The previously released ML-based noise suppression is now enabled by default for most of our customers. ML-based noise suppression considers any non-speech signal picked up by the microphone as noise which
should be suppressed. To avoid suppressing music and to allow users to enable the new high-fidelity music mode, we have built an ML-based music detector which will inform the user whenever music is present through a notification. This will give the end user the choice whether music is indeed an
unwanted background noise, such as when calling into a meeting from a coffee shop, or if music is a desired signal, such as when participating in a music lesson.