Help reading audio data from I2S microphones

Question

Using the Adafruit I2S MEMS microphone breakout board and the Raspberry Pi 2B, I am trying to read in stereo input using PyAudio and pass the data through a FFT to analyse. However I am not able to record the data from the microphone and plotting the data shows a high frequency square wave.

I have installed the I2S drivers according to the Adafruit website's tutorial https://learn.adafruit.com/adafruit-i2s-mems-microphone-breakout/ and I am able to record stereo audio using the following arcord command in terminal.

arecord -D plughw:1 -c2 -r 48000 -f S32_LE -t wav -V stereo -v file_stereo.wav

If anyone could help point me into the right direction that would be very helpful. Here is the code that I am using.

import pyaudio
import numpy as np

CHUNKSIZE = 1024 # fixed chunk size
RATE = 44100

# initialize portaudio
p = pyaudio.PyAudio()

# Get input device number
info = p.get_host_api_info_by_index(0)
numdevices = info.get('deviceCount')
for i in range(0, numdevices):
        if (p.get_device_info_by_host_api_device_index(0, i).get('maxInputChannels')) > 0:
            print "Input Device id ", i, " - ", p.get_device_info_by_host_api_device_index(0, i).get('name')

stream = p.open(format=pyaudio.paInt16, channels=2, rate=RATE,
                input=True, frames_per_buffer=CHUNKSIZE,
                input_device_index = 2)

# do this as long as you want fresh samples
data = stream.read(CHUNKSIZE)
numpydata = np.fromstring(data, dtype=np.int16)

# close stream
stream.stop_stream()
stream.close()
p.terminate()

# Output as a WAV file
import scipy.io.wavfile as wav
wav.write('out.wav',RATE,numpydata)

Answer 1

I had a similar issue but was able to get a non-zero recording by using the pyaudio.paInt32 format. My sampling frequency and block size are different in my code but I doubt that will be a problem. If you look at the data blocks when using paInt16, they are all 0 valued, i.e., x00. Maybe something inherent in the DMIC? There is also a fixed DC off-set as you can see in the image below.

Hopefully, the dual channel case also works. I don't have an extra DMIC to test it on.

#!/usr/bin/env python3
import pyaudio
import wave

CHUNK = 4096
FORMAT = pyaudio.paInt32
CHANNELS = 1
RATE = 16000 
RECORD_SECONDS = 10
WAVE_OUTPUT_FILENAME = "test.wav"

p = pyaudio.PyAudio()

stream = p.open(format=FORMAT,
            channels=CHANNELS,
            rate=RATE,
            input=True,
            frames_per_buffer=CHUNK) 

frames = []

for i in range(0, int(RATE / CHUNK * RECORD_SECONDS)):
     data = stream.read(CHUNK)
     print(data)
     frames.append(data)

stream.stop_stream()
stream.close()
p.terminate()

wf = wave.open(WAVE_OUTPUT_FILENAME, 'wb')
wf.setnchannels(CHANNELS)
wf.setsampwidth(p.get_sample_size(FORMAT))
wf.setframerate(RATE)
wf.writeframes(b''.join(frames))
wf.close()