Course contentsShow
← Machine Learning and Deep Learning
Lesson 2434 of 3,53853. Speech and Audio ProcessingFree lesson

Sampling Rate and the Nyquist Theorem

Understanding sampling frequency, aliasing prevention, and why 44.1kHz/48kHz are common standards.

Sampling Rate and the Nyquist Theorem

What you'll learn: How sampling rate determines audio quality and why the Nyquist theorem prevents information loss in digital audio.

The Core Idea

When we convert continuous sound waves into digital audio, we sample them—taking snapshots of the wave's amplitude at regular intervals. The sampling rate (measured in Hz) tells us how many snapshots per second we capture. For example, 44.1kHz means 44,100 samples every second.

But here's the critical question: How fast do we need to sample to capture all the information in the sound?

The Nyquist Theorem

The Nyquist-Shannon sampling theorem provides the answer: to perfectly reconstruct a signal, you must sample at at least twice the highest frequency present in that signal.

This minimum rate is called the Nyquist rate. If you sample below this threshold, you get aliasing—where high frequencies masquerade as lower ones, creating distortion and artifacts you can't remove.

Why 44.1kHz and 48kHz?

Human hearing typically ranges from 20Hz to 20kHz. According to Nyquist:

  • Minimum sampling rate = 2 × 20kHz = 40kHz

So why 44.1kHz and 48kHz?

  • 44.1kHz (CD quality): Provides a safety margin above 40kHz, with room for anti-aliasing filters
  • 48kHz (professional audio/video): Even more headroom, standard for film and broadcast

Think of it like taking photos of a spinning wheel: sample too slowly, and it might appear to spin backward (aliasing). Sample fast enough, and you capture its true motion.

Key Takeaway: The Nyquist theorem requires sampling at twice the highest frequency to avoid aliasing; 44.1kHz and 48kHz became standards because they safely exceed the 40kHz Nyquist rate for human hearing (20kHz max), with practical margins for filtering.