Using sound file as pyaudio input device

Question

With PyAudio it is fairly simple to open an input stream comming from a physical sound input device.

I wonder if there is a way of open an existing sound file as a stream, which has the same properties as the sound-device-stream?

Instead of

    self.p = pyaudio.PyAudio()
    self.inStream = self.p.open(format=pyaudio.paInt16,
                                channels=1,
                                rate=self.RATE,
                                input=True,
                                frames_per_buffer=self.BUFFERSIZE)

I'd like to do

    self.p = pyaudio.PyAudio()
    self.inStream = self.p.open('testfile.wav',
                                input=True,
                                frames_per_buffer=self.BUFFERSIZE)

Regards, Torsten

No, in your own applcation. Just make a convenience function that has the interface that you want, which maps to PyAudio's way of doing it underneath. — Games Brainiac, May 22 '14 at 08:08
I understand. But if there was a transparent emulation mode you could modify every PyAudio application easily to work with sound files instead of devices. — Torsten2001, May 22 '14 at 11:46
I now tried to create a class `PyAudioFileStream`which has the same interface as `PyAudio.Stream`. This gave me some good results. I was able to use the wav-file in an PyAudio application by just replacing one line. — Torsten2001, May 22 '14 at 15:52

galarius · Answer 1 · 2015-10-23T09:43:30.337

As far as I know you can't use PyAudio to open file directly, however it's possible to combine pyaudio and wave modules in case your app reroutes the processed sound to SoundFlower for example.

import pyaudio, wave

def open_stream(self):
        self.p_audio = pyaudio.PyAudio()
        self.file_source = wave.open('input.wav', 'rb')
        s_width = self.file_source.getsampwidth()
        format = self.p_audio.get_format_from_width(s_width)
        channels = self.file_source.getnchannels()
        rate = self.file_source.getframerate()
        self.stream = self.p_audio.open(format=format,
                                        channels=channels,
                                        rate=rate,
                                        frames_per_buffer=1024,
                                        input=False,
                                        output=True,
                                        output_device_index=detect_build_in_output_device_idx(self.p_audio),
                                        stream_callback=self.__recording_callback)


def __recording_callback(self, in_data, frame_count, time_info, status):

        # read frames
        in_data = self.file_source.readframes(frame_count)
        # decode frames
        left, right = audio_decode(in_data, self.file_source.getnchannels())
        # process frames
        left, right = self.core.process(left, right)
        # encode back
        processed_data = audio_encode((left, right))

        return processed_data, pyaudio.paContinue

Using sound file as pyaudio input device

1 Answers1