Customizing the TF-lite android

I’m trying to customize the source code for TF-lite audio classification for android,
can I process the recognized sound data and extract the sound intensity with this?

Hi @mschoi ,

Welcome to The TensorFlow Forum .

You will have to write some custom code to caluclate the Root Mean Square of the audio buffer as a simple measure of intensity. Not sure how accurate this will be but this is one way to implement it .

Thank You .