Hello, many thanks for noticing… the related release note is:
- 3rd generation sound recognition neural networks
- New classes: laughter, cough and sneeze, baby crying
- Better classification on more devices with less false positive
- Detailed sound classification settings in Settings > Noise recording > Sound recognition
No you can go the the settings and fine tune sensitivities and thus also minimize false positives for each class…
We have that the new categories can give some useful additional insight. The sneeze and caught can indicate onset of a flue or similar so you know to take vitamins in the morning for instance, baby crying can a big influence on sleep quality of course - this is than also used for awake detection, laughter is a more a fun class some people tent to laugh in their sleep so now you can listen to it back in the morning when this happens…
Looking forward on your feedback on the new sound recognition…I our tests this new networks should perform significantly better on different devices. Some issues n problematic devices with lots of false positives should now be resolved.