Oversampling Data (Explained with Audio) - Computerphile
Computerphile
Apr 9, 2026
When digital audio processing creates harmonics that exceed the Nyquist-Shannon sampling limit, they 'mirror' back into the audible spectrum as harsh, inharmonic noise known as aliasing. Oversampling solves this by increasing the sample rate and using low-pass filtering to remove these artifacts, ensuring high-fidelity sound reproduction.
Key insight: If your sample rate is 10,000 Hz, the maximum frequency you can produce is 5,000 Hz; if you input a 6,000 Hz signal, the system doesn't just cut it off—it reflects it back down to 4,000 Hz, creating audible, inharmonic distortion.
