provide “a rich metaverse experience without requiring massive bandwidth improvements.”

Currently, the new Meta technology is under research, but it points to good prospects that high-quality audio will use less bandwidth; this would be great news for ISPs whose networks are overloaded with streaming media.

The exact dates of the codec distribution are unknown.