AudioUtils
Audio Glossary

What Is PCM Audio?

PCM (Pulse Code Modulation) is the raw representation of digital audio: a stream of evenly spaced samples, each storing the amplitude of the waveform at that moment. Every WAV file, every audio CD, every track loaded into a DAW exists as PCM at the moment of playback. This page is a reference card for the math, the common rates, and the role PCM plays in the conversion pipeline.

The PCM Sampling Model

Audio is a continuous waveform. PCM digitises it by measuring the amplitude at fixed time intervals — the sample rate — and storing each measurement as an integer of fixed width — the bit depth. A 44.1 kHz / 16-bit PCM stream takes 44,100 amplitude readings per second, each from a 65,536-step scale (-32,768 to +32,767). No bit is added or removed for compression: the file is exactly as wide as the sample count multiplied by the bit width multiplied by the channel count. PCM is the unambiguous reference every other audio format eventually decodes back into. MP3, AAC, Opus, FLAC — all decode to PCM in memory before reaching the DAC. See [pcm audio explained](/blog/pcm-audio-explained) for diagrams.

Where PCM Lives

PCM is the substrate, not the file format. WAV (.wav) is the most common PCM container — Microsoft's RIFF wrapper plus a header describing rate, depth, and channel layout. AIFF (.aiff) is Apple's equivalent: same PCM payload, big-endian byte order, identical fidelity. Audio CDs store 16-bit 44.1 kHz stereo PCM in a slightly different framing (CD-DA) — ripping to WAV with [MP3 to WAV](/convert/mp3-to-wav)'s sister tool [FLAC to WAV](/convert/flac-to-wav) is bit-perfect. Inside a DAW, samples are floating-point PCM (32-bit float typically) for headroom during processing. Hardware: every audio interface DAC reads 16-bit or 24-bit linear PCM at the chosen sample rate. PCM is the lingua franca; everything else is a transport format.

Sample Rate Reference

8 kHz: telephone-quality voice, narrowband. 16 kHz: wideband VoIP, modern voice codecs at low rates. 22.05 kHz: low-quality music, retro game audio. 32 kHz: FM broadcast, some long-form voice archives. 44.1 kHz: CD-DA standard, every consumer music release. 48 kHz: industry standard for film, video, broadcast, gaming, DAW projects with video. 88.2 kHz / 96 kHz: professional recording with processing headroom; 96 kHz is the most common 'high-res' rate. 176.4 / 192 kHz: archival and audiophile, rarely audibly distinguishable from 96 kHz. The Nyquist theorem caps representable frequency at half the sample rate; 44.1 kHz captures up to 22.05 kHz, comfortably above the 20 kHz upper limit of human hearing.

Bit Depth Reference

Bit depth controls dynamic range. Each additional bit adds ~6 dB of representable range. 8-bit PCM: 48 dB range — voice memos, vintage games. 16-bit PCM: 96 dB range, the CD and consumer-audio standard. 24-bit PCM: 144 dB range, the recording-and-editing standard; gives 24 dB of safety headroom over 16-bit so loud transients do not clip during tracking. 32-bit float: dynamic range exceeds any analogue source; clipping inside a DAW is mathematically reversible because peaks above 0 dBFS are still represented. Always record at 24-bit and master to 16-bit at the very end; converting [WAV to MP3](/convert/wav-to-mp3) after a 24-bit master keeps the dither stage clean.

File Size and Bitrate Math

Bitrate of PCM equals sample_rate x bit_depth x channels. Worked examples: 44.1 kHz x 16-bit x 2 channels = 1,411,200 bits per second = 1411 kbps = ~10.1 MB/min. 48 kHz x 24-bit x 2 channels = 2,304,000 bps = 2304 kbps = ~16.5 MB/min. 96 kHz x 24-bit x 2 channels = ~33 MB/min. Multiply by minutes for total size. Compared to a 192 kbps MP3 (~1.4 MB/min), CD-quality PCM is ~7x larger; 96/24 PCM is ~24x larger. This is why lossy codecs exist — the perceptual data per sample is far less than 16 bits in most musical material. See [why wav files are large](/blog/why-wav-files-are-large) for a fuller breakdown.

PCM in DAW and Plugin Chains

Every plugin and every DAW process operates on PCM, almost always 32-bit floating-point internally. When you load an MP3 into Logic, the file is decoded to float PCM on import; mixing happens in PCM; the bus sums in PCM; the export converts PCM back to your chosen delivery codec. Working with PCM source files (WAV, AIFF, FLAC) skips the decode step at load time, reduces CPU at the start of playback, and avoids any decoder-specific colouration. For multi-take editing, comping, and time-stretching, PCM is mandatory in practice — lossy formats produce smear and pre-echo when stretched or pitched. Use [MP3 to WAV](/convert/mp3-to-wav), [FLAC to WAV](/convert/flac-to-wav), or [M4A to WAV](/convert/m4a-to-wav) before importing into a DAW for any non-trivial edit job.

When PCM Is the Wrong Choice

PCM is the right master and edit format, but the wrong distribution format. A 4-minute stereo song at 44.1/16 PCM is ~40 MB — too big to email, too costly to stream, too slow on slow connections. Always render PCM masters first, then derive distribution copies via lossy or losslessly compressed codecs. Lossy distribution: [WAV to MP3](/convert/wav-to-mp3) at 192 kbps cuts size by 90% with imperceptible quality loss for most listeners. Lossless distribution: [WAV to FLAC](/convert/wav-to-flac) compresses by 40-60% with bit-perfect decode. If you need to keep PCM but trim its footprint, [compress a WAV file](/compress-wav) by lowering bit depth or sample rate. Keep PCM for working files and the long-term archive; derive everything else on demand. See [flac vs wav for music production](/blog/flac-vs-wav-for-music-production).