search

ChemiKhazi / Samaritan

Samaritan interface.
10 stars 8 forks source link

Use microphone API to make Samaritan respond to changes in noise level. #6

Open ChemiKhazi opened 9 years ago

ChemiKhazi commented 9 years ago

(Also, turn the Samaritan app into a thing that listens into all the PoI fans so that Samaritan can mark them as deviants… wait a minute)

Wojbie commented 9 years ago

Could google speach to text api be used to "read" spoken message? Next thing you know you will get random "Donation" from unknown source for new UI for system.

ChemiKhazi commented 9 years ago

It looks like both speech to text and text to speech are doable right now with HTML5.

I'm not sure how speech to text would work for this though, parsing it to make Samaritan respond coherently seems like too much work.

Speech to text is more easily usable but going by the show canon, The Machine is the AI that speaks to its agents using sound.

Wojbie commented 9 years ago

Could just go and make that few keywords like:

"Episode" - prints latest episode name and relase date. (would require some kind of web lookup for that) "Status" - "Operational" "Ok Google" - "Google Can't Save you now" "Siri" - "She is Gone" Any character name from Series - "Irrelevant" "Machine" - "Tracking in progress" "Reddit: - "Cat Cat Cat"

I will look into speech to text stuff day after tomorrow and see what i can cobble together.

Wojbie commented 9 years ago

Looked into it in free time today. speech to text works only on desktop chrome as of this moment. Sadly android chrome (where i was aiming that part at) does not support it right now. Will still take time to code something for desktop chrome in coming days! And some kind of workaround to use build in voice keyboard on android... like semi-invisible text input instead of microphone icon? Will research more later.