mini-voicecode

Soothe your tendinitis and code faster by voice.

All platforms

There's several dependencies which may be best built from source. They're included as git submodules. To download everything type:

git submodule init
git submodule update

Cooperating with other dictation software

You do not need other dictation software such as Dragon, Apple enhanced dictation, or Windows dictation. But you definitely might want to check them out. In order to cooperate between mini voice code and other software:

  • assign the F1 key to toggle your dictation software
  • assigned the voice command "awaken" to do nothing in your dictation software

Linux Install (Ubuntu 14.10)

sudo aptitude install python-pyaudio libxtst-dev

Get my fork of autopy: git clone https://github.com/gitfoxi/autopy.git

Build with: python setup.py install --user

Make sure your microphone is working (could be a fight):

parecord tst.wav
paplay tst.wav

Finally: ./voicecode

OS X install

brew install wxmac wxpython portaudio python autoconf libtool automake swig
brew install homebrew/python/numpy
pip install pyobjc
pip install py-applescript

I couldn't get homebrews version of sphinx to work so I've been compiling does from source. There are some modules directories for each of sphinxbase, pocketsphinx and sphinxtrain. go into each one and type:

./autogen.sh
./configure
make
make install

Install PyAudio from http://people.csail.mit.edu/hubert/pyaudio/

Get my fork of autopy in the autopy directory.

Build with: python setup.py install

Adjust your microphone levels. A good headset microphone is recommended. I've also had luck with a good table mic and the new macbooks have pretty-good mics built-in.

For best results set input output and sound effects to your headset in sound preferences.

You can use the famous line-in program to check what you sound like. https://www.rogueamoeba.com/freebies/

Go to system preferences. Under dictation and speech, turn dictation on and check "use enhanced dictation", and set the shortcut to F1. Under accessibility create a dictation command, "awaken". (First, enable 'advanced commands'.) Have it perform a paste text action, And leave the text blank.

Finally: ./voicecode

If it doesn't work the first time you read it, go to system preferences/ security and privacy and allow your terminal to control the computer in the privacy tab.

Helpful applications

OSX

  • KeyCastr gives visual feedback on the hidden kitty press
  • Spectacle shoves your windows around without you having to touch the mouse
  • Vimari a vim mode for Safari