Pulse code modulation (PCM) is a method for converting analog signals, like sound waves, into digital data. It works by measuring the amplitude of a signal at regular intervals and expressing each measurement as a binary number. This simple three-step process (sampling, quantizing, and encoding) is the foundation of nearly all digital audio you encounter today, from phone calls to CDs to streaming music.
How PCM Works
An analog sound wave is a smooth, continuous curve. To turn it into something a computer can store and transmit, PCM breaks that curve into thousands of individual snapshots per second. Each snapshot captures the wave’s amplitude at that instant and rounds it to the nearest value on a fixed scale. That rounded value is then written as a string of binary digits (bits). The result is a stream of numbers that, when played back, reconstructs the original sound.
Three settings define any PCM signal:
- Sample rate: how many snapshots are taken per second, measured in hertz (Hz).
- Bit depth: how many possible amplitude values each snapshot can be assigned, which determines precision.
- Number of channels: mono (one) or stereo (two), for example.
A higher sample rate captures more detail in the time dimension, while a higher bit depth captures more detail in the loudness dimension. Together they determine both the fidelity and the file size of the resulting digital audio.
The Sampling Rule That Makes It All Work
For PCM to faithfully reproduce an analog signal, the sample rate must be at least twice the highest frequency in that signal. This principle, known as the Nyquist-Shannon sampling theorem, sets a hard floor. Human hearing tops out around 20,000 Hz, so a sample rate above 40,000 Hz is enough to capture everything you can hear. If the sample rate is too low, high-frequency content folds back into the audible range as a false, distorted tone called an alias.
This is why the standard CD sample rate lands at 44,100 Hz: it comfortably exceeds twice the upper limit of human hearing, leaving a small margin for the filters that prevent aliasing.
Bit Depth and Audio Quality
Bit depth controls how precisely each sample’s amplitude is recorded. With 16 bits, there are 65,536 possible values for each snapshot. With 24 bits, that jumps to over 16 million. The practical effect is dynamic range, the gap between the quietest sound the system can capture and the loudest before distortion. Each additional bit of depth adds roughly 6 dB of dynamic range. A 16-bit recording offers about 96 dB of range, while a 24-bit recording reaches around 144 dB.
The rounding that happens when a continuous amplitude is forced onto a fixed scale introduces a tiny amount of error called quantization noise. It’s the digital equivalent of tape hiss. Every time you add a bit of depth, the quantization noise level halves, pushing it further below the audible signal. At 16 bits, this noise is already well below what most listeners can detect in a normal environment. At 24 bits, it’s essentially inaudible under any condition, which is why professional studios record at 24 bits: the extra headroom makes mixing and editing forgiving without any risk of audible artifacts in the final product.
PCM in Everyday Technology
CDs and Home Audio
The Red Book standard, established in the early 1980s for audio CDs, specifies 2-channel (stereo) linear PCM sampled at 44,100 Hz with 16-bit depth. This format, often called CD-DA, remains one of the most widely recognized digital audio standards. “Linear” PCM (LPCM) simply means the quantization steps are evenly spaced. It’s the default uncompressed audio format on Blu-ray discs and HDMI connections as well.
Phone Calls
Digital telephone systems use a leaner version of PCM defined by the ITU-T G.711 standard. The voice signal is sampled at just 8,000 Hz (adequate for speech, which carries most of its intelligibility below 4,000 Hz) and compressed from 16 bits down to 8 bits using a logarithmic curve. North America and Japan use one variant of this curve (called mu-law), while most of Europe and the rest of the world use another (A-law). Both achieve 2:1 compression, cutting the data rate to 64,000 bits per second per channel while maintaining what the telecom industry calls “toll quality” voice.
High-Resolution Audio
Formats marketed as “hi-res” typically use PCM at 24 bits with sample rates of 96 kHz or 192 kHz. The higher sample rate pushes the Nyquist limit far above human hearing, which matters less for playback and more for digital signal processing. When audio is edited, effects like equalization and reverb perform mathematical operations on the samples, and having extra bandwidth above the audible range reduces the chance of artifacts folding back into the signal during those calculations.
Why PCM Became the Standard
The idea dates to 1937, when British engineer Alec Reeves was struggling with a fundamental problem in long-distance telephony: every time an analog signal passed through an amplifier, noise accumulated. Amplify a signal ten times across a continent and you amplify ten layers of static along with it. Reeves realized that transmitting a digitized version of the voice signal would sidestep this entirely. A digital signal only needs to be recognized as a 1 or a 0, so each relay point can regenerate a clean copy instead of amplifying a degraded one. He filed a French patent in 1938 and received U.S. Patent 2,272,070 in 1942.
The technology was ahead of its time. The vacuum-tube electronics of the 1940s couldn’t practically handle the switching speeds PCM demanded. It wasn’t until transistors and integrated circuits matured in the 1960s and 1970s that PCM became economically viable for telephone networks, and then for consumer audio in the 1980s with the arrival of the compact disc.
PCM vs. Other Digital Audio Approaches
PCM isn’t the only way to digitize sound, but it’s the most direct. Delta modulation, for instance, records only the difference between consecutive samples rather than each sample’s absolute value. This uses fewer bits per sample but sacrifices accuracy for complex signals. Pulse-width modulation (PWM) encodes information in the varying width of pulses rather than in binary amplitude values. Compared to both, PCM places its switching noise at much higher frequencies, making it far easier to filter out. In direct comparisons, PCM-based converters have achieved 9-bit effective resolution where equivalent PWM designs reached only 7 bits.
Compressed formats like MP3 and AAC start with PCM data and then discard information deemed inaudible to shrink the file. They aren’t alternatives to PCM so much as layers built on top of it. When you rip a CD or record in a studio, the raw capture is PCM. Compression happens afterward.
How Data Rate Is Calculated
The bitrate of an uncompressed PCM stream is straightforward arithmetic: sample rate × bit depth × number of channels. For a standard CD, that’s 44,100 × 16 × 2, which equals 1,411,200 bits per second, or about 1.4 Mbps. One minute of stereo CD audio takes up roughly 10.6 MB of storage. A 24-bit, 192 kHz stereo file runs at about 9.2 Mbps, nearly seven times the data rate of a CD. This is why lossless compression formats like FLAC exist: they reduce PCM file sizes by 30 to 60 percent without discarding any data, making high-resolution audio more practical to store and stream.