Mixing involves balancing / processing / FX etc… Mastering is everything that happens once you have completed your mix; it cannot change the balance of instruments, the parameter settings on pre-existing effects etc…

Key mastering processes

• Dynamic range – compression / limiting / multiband compression are used to control the dynamic range of the recording; this increases (average) volume / perceived loudness whilst preventing distortion / peaking.
• Frequency response – subtle adjustments are made to the overall EQ (not instruments). This often boosts low frequencies and enhances high frequencies, giving power and clarity. Linear phase EQ used to avoid smearing; an HPF is used to remove rumble with a cutoff frequency of less than 35Hz (avoiding compression pumping / to get a louder master) / LPF can be used to remove hiss (remastering). Notch filters remove resonances / hum.
• Stereo width – LFs are normally narrow; HFs are spread to give width. 60s mixes are sometimes rereleased in mono because stereo is too polarised. Stereo reverb can be used to give width and mastering suites e.g. Ozone include imagers which enhance the stereo width. MS processing can treat the middle differently to the sides.
• Top and tail – master starts within 1s; last note decays naturally / fade out prepared – leaves reverb tail!

Other mastering processes

• Mastering normally take place in an acoustically treated room; mix translation is checked on different speakers
• Mastering engineers might use a little reverb in the final master to glue the mix together in the same space
• MS (mid-side) processing can be used to treat the middle and sides of the mix in different ways
• Radio edits are created – short versions of tracks designed for better radio play that cut long solos or drum intros
• Noise reduction is used to reduce hiss / clicks / hum / dropouts
• Software can be used to assist mastering e.g. Ozone or SoundCloud’s online mastering service using rules / AI
• When preparing an album, it is normal to match volume between each track; reference tracks can be used
• Dithering is used for digital masters which adds a layer of noise to randomise quantisation distortion
• Exciters can be used to add extra harmonics at the top end – normally lost as a result of tape recording in the 70s
• Remastering tries to improve the quality of the original recording / make it cater to modern trends or norms
• Masters should be bounced at high quality; at the very least an uncompressed WAV file

Mastering trends

• 60s: masters prepared on analogue tape and transferred to vinyl. Processes involved some use of HPF / LPF to avoid hum / hiss and some modest EQ. The mono master was the key one. Vinyl mastering used RIAA curve.
• 70s: warm sound; less upper-mids / highs. Noise reduction (more tracks). Masters transferred to cassette / vinyl
• 80s: masters in the 1980s had more upper-mid / high frequencies; ‘bass-light’ when compared with today. Digital technology started to become part of mastering e.g. digital reverb; use of SSL bus compressor.
• 90s: DAT / ADAT gave all digital mixing and mastering chains; loudness wars meant CD mastering engineers tended to view 0dBFS as a target not a limit.
• 00s: Low and high-frequency heavy compared to old masters; loudness wars intensified post-2000 and reduced post-2010 with LUFS / loudness penalty and the rise of streaming. Dolby Atmos / AI mastering.

Mastering for different formats

• Vinyl: Low frequencies reduced to prevent stylus from moving too much; EQ applied on playback to counter (RIAA curve). Issues can occur with sibilance and low end imaging. More compression is needed the slower a record turns. The louder the signal, the bigger the groove so less time fits on a side. Sound quality deteriorates at the centre of the record so the track order has an impact on quality.
• Tape: Tape bias (high frequency high energy signal) ensures that the iron oxide particles respond evenly to the magnetic signal transferred to tape. Saturation (on master tape) can be desirable or cause issues at higher volume. Noise reduction (normally Dolby or dbx) is used to minimise hiss.
• CD: CDs have a wider and more accurate frequency response and less noise than other media because they are digital. They hold 74-80 minutes of stereo LPCM audio (16 bit / 44.1kHz). CDs also have a better signal-to-noise ratio than tape or vinyl, with a wider dynamic range. In the 90s, engineers saw 0dBFS as a target not a limit (loudness wars). Includes sequencing the tracks, writing metadata and deciding on the gaps between the tracks.
• Digital audio: uncompressed / lossy / lossless format choice affects mastering process; some streaming services now use a process called loudness normalisation. This analyses the average volume of a song and turns it up or down to avoid the listener having to change the volume to compensate for louder or quieter masters.