I liked that with Google Home speakers, there was a small delay (1 second maybe?) between when you finish speaking and when the voice is sent to be transcribed and processed.
In ESPHome based satellites, there is an option called "Finished speaking detection" which adds a delay like that, but it doesn't seem like wyoming-satellite does. As a result, if you're talking slowly or make pauses between words (for example if a person has just woken up, or if they are a non-native speaker), this results in incomplete voice recordings and thus incomplete requests.
It would be preferred if the option could be set as a number (float?) in seconds to wait before declaring that the user finished speaking.
I liked that with Google Home speakers, there was a small delay (1 second maybe?) between when you finish speaking and when the voice is sent to be transcribed and processed. In ESPHome based satellites, there is an option called "Finished speaking detection" which adds a delay like that, but it doesn't seem like wyoming-satellite does. As a result, if you're talking slowly or make pauses between words (for example if a person has just woken up, or if they are a non-native speaker), this results in incomplete voice recordings and thus incomplete requests. It would be preferred if the option could be set as a number (float?) in seconds to wait before declaring that the user finished speaking.