Sample rate

Sample rate is the number of times per second that an analog audio signal is measured and stored as a digital value, expressed in hertz (Hz) or samples per second. In digital audio, it determines the highest frequency a system can represent: according to the sampling theorem, a sample rate must be more than twice the highest frequency in the signal, a limit known as the Nyquist frequency.

Together with bit depth, sample rate defines the resolution of a PCM recording. Sample rate governs the time and frequency axis, while bit depth governs the amplitude axis and dynamic range. Common values are 44.1 kHz for music distribution and 48 kHz for video, broadcast and most professional production.

How sampling captures a continuous signal

Sampling converts a continuous waveform into a sequence of discrete measurements taken at evenly spaced moments. An analog-to-digital converter reads the instantaneous voltage at each tick of a precise clock, and each reading is then quantized to a number. At 48 kHz, the interval between two samples is about 20.8 microseconds.

The Nyquist-Shannon sampling theorem states that a band-limited signal can be reconstructed exactly from its samples if the sample rate exceeds twice its highest frequency. Reconstruction does not mean connecting the dots with straight lines; the digital-to-analog converter uses a reconstruction filter that restores the smooth original waveform within the band.

Any content above half the sample rate cannot be represented correctly. Instead it folds back into the audible band as aliasing, a form of distortion that is not harmonically related to the original sound. For this reason every converter places an anti-aliasing low-pass filter before the sampling stage.

The table below shows how sample rate relates to the sample interval and to the theoretical frequency limit.

Standard sample rates and where they are used

Standard sample rates exist because every device in a chain must agree on one rate, and each industry settled on values that suited its technology. The two dominant families are multiples of 44.1 kHz and multiples of 48 kHz.

The 44.1 kHz rate comes from the compact disc standard, chosen to cover human hearing up to about 20 kHz with room for a practical filter and to fit early video-based recording equipment. The 48 kHz rate became the standard for film, television and professional digital audio, and it divides cleanly into common video frame rates.

Some delivery formats, such as DSD, use a very different approach: a 1-bit stream at several megahertz with noise shaping. Rates like 2.8224 MHz are not directly comparable with PCM sample rates.

Sample rate compared with bit depth and Nyquist frequency

Sample rate is often confused with neighboring concepts in digital audio, but each describes a different dimension of the signal. The table separates them clearly.

For uncompressed PCM, the bit rate follows directly from these values: sample rate × bit depth × channels. A stereo file at 48 kHz and 24 bits therefore needs about 2.3 Mbit/s, or roughly 17 MB per minute. Doubling the sample rate doubles the file size without changing the dynamic range.

Choosing a sample rate for a project

The right sample rate is usually the one the final delivery requires, because every conversion between rates adds processing and a small risk of artifacts. Higher rates help in a few specific cases, but they cost storage, CPU load and plugin capacity.

Higher rates can reduce aliasing in nonlinear plugins such as saturators and compressors, but many modern plugins already use internal oversampling. In most production work, a well-designed converter at 48 kHz is audibly transparent for listeners.

Common misconceptions and mistakes

Most sample rate problems come from mismatched settings or exaggerated expectations rather than from the rate itself. The following points cover the errors that appear most often in practice.

• Mismatched rates between devices: an interface clocked at 44.1 kHz playing a 48 kHz session makes audio play about 9 percent fast and sharp unless the software converts it.
• Believing higher rates mean smoother waves: within the audible band, a correctly reconstructed 44.1 kHz signal is not a staircase; the extra samples only extend the frequency range.
• Upsampling to improve quality: converting a 44.1 kHz file to 96 kHz adds no information and only enlarges the file.
• Ignoring clock synchronization: digital devices that are not locked to one word clock produce clicks and dropouts.
• Poor sample rate conversion: low-quality converters introduce aliasing or ripple, so use a high-quality resampler when conversion is unavoidable.

The short version

Sample rate is how many snapshots of sound a recorder takes every second. Think of a film camera that takes many still frames per second so that motion looks continuous. A digital recorder does the same with sound, measuring the wave tens of thousands of times each second.

More snapshots let the recording capture higher pitches. Human hearing tops out around 20,000 vibrations per second, so the common rates of 44,100 and 48,000 snapshots per second already cover everything people can hear. Going higher mostly creates bigger files.

Where it makes a difference

A video editor receives music stems recorded at 44.1 kHz and drops them into a film timeline set to 48 kHz. The editing software flags the mismatch and converts the files on import. Had the conversion been skipped, the music would have played about 9 percent too fast and roughly a semitone and a half sharp, drifting out of sync with the picture.

On the same film, a sound designer records breaking glass at 96 kHz with a microphone that captures content above 20 kHz. When the recording is slowed to a quarter of its speed for a dramatic effect, the ultrasonic detail moves down into the audible range and the slowed sound stays bright and detailed instead of turning dull.