The video from Real Engineering explores how the Shazam app identifies songs in noisy environments. It demonstrates Shazam recognizing a song and explains the complexities of programming a computer to do this. While our brains can quickly identify songs through complex neural pathways and pattern recognition, programming a computer to match audio data is a significant challenge.
Shazam uses a process called “fingerprinting,” where a song is transformed into a visual representation called a spectrogram. This spectrogram is simplified into a “fingerprint” highlighting key frequencies, reducing data for quicker matching.
The video also discusses how the human brain’s recognition of music differs from a computer’s method, emphasizing the importance of timbre (the unique sound quality of instruments) and overtones in identifying music. The ingenious coding behind Shazam’s technology led to the app’s acquisition by Apple for $400 million.