When you make a voice profile by saying ok Google 3x times it creates a key to compare the input to, it’s very efficient and takes up few CPU<\/p>\n
There is three states your microphone can be in for \u201cok Google\u201d detection. Most of the time when your screen is off Google is not listening to anything. Then there is passive listening to see if there is a match for your trained \u201cok Google\u201d then it switches to active listening to more carefully and accurately take your dictation and uploads to the server for analysis. Then you receive from the server what it thinks you said. That active listening is more intensive than the passive listening it does to detect \u201cok Google\u201d. That’s why you only train it to recognize \u201cok Google\u201d and you’re not reading multiple paragraphs to train it.<\/p>\n
Okay Google is always ready for your command
\nalthough you can enable and disable it
\nSometimes it is activated to hear the command from any other Screen( Any other launcher’s home screen)\u00a0 whereas on the other side it can be activated to hear only from home screen ( Google Now Launcher)<\/p>\n
If you want to disable it:-\u00a0<\/b><\/p>\n