All of these programs, I noticed one drawback.
The user must press a certain button that would activate the
speech recognition process.
Why can not this activation process also hang on a voice command?
For example: In the phone settings recorded voice one code word.
Those. configure the user presses the codename Voice Assistant, then pronounces the name
and the program writes the name of his memory.
Next, the program simply has to listen to the microphone for
Detection of such words and activate the main voice assistant,
which may already be working through an Internet or still as audio-Tb.
As a result, we get a very convenient scheme of work.
Pronounce codename (phone or headset), such as "piece of iron" - and then say what we want, "they say, a piece of iron, called my mother." ....
As a result, we ring in HTC and flies at $ 1000 on roaming. - this is a joke.
Post has been editedkupava - 02.11.12, 01:47