Additional cmdline argument for pyaudio

README.md

@@ -86,6 +86,13 @@ According to this table, the previous command can be run as:
 
     rec -q -t raw -r 16000 -c 1 -b 16 -e signed - | auditok -i -
 
+### PyAudio
+
+When capturing input with PyAudio, you may need to adjust the device index with -I if multiple input devices are available. Use `lsusb -t` to get the list of usb devices, or use `arecord -l` if you're using a non-usb input device. If you don't know what index to use, just try `0`, `1`, `2` and so on, outputting the audio using `-E` (echo) until you hear the sound.
+
+You may also get an error `[Errno -9981] Input overflowed` from PyAudio. If that's the case, you need a bigger frame buffer.
+Use `-F` with 2048 or 4096 (the default is 1024).
+
 ### Play back detections
 
     auditok -E

auditok/cmdline.py

@@ -576,6 +576,8 @@ def main(argv=None):
         group.add_option("-r", "--rate", dest="sampling_rate", help="Sampling rate of audio data [default: %default]", type=int, default=16000, metavar="INT")
         group.add_option("-c", "--channels", dest="channels", help="Number of channels of audio data [default: %default]", type=int, default=1, metavar="INT")
         group.add_option("-w", "--width", dest="sample_width", help="Number of bytes per audio sample [default: %default]", type=int, default=2, metavar="INT")
+        group.add_option("-I", "--input-device-index", dest="input_device_index", help="Audio device index [default: %default] - only when using PyAudio", type=int, default=None, metavar="INT")
+        group.add_option("-F", "--audio-frame-per-buffer", dest="frame_per_buffer", help="Audio frame per buffer [default: %default] - only when using PyAudio", type=int, default=1024, metavar="INT")
         parser.add_option_group(group)
 
         group = OptionGroup(parser, "[Do something with detections]", "Use these options to print, play or plot detections.") 
@@ -609,7 +611,9 @@ def main(argv=None):
             try:
                 asource = PyAudioSource(sampling_rate = opts.sampling_rate,
                                         sample_width = opts.sample_width,
-                                        channels = opts.channels)
+                                        channels = opts.channels,
+                                        frames_per_buffer = opts.frame_per_buffer,
+                                        input_device_index = opts.input_device_index)
             except Exception:
                 sys.stderr.write("Cannot read data from audio device!\n")
                 sys.stderr.write("You should either install pyaudio or read data from STDIN\n")

auditok/io.py

@@ -351,10 +351,12 @@ class PyAudioSource(AudioSource):
     def __init__(self, sampling_rate=DEFAULT_SAMPLE_RATE,
                  sample_width=DEFAULT_SAMPLE_WIDTH,
                  channels=DEFAULT_NB_CHANNELS,
-                 frames_per_buffer=1024):
+                 frames_per_buffer=1024,
+                 input_device_index=None):
 
         AudioSource.__init__(self, sampling_rate, sample_width, channels)
         self._chunk_size = frames_per_buffer
+        self.input_device_index = input_device_index
 
         import pyaudio
         self._pyaudio_object = pyaudio.PyAudio()
@@ -370,6 +372,7 @@ def open(self):
                                                        rate=self.sampling_rate,
                                                        input=True,
                                                        output=False,
+                                                       input_device_index=self.input_device_index,
                                                        frames_per_buffer=self._chunk_size)
 
     def close(self):