Microphone (and wake word) not working

[Benjamo44]

Hi all, this is my first ever Github post, for my first ever Github project. Please forgive me if I am asking this in the wrong section.

I have followed the instructions to build the XiaoZhi AI chatbot, and everything is working perfectly, except the INMP441 microphone.

I have tried the project using three different ESP-32 dev kits - 2 x ESP32-S3 and 1 x ESP-32, and while the LED screen works perfectly and the speaker/amplifier also work perfectly, I cannot get the microphone to recognise wake words or anything at all.

I have been able to switch the chatbot to 'Listening' by connecting GIO5 to Ground, but all the chatbot seems to 'hear' is the word 'I' or 'Yes'. I have been able to create an agent at xiaozi.me in the Console, and the chatbot plays songs and tells me jokes, but I cannot do anything to interact with it.

I am Australian and no doubt have a very strong accent, but is anybody able to help me understand why the chatbot does not recognise me?

Finally, I have tested the 3v3 and ground going in to the microphone and both seem ok. I have attached a photo of my setup with the ESP-32D.

Thank you so much for everything, I am LOVING this project!