How Digital Audio Sampling Actually Works: The Mechanics of the Nyquist-Shannon Theorem

Audiophiles look at a vinyl record and see a continuous, infinite-resolution physical wave, arguing that digital audio is a jagged, discrete staircase of 1s and 0s that misses the spaces between the samples. Digital engineers look at the exact same digital file and see mathematical perfection, arguing that analog media is a noisy, distorted approximation of the original performance. The disagreement is not just a matter of taste; it is a collision between the physics of mechanical playback and the mathematics of signal processing.[6]

The "staircase myth"—the idea that digital audio is a blocky, low-resolution approximation of a smooth analog wave—is perhaps the most persistent misconception in consumer audio. As Christopher "Monty" Montgomery of the open-source multimedia foundation Xiph.Org explains, "a digital waveform is not a stairstep, and you certainly don't get a stairstep when you convert from digital back to analog."[5]

To understand why the staircase is a myth, you have to look at the mathematics of digital signal processing. In 1949, mathematician Claude Shannon built upon the work of Harry Nyquist to prove one of the most important results in information theory: the Nyquist-Shannon sampling theorem.[3]

The theorem states that a continuous analog signal can be perfectly reconstructed from discrete samples, provided the sample rate is more than twice the highest frequency present in the signal. As the digital signal processing curriculum at Carnegie Mellon University notes, "sampling does not degrade the sound in any way, as long as we sample fast enough." If this condition is met, no information is lost between the samples.[1][3]

This is why the compact disc standard was set at a sample rate of 44.1 kHz. Human hearing tops out at roughly 20 kHz. By sampling at 44.1 kHz, digital audio captures more than twice the highest frequency a human can hear, leaving a small 2.05 kHz safety buffer for the anti-aliasing filter to do its work. The CD format also established a 16-bit depth, providing a dynamic range of 96 decibels—far exceeding the physical capabilities of a vinyl record.[3][4]

When a digital file is played back, the digital-to-analog converter (DAC) does not output a jagged staircase of discrete voltages. Instead, it passes the discrete samples through a low-pass reconstruction filter. This filter mathematically connects the dots, resulting in a smooth, continuous analog wave that is identical to the original signal that entered the microphone. The "gaps" between the samples are mathematically redundant; there is only one possible continuous wave that can pass through those specific points without exceeding the Nyquist frequency.[2][5]

If the sample rate drops below the Nyquist threshold, the system fails. As the engineering documentation from MathWorks explains, "When a signal is sampled below the Nyquist rate, high-frequency components are 'folded' back into lower frequencies, creating inaccurate data in the reconstructed digital signal." This phenomenon, known as aliasing, permanently destroys the original audio.[2]

If digital audio is mathematically perfect, why do so many listeners insist that vinyl sounds better? The answer lies not in vinyl's accuracy, but in its physical imperfections. Vinyl playback is an inherently mechanical process, and that physical friction leaves a sonic fingerprint. As the audio publication STMedia notes, "Vinyl captures audio as a continuous physical waveform etched into grooves. This waveform is an exact physical representation of the original sound wave."[6]

Analog systems naturally introduce low-order harmonic distortion. When a stylus tracks a groove, it adds subtle harmonic frequencies that were not present in the original recording. These harmonics—particularly second-order and third-order harmonics—are mathematically related to the fundamental notes, making instruments like upright bass, saxophone, and human vocals sound richer and more intimate.[6]

Furthermore, vinyl physically smooths out harsh high frequencies. Digital formats can reproduce piercing, high-frequency transients with absolute precision, which can sometimes sound metallic or fatiguing to the human ear. Vinyl's mechanical limitations naturally roll off these extreme highs, resulting in a "warmer" sound profile.[6]

The physical limitations of the vinyl groove also prevent the extreme dynamic compression that plagues modern digital mastering. If a vinyl record is mastered too loudly or with too much bass, the physical stylus will literally jump out of the groove. Uncompressed CD audio streams at a bitrate of 1.4 megabits per second (Mbps), which allows for massive dynamic swings, but modern pop mastering often squashes that range to make tracks sound louder on streaming services.[6]

Because of the physical constraints of the needle, vinyl masters are forced to retain a wider dynamic range. The quiet passages remain quiet, and the loud crescendos hit with more impact. This dynamic contrast is a major reason why vinyl records often feel more "alive" than their heavily compressed digital counterparts.[6]

The analog versus digital debate is not a question of which format is more accurate. Digital audio is objectively, mathematically more accurate, capable of perfectly reconstructing the original sound wave with a lower noise floor and zero wow or flutter. The real question is whether absolute accuracy is what the human ear actually prefers, or if the mechanical imperfections of a physical groove are the very thing that makes music feel human.[7]