Cat or Dog, Bat or Log

Here is the standalone version here

Why do this? So even though speech recognition is already well known and very accurate.  This is just looking at raw audio not words or speech. Using Cat and Dog just makes the test definitive with unique "sounds".

This compares incoming audio in comparison to the current background noise.

Think of an app that can detect and learn any type of sound that is new to the background noise and use that as a trigger.

Maybe your car is making a weird sound... perhaps the AI can learn by listening to normal car sounds and when different sounds occur and just one person or mechanic then identifies what it was and teaches the AI what it was, that now becomes identifiable. Imagine your car listens to itself, then warns you to get your tires balanced!  An unbalanced tire does create a recognizable resonance or shake in a car.

Rich Keller