This repository contains scripts to collect data for speech recognition.

The script manual.py records audio and also records all X11 keystrokes as they
are manually entered. (It should work on other platforms, but has only been
tested on Linux.) This is useful if your speech is being interpreted by another
speech recognition system, or by a human typist. To use it, try:

    $ PYTHONPATH=pynput/lib/ python manual.py -L -o /tmp

This will generate .flac and .txt files with audio recordings and keypresses.
Press CTRL-C to stop the recording, and it will exit safely. Due to a bug in
the keystroke handler, you may need to press one additional key (e.g. CTRL)
before the keystroke thread exits.

To check microphone device numbers, use

    $ python list-mics.py