I apologize if this has been answered before. I have added assist_pipeline: debug_recording_dir: /share/assist_pipeline
to the configuration.yaml file to fine tune the microphone settings. When the "Wake sound" is enabled the wake sound is at the beginning of the recordings. Previously I have always assumed that the wake sound was not sent to the SST. But I thought I would get an official answer. If the wake sound is passed to the SST, I would think this could be a problem with a wake sound that includes speech.
#Is the wake Sound sent to STT?
dim carbon
elder token
I apologize if this has been answered before. I have added```
assist_pipeline:
...
I guess that if the microphone is activated whilst the wake sound is still playing then it will pick itself up. the voice-pe counters this using the XMOS chip which can filter out stuff its outputting from the microphone.
that said i dont think i have ever heard of anyone using speech in their wake sound
dim carbon
I guess that if the microphone is activated whilst the wake sound is still playi...
Thanks for the info. I think there is a few of us out there that use speech for a wake sound. I do for the voice -pe, re-speaker-lite (FormatBCE), ESP32-S3-Box-3 (BigBobbas) and Satellite1 (Future Proof Homes). I have an example video in my https://github.com/mikey60/Voice-Assistant-Customizations/tree/main repository. It seems to work pretty well and I don't see the wake sound in the STT output, but occasionally I see a weird STT output. I don't believe that I hear the wake sound in the recordings from BigBobbas ESP32-SE-Box-3 voice assistant.
static dagger
Ok, now I want a custom wake sound that is the bat computer from the 1960s Batman TV show. "working....." 🤣