Replace constant audio bytearray with generated sine wave sample to get more robust results across test environments. Use zero-crossings as an audio fingerprint and compare it with sine wave frequency.