MQTT Capture Incoming Voice

wilsomrm2 · August 17, 2021, 4:28pm

Hi All,

This is my first post; although I have been going through the forum for a while now.

I am trying to capture the incoming voice via an MQTT channel to convert to it a stored WAV file.

Would any of you know how to do this?

Any direction would be great as I have been on this one on and off for a month now.

All the Best
Mike

Daenara · August 17, 2021, 4:48pm

I did something like that with the wakeword here, that might help you out. Basically I buffer the mqtt audio messages and then save the last 10 or so seconds once the wakeword was detected. Same principle if you want to capture other parts of the audio, you just need another trigger and duration to save, or even add in a mqtt trigger for when to start listening.

Note that this does not work with udp audio activated because then only audio traffic after the wakeword will go over the mqtt server.

romkabouter · August 17, 2021, 5:08pm

I have a python script doing that, have to check if I can find it

romkabouter · August 17, 2021, 6:52pm

github.com

Romkabouter/ESP32-Rhasspy-Satellite/blob/voco/record.py

#!/usr/bin/python

import time
import sys
import os
import paho.mqtt.client as paho
import wave
import io
import pyaudio
broker = "192.168.43.79"

# folder = './tmp/'
# file_name = folder + 'wav_'

output = io.BytesIO()
p = pyaudio.PyAudio()
streamOut = p.open(format=pyaudio.paInt16,
                    channels=1,
                    rate=16000,
                    output=True)

This file has been truncated. show original

wilsomrm2 · August 17, 2021, 8:16pm

Very Interesting.

Thank you for that.

I will take a look.

wilsomrm2 · August 17, 2021, 8:19pm

Thank you for that. I will take a look.

wilsomrm2 · August 17, 2021, 9:08pm

Yes very very help thank you

All I needed to do is change the IP and edit the queue

hermes/audioServer/+/audioFrame

Very much appreciated.