https://5dok.net/document/dzx5g5oq-audio-fingerprinting-speech-reconstruction-recognition-noisy-environments.html