are there any audio experts out here that might be able to lend me a hand? Specifically I am looking for advice around extracting features from audio data that can then be fed into an AI algo to detect classifications or anomalies.