# *-* coding: utf-8 -*-
"""Audio module to manage I/Os and streams.
This module manages the audio I/Os of the computer and send the stream to the
speech recognition engine. It's purpose is to unify all type of sources
(physical or application streams) into one unique system without having
worriying about source types.
Note:
At the moment, this module takes only the input from the default microphone
of the laptop. This has to be improved in the future and find a way
configure easily the defaults.
Todo:
This needs to be abstracted in order to be able to select any type of
source on the computer and send the stream to a speech_recognition class
that will be an interface to any type of speech recognition engine.
- Give the possibility of creating audio dump in a wav format by giving a
flag in the command line
"""
import pyaudio
import time
p = pyaudio.PyAudio()
[docs]def callback(in_data, frame_count, time_info, status):
""" Audio callback of the hardware
This function is the callback (non-blocking) of the audio function and its
purpose is to connect the input to the output.
Returns:
input data
a bit at 0 in order to continue to deliver the buffers.
Todo:
This function should be renamed in order to have a better idea about
its purpose. It's here as well that the signal processing should be
done. It would be nice in the future to be able to load whatever
signal blocks we want and create a desired chain at startup (or even as
hot plug).
"""
return (in_data, pyaudio.paContinue)
[docs]def start_recording(
width=2,
channel_number=1,
sample_rate=44100):
""" Start the stream.
This function starts the audio stream according to few parameters. It stays
active thanks to the infinite loop.
Todo:
Several things need to be done:
- Rename the function as it doesn't perform recording but more
streaming
- Give the possibility to add more parameters (see PyAudio
documentation... frame size, device index, ...)
- Write a small function for the width conversion that isn't really
easy to handle (audiobits2bits)
"""
stream = p.open(format=p.get_format_from_width(width),
channels=channel_number,
rate=sample_rate,
input=True,
output=True,
stream_callback=callback)
stream.start_stream()
while stream.is_active():
time.sleep(1.1)
stream.stop_stream()
stream.close()
p.terminate()