by Renman Mattias
Voice activate your smart assistant with any greeting and start giving your voice instructions.
Try three different flows triggering the speech recognition block using any hotword:
1. with only speech recognition (3 blocks)
2. at fixed sound level (6 blocks)
3. at dynamic sound level (31 blocks)
Only using the speech recognition block as in flow 1 is a simple approach that already adapts for surrounding noise or different microphone levels although it may not be suited for "always on" hotword listening.
Flow 2 and 3 is an attempt to bridge this issue without implementing an external hotword listening app, like the Tasker Hotword plugin.
Flow 2 listens for how much the microphone level is changing. When the level is changed above an fixed change limit "40", speech recognition is started.
The "40" limit must be adjusted manually to adapt to different microphone base levels cased by background noise or by the microphone itself.
Flow 3 also listens for how much the microphone level is changing by a louder noise, but will gradually adapt the sound change trigger limit by also listening to the microphone sound or "base" level previous to the hotword is spoken.
The change level is thus dynamically calculated in flow 3 and will automatically adapt for different microphones or changing background noise.
Flow 3 is also adapted for "always on" use.
Rate and review within the app in the Community section.