dascontrol

Computer listens people.

python RUNME.py

Proof of concept

[x] Get voice sample
[x] Convert sample into text (recognize)
[x] Select command (match)
[x] Execute command

User experience

[x] Single top level command to start the process
[ ] Signal to start and end recording
[ ] Show "..processing.." state while recognizing
[ ] Show text

Components

[x] Recorder   (mic -> sample.wav)
  [ ] (lev.2) UI (press button -- record -- unpress)
[x] Recognizer (sample.wav -> sample.txt)
[x] Matcher    (sample.txt -> action.txt)
[x] Executor   (action.txt -> ...)

Out of scope

[ ] Continuous voice scan
     (no technology available, no time to reseach
      current version of Sphinx CMU)

Phase One (Done)

Requisites:

• Linux
• curl
• jshon
• sox
• clisp

Every component is shell script:

[x] Console scripts

Phase Two (Done)

Requisites:

• sox
• Python

Go cross-platform

[x] Python (Lin/Win/OS) recording
[x] Python Google API request (requests lib)
[x] Python Matcher
[x] Python Action Executor

Phase Three

Graphics UI

• pyglet/pygame

• or anything (web/kivy + websockets)

  [ ] three static screens [ ] state machine [ ] animations