Reading soundcard output with python

Question

i have successfully utilized pyaudio to read and record microphone data. I would like to do the same for the soundcard output. (i.e. playing a wavfile and open a stream with pyaudio and read frame by frame the soundcard output).

For reading mic i opened a pyaudio stream like the following

stream = pyaud.open(
    format = pyaudio.paInt16,
    channels = 1,
    rate = 22050,
    input_device_index = 0,
    input = True)

when i change the input_device_index to 1 i get the following error

IOError: [Errno Invalid number of channels] -9998

Below is the system_info for the macbook soundcard

Default Devices:
================
Default Input Device : 0
    defaultSampleRate: 44100.0
    defaultLowOutputLatency: 0.01
    defaultLowInputLatency: 0.00199546485261
    maxInputChannels: 2
    structVersion: 2
    hostApi: 0 (Core Audio)
    index: 0
    defaultHighOutputLatency: 0.1
    maxOutputChannels: 0
    name: Built-in Microph
    defaultHighInputLatency: 0.0121541950113
    --------------------------------
Default Output Device: 1
    defaultSampleRate: 44100.0
    defaultLowOutputLatency: 0.00655328798186
    defaultLowInputLatency: 0.01
    maxInputChannels: 0
    structVersion: 2
    hostApi: 0 (Core Audio)
    index: 1
    defaultHighOutputLatency: 0.0167120181406
    maxOutputChannels: 2
    name: Built-in Output
    defaultHighInputLatency: 0.1
    --------------------------------

I wonder... You have a wav file. Why do you want to pass it through the sound card first? Read it with the 'wave' python module directly! You have direct access to the frames, and all the wave parameters. Much easier (seems to me, at least)! — jcoppens, Apr 05 '15 at 14:12
In my scenario there is not a wavfile available. Let's say that i want to process an audio stream (i.e. a voip call or the audio of a youtube video) and do some signal processing in realtime (i.e. compute loudness and do some annotation). — Panos, Apr 05 '15 at 14:47

score 0 · Accepted Answer · answered Apr 05 '15 at 15:18

Some soundcards have 'Capture' channels, which is the output of the soundcard. They may or may not work correctly (there seems to have been pressure from the music industry to eliminate those ports).

The following is for Linux... Eg. on my system, I only have a standard, on-board ALC887 audio port. In /proc/asound/card0, I find a port pcm2c, which has the following info:

card: 0
device: 2
subdevice: 0
stream: CAPTURE
id: ALC887-VD Alt Analog
name: ALC887-VD Alt Analog
subname: subdevice #0
class: 0
subclass: 0
subdevices_count: 1
subdevices_avail: 1

These capture devices should show up in your audio mixer.

Reading soundcard output with python

1 Answers1