Neural network for speech/audio volume recognition?

Question

Mobile development

Good day, dear toasters.
As one pet-project, I wanted to do something similar to the game that I saw in a musical comedy program on some federal channel. The essence of the game is quite simple - the louder you yell, the higher the ball jumps.
Starting to think about how I could implement something like that, I got stuck on the fact that I did not know how to implement the sound processing part. There are two options:
1) Determine only the volume of the sound, which should be easier.
2)speach-to-text recognition as a further development of the idea.
Tell me, please, in which direction to look, in order to still realize the idea. Should this be some kind of neural network or another ML model?
Many thanks in advance for your advice!

Answer 1

Loudness is the normal amplitude of the signal. It is extracted from any media format or directly from the microphone almost as primary data - there is not even any processing required. So google - "getting the amplitude from the microphone." What the hell is a neural network?

Answer 2

here