karen.listener.Listener

Listener device to capture audio from microphone and convert any speech to text and send to callback method.

__init__(self, speechModel=None, speechScorer=None, audioChannels=1, audioSampleRate=16000, vadAggressiveness=1, speechRatio=0.75, speechBufferSize=50, speechBufferPadding=350, audioDeviceIndex=None, callback=None) special

Listener Initialization

Parameters:

  • speechModel (str) – Path and filename of Deepspeech Speech Model file. If not set then listener will do a basic seach for the PBMM or TFLite file.

  • speechScorer (str) – Path and filename of Deepspeech Scorer file. Okay for this to be None as scorer file is not required.

  • audioChannels (int) – Audio channels for audio source. VAD requires this to be 1 channel.

  • audioSampleRate (int) – Audio sample rate of audio source. VAD requires this to be 16000.

  • vadAggressiveness (int) – Voice Activity Detection (VAD) aggressiveness for filtering noise. Accepts 1 thru 3.

  • speechRatio (float) – Must be between 0 and 1 as a decimal

  • speechBufferSize (int) – Buffer size for speech frames

  • speechBufferPadding (int) – Padding, in milliseconds, of speech frames

  • audioDeviceIndex (int) – Listening device index number. If not set then will use default audio capture device.

  • callback (function) – Callback function for which to send capture text

start(self, useThreads=True)

Starts the listener to listen to the default audio device

Parameters:

  • useThreads (bool) – Indicates if the brain should be started on a new thread.
Returns:

  • (bool) – True on success else will raise an exception.

stop(self)

Stops the listener and any active audio streams

Returns:

  • (bool) – True on success else will raise an exception.

wait(self, seconds=0)

Waits for any active listeners to complete before closing

Parameters:

  • seconds (int) – Number of seconds to wait before calling the "stop()" function
Returns:

  • (bool) – True on success else will raise an exception.