Audio mastering

Mastering converts an approved mix into a technically consistent and distributable release. A mastering engineer first checks the supplied files, reference mixes, intended format, and any known problems such as clicks, clipped samples, excessive noise, or intersample peaks. Listening takes place on accurate monitoring systems and at several playback levels, because decisions must hold across headphones, cars, consumer speakers, and broadcast-style environments.

Processing may include equalization, compression, limiting, expansion, stereo-field adjustment, harmonic coloration, and precise fades. The engineer normally works conservatively: mastering cannot reliably repair a badly balanced arrangement or a distorted mix. Communication with the mixing engineer is therefore part of the process, and a revised mix is often preferable to aggressive corrective processing.

For an album or EP, mastering also sets track order, spacing, start and end points, transitions, and overall level relationships. The result is commonly approved through reference files or a listenable test master before final delivery.

Loudness management balances perceived level with musical dynamics and delivery requirements. Peak level describes instantaneous signal amplitude, whereas loudness estimates are designed to correspond more closely to human perception; the ITU-R BS.1770 family specifies measurement methods based on weighted audio energy and channel factors.¹ The EBU R 128 framework applies related loudness and true-peak practices to broadcast production.²

A limiter can prevent peaks from exceeding a chosen ceiling, but excessive limiting may produce pumping, distortion, flattened transients, and listener fatigue. Streaming services may alter playback gain through loudness normalization, so maximizing level does not necessarily make a track sound louder in context. Mastering instead weighs transient impact, spectral balance, genre conventions, and the artist's intent. A technically compliant master can still translate poorly if low bass, vocal presence, or stereo information changes substantially on smaller systems.

Delivery requirements determine which master is made and how it is checked. Digital releases may require uncompressed PCM files at specified sample rates and bit depths, while compact discs use 16-bit, 44.1 kHz PCM and vinyl involves physical limitations that can influence side length, low-frequency content, stereo bass, and cutting level. The mastering engineer may create separate versions rather than forcing one master into every medium.

Quality control examines file headers, sample rate, bit depth, channel order, duration, fades, silence, clipping, and track boundaries. Metadata may include title, artist, composer, genre, copyright, ISRC, and artwork, although different distributors accept and transform different fields. Archival practice favors well-documented, lossless files and preservation metadata; the Library of Congress identifies format specifications and sustainability factors as important parts of digital audio stewardship.³ Industry standards and recommended practices are maintained through organizations such as the Audio Engineering Society.⁴

Mastering is also an editorial and forensic discipline, not merely a loudness-increasing step. Engineers may prepare instrumental, clean, radio-edit, performance, immersive, and alternate-resolution versions, each requiring its own verification. A hidden defect can arise at the boundaries: a truncated reverb tail, an abrupt fade, a polarity reversal, or a brief digital glitch may escape casual listening but become obvious in a gapless album sequence.

True-peak control matters because digital-to-analog conversion can create peaks between stored samples even when sample peaks remain below 0 dBFS. Channel layout and downmix behavior become especially significant for multichannel and immersive releases. For historical recordings, restoration and mastering overlap: clicks, hum, tape azimuth, wow and flutter, and changing source levels may require specialist intervention before musical processing. The original mix, approved master, delivery master, and lossy encoded copy are distinct objects; retaining each version improves both accountability and future reissue work.

True peak

An estimate of the highest signal level after digital reconstruction, including peaks that occur between stored samples.

Loudness normalization

Playback gain adjustment intended to make programs with different measured loudness appear more comparable in level.

Dither

Deliberate low-level noise added during word-length reduction to decorrelate quantization distortion.

ISRC

International Standard Recording Code, an identifier assigned to an individual sound recording.

Pre-emphasis

A historical high-frequency boost applied before recording or transmission and reversed during playback; some legacy digital material requires special handling.