Need help in converting an audiosignal to a specific format for streaming

[elektromeier]

Hi there…

this is somehow a very specific question, and i admit i dont understand the basics of audio processing.

I need to stream an incoming audiostream (coming live from a microphone) to a voicerecognition server via mqtt.
the server needs a very specific how to prepare the audio in a way it can be sent as mqtt payload.
if the server recogniced a sentence it will publish a message also via mqtt (asr/match).

the guy prepared the voicerecogntion part described a bit how to do it, and he sent me a python script
which takes an audiostreams, prepares it and streams it via mqtt to the server.

i tried to repatch this script in vl but i guess im doing something wrong in converting the audiostream, as the voicerecognition does not react at all.

i attached my patch and the python script. [Audio_MQTT.zip|attachment]
Audio_MQTT.zip (529.1 KB)

im a bit desperate on this, if someone can help me i would be super thankful!
we also have one machine with the voicerecognition running as mqtt broker in the web. i deleted ip, user and password in the patch, but if someone wants to help me i would send this infos privately.

[motzi]

hey @elektromeier,

just to make sure - are you even connected to the broker? at least in the screenshot the “connected” output at the client is false and you have quite a count of messages queued up in the “Pending messages count”. (i guess you did not want to set a user/password for the screenshot but you probably one for your broker).