如何在内置输出上调用python声音设备包

Question

I am trying to record computer's audio output (not input to the microphone, but input to the speakers - before it leaves the device) on Mac OS X 10.14.1 with python3 and sounddevice.

As you can see from the following traceback, I have tried every sensible number of input channels and default number of input channels, with no luck. I have also looked at the source code, but haven't been able to determine my problem.

Is there a way to record the audio output stream on my system?

Python 3.6.2 (v3.6.2:5fd33b5926, Jul 16 2017, 20:11:06) 
[GCC 4.2.1 (Apple Inc. build 5666) (dot 3)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import sounddevice as sd
>>> sd.query_devices()
> 0 Built-in Microphone, Core Audio (2 in, 0 out)
< 1 Built-in Output, Core Audio (0 in, 2 out)
>>> sd.default.device = 1
>>> print("Channels should be 0 (number of input channels) or 2 (number of output channels)")
Channels should be 0 (number of input channels) or 2 (number of output channels)
>>> duration = 10
>>> fs = 44100
>>> x = sd.rec(int(duration * fs), samplerate=fs, channels=2)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 224, in rec
    ctx.input_dtype, callback, blocking, **kwargs)
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 2417, in start_stream
    **kwargs)
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 1301, in __init__
    **_remove_self(locals()))
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 780, in __init__
    'Error opening {0}'.format(self.__class__.__name__))
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 2572, in _check
    raise PortAudioError(errormsg, err)
sounddevice.PortAudioError: Error opening InputStream: Invalid number of channels [PaErrorCode -9998]
>>> x = sd.rec(int(duration * fs), samplerate=fs, channels=0)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 224, in rec
    ctx.input_dtype, callback, blocking, **kwargs)
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 2417, in start_stream
    **kwargs)
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 1301, in __init__
    **_remove_self(locals()))
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 780, in __init__
    'Error opening {0}'.format(self.__class__.__name__))
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 2572, in _check
    raise PortAudioError(errormsg, err)
sounddevice.PortAudioError: Error opening InputStream: Invalid number of channels [PaErrorCode -9998]
>>> x = sd.rec(int(duration * fs), samplerate=fs)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 215, in rec
    ctx.frames = ctx.check_out(out, frames, channels, dtype, mapping)
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sounddevice.py", line 2351, in check_out
    'Unable to determine number of input channels')
TypeError: Unable to determine number of input channels

Answer 1

It looks like you are telling it to record:

sd.rec(...

and the device you are recording has no inputs, which is true.

A good example of playing sounds is shown here:

https://python-sounddevice.readthedocs.io/en/0.3.12/examples.html#play-a-sound-file

or use the very concise example provided by Mark below.

As far as getting the output back into an input that takes some advanced work like down at the driver level. I note that there is an OSX virtual device called SoundFlower which does this. See:

https://apple.stackexchange.com/questions/221980/os-x-route-audio-output-to-audio-input

The source can be found here:

https://github.com/mattingalls/Soundflower

Take your choice on leveraging this existing tool or a much steeper path of trying to understand and extract the subset of behavior that you desire (the guts are in SoundflowerEngine::createAudioStreams).

Answer 2

I have wrote the following script which runs in both Python 2.x and 3.x.

The following code runs perfectly in windows but have not yet tested in MAC OS X. Hope this might work for you.

And most importantly don't forget to install sounddevice, soundfile, numpy

import sounddevice
import soundfile

rec_rate = 40000  # Hertz
rec_duration = 10  # seconds
rec_name = 'names.wav'  # Name for file
rec_data = sounddevice.rec(int(rec_rate * rec_duration), samplerate=rec_rate, channels=1, blocking=True)  # Recording ...
soundfile.write(rec_name, rec_data, rec_rate)  # Writing recorded sound in a file