Domoticz

Hello,

I'm appealing to the community to see if anyone knows how to detect a specific audio sound and then trigger a virtual switch.
For example I configure the words "turn on" and "turn off", and when I say these words, and only at the sound of my voice, not of another person, the virtual switch closes or opens, and only when I say these 2 words, the switch must not react to any other sound or audio noise.

Thank you for your idea

Hi, @cyberbob69,

You touch 2 problems:

1. Detect your voice, reliable
2. Detect the two words "turn on" and "turn off"

Ad 1:
From experience, it is possible to learn the Google Assistant to discriminate your voice from others.
This requires some investment, but you need a microphone/loudspeaker anyhow.

Ad 2:
Node Red, together with SmartNora node or any other Google integration with either Node Red or Domoticz is able to give you your vitual output.
With SmartNora, you can use the open/close node. With other integration I do not have any experience, but maybe others are able to help you.

This is only a suggestion, but I never tried that.

Regards

My wish is also that the system recognizes a specific sound. As each sound has a wave of its own, we can compare it and if it corresponds with a certain tolerance of difference, we make the scenario and activate the virtual switch that we want.
Let us take the example of a dog or a domestic cat, when it emits a sound, the same frequencies are always almost the same, it would suffice for the system to compare this sound and for an action to be made on a virtual switch.

I know it's possible but I don't see too much of a solution.

Thanks again for your answers

You could do something with for example https://github.com/bishoph/sopare and have it set a switch with a Domoticz http api call.
But I checked offline voice/sound recognition a little bit and the hardest thing is to find a good microphone. I apparently works if you are close by but not from a couple of meters away.