Audio file formats

A digital audio file stores sound as a sequence of numbers obtained by sampling an analogue waveform many thousands of times per second and recording the amplitude of each sample. The most direct representation is pulse-code modulation (PCM), in which the raw sample values are stored without compression. PCM is the basis of uncompressed formats and is defined by two parameters: the sample rate, which sets the highest frequency that can be captured, and the bit depth, which sets the resolution and dynamic range of each sample.

Raw samples need a container that records how to interpret them. The WAV format wraps PCM audio in Microsoft's RIFF structure, a chunked layout that stores the sample rate, channel count, bit depth and the audio data together. Because WAV is uncompressed it preserves the original samples exactly but produces large files. Other containers and formats add compression. Lossless schemes shrink the data while allowing the exact samples to be recovered, whereas lossy schemes achieve much smaller files by discarding detail.

The widely used MP3 format, part of the MPEG family, applies a perceptual model to drop sound the ear is unlikely to notice, giving small files at the cost of some fidelity. WMA is a comparable lossy format from Microsoft. For film and surround sound, multichannel formats such as AC-3 (Dolby Digital) carry several audio channels in a compressed stream, as used on DVD and DVD-Audio. Choosing a format is a balance between file size, sound quality and compatibility with the playback equipment.

Frequently asked questions

What is PCM?

Pulse-code modulation stores digital audio as the raw amplitude values of each sample without compression; it is the basis of formats such as WAV.

What is the difference between lossless and lossy formats?

Lossless compression reduces file size while letting the exact original samples be restored, whereas lossy compression discards some detail permanently to achieve smaller files.

What does a WAV file contain besides audio?

Within its RIFF structure a WAV file records the sample rate, number of channels and bit depth alongside the sample data, so a player knows how to interpret it.