You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Fact is some users prefer to use OpenWakeWord instead for now because creating your own MicroWakeWord model that works good is still hard.
End-user should have the option to use on-device wakeword (MicroWakeWord)) or server-side wakeword (OpenWakeWord) or no wake word at all (in case some end-user just want to use it as a media player output device for whatever reason).
The text was updated successfully, but these errors were encountered:
@Hedda thanks for making this recommendation. It makes complete sense to do what we can to support on-device and server-side wake word. I'll certainly look into this.
The link you provided is related to the VA component, which SAT1 is using too.
The PE software, on the other hand, doesn't have any OWW-related setup. It just relies on MWW, unlike previous configs from NabuCasa (i.e. for ESP32-S3_BOX3, where you could choose which engine to use).
Summary
Guard implementation against no MWW (Micro Wake Word) in case user want to use OpenWakeWord instead, or use no wakeword.
Basic example
Looks like this is implemented in home-assistant-voice-pe repository -> esphome/home-assistant-voice-pe#291
Motivation
Fact is some users prefer to use OpenWakeWord instead for now because creating your own MicroWakeWord model that works good is still hard.
End-user should have the option to use on-device wakeword (MicroWakeWord)) or server-side wakeword (OpenWakeWord) or no wake word at all (in case some end-user just want to use it as a media player output device for whatever reason).
The text was updated successfully, but these errors were encountered: