Is there any way to use SpeechRecognition with wasapi loopback?

[ghost]

What problem are you facing?

• audio isn`t recorded
• audio is recorded with artifacts
• problem with "silence"
• [ x] other

What is the cause of the error (in your opinion)?

• PyAudio\PortAudio bug
• [ x] I just need help(or answer)

❗ Before creating, complete the checklist:

• I have provided a minimal reproducible example (and checked its correctness)
• [ x] I'm sure I fully described the problem

If the problem is related to the inability to record sound from speakers,
then before creating this issue, try to first record audio using the provided examples
(for example, simple_recording_app.py)

Start your issue here 👇

Is there any way to use SpeechRecognition with wasapi?
like this
`
import pyaudiowpatch as pyaudio
import speech_recognition as sr

p = pyaudio.Pyaudio()
recognizer = sr.Recognizer()

default_speakers = p.get_default_wasapi_loopback()

rate = int(default_speakers["defaultSampleRate"])
channels = default_speakers["maxInputChannels"]
print(f"Recording from: ({default_speakers['index']}) {rate} {channels} {default_speakers['name']}")

stream = p.open(
format=data_format,
channels=channels,
rate=rate,
frames_per_buffer=1024,
input=True,
input_device_index=default_speakers["index"],
)

recognizer.listen_in_background(stream.read(4096)) # i tried but doesnt work

`
any help would be appreciated!

[ghost]

(in real time)

[s0d3s]

Hi🖐 Its depends on SpeechRecognition. If it can do STT in real-time, then you can do it. In turn, PyAudio can do what you need, but maybe you will need to increase the frames_per_buffer size, or do manual buffering

[ghost]

Could you provide me for the code as an example ? Much appreciated!

[s0d3s]

@imposterfromtheamongus You can use this project as reference:
https://github.com/SevaSk/ecoute/
https://github.com/SevaSk/ecoute/blob/main/AudioRecorder.py

There inside custom version of SpeachRecognition but only reason for it - is to switch from PyAudio to WPatch-version. I think here exists option to use only custom Microphone class(not whole SpeachRecognition), which will be overriden with WPatch. But you have to work it out on your own. Good luck!