[Python] "Microphone access error for real-time voice recognition in Google Colab...

I am trying to implement real-time voice recognition using Google Colab. However, I am encountering an issue where Colab cannot access the microphone, resulting in the following error:

PortAudioError: Error querying device -1

From my understanding, Colab might not support direct microphone access. I want to process live audio input rather than recording a file and processing it later.

Is there a workaround to enable microphone access in Colab, or is there an alternative approach to achieve real-time speech recognition in a cloud-based notebook environment?

Below is my current code:

import sounddevice as sd
import queue
import json
from vosk import KaldiRecognizer

q = queue.Queue()


def callback(indata, frames, time, status):
if status:
print(status, flush=True)
q.put(bytes(indata))

def recognize_speech():
device_info = sd.query_devices(None, "input")
samplerate = int(device_info["default_samplerate"])

rec = KaldiRecognizer(model, samplerate)
rec.SetWords(True)

with sd.RawInputStream(samplerate=samplerate, blocksize=8000, dtype="int16",
channels=1, callback=callback):
print(" ")
while True:
data = q.get()
if rec.AcceptWaveform(data):
result = json.loads(rec.Result())
print(result["text"])


recognize_speech()


The error

PortAudioError Traceback (most recent call last)
<ipython-input-10-83e5d8e96a1a> in <cell line: 0>()
31
32
---> 33 recognize_speech()

1 frames
/usr/local/lib/python3.11/dist-packages/sounddevice.py in query_devices(device, kind)
570 info = _lib.Pa_GetDeviceInfo(device)
571 if not info:
--> 572 raise PortAudioError(f'Error querying device {device}')
573 assert info.structVersion == 2
574 name_bytes = _ffi.string(info.name)

PortAudioError: Error querying device -1

Continue reading...