Skip to content

Glossary of mixing and mastering terms

Send feedback

Last updated 34 min readTags:Glossary

Look up mixing and mastering terms, including LUFS, phase, true peak, compression and dither.

Mixing shapes and balances the tracks of a song. Mastering prepares the resulting mix for release, including final tone, peaks, loudness and delivery files. See what is mastering?.

Find a term by letter. Sonalix entries distinguish the current report from terms you may still see in older reports or guides.

0–9 · A · B · C · D · E · F · G · H · I · K · L · M · N · O · P · Q · R · S · T · U · V · W · Z

0–9

0 dBFS. Full scale for an integer PCM sample. Fixed-point values peak at this limit. Floating-point audio can hold samples above it, and a reconstructed waveform can have a positive true peak even when the individual samples are below full scale. dBFS vs dBTP vs LUFS

5-band energy. An earlier Sonalix spectrum display covering Sub (20–100 Hz), Low-mid (100–500 Hz), Mid (500 Hz–4 kHz), Presence (4–8 kHz) and High (8–20 kHz). The current report uses seven measurement bands and a separate set of tonal comparison regions. 5-band energy

A

AAC. Advanced Audio Coding, a lossy codec used by Apple Music and other services. Compare an AAC preview with the original master if you hear distortion on loud peaks. How lossy codecs affect your master

A/B comparison. A comparison between two versions, or between your mix and a reference. Match their playback levels before judging the change. How to use reference tracks when mixing

Acoustic treatment. Absorbers, diffusers and bass traps that change how a room reflects sound so that you hear your speakers more accurately. In a small room, bass control comes first. Home studio acoustics

AIFF. Audio Interchange File Format. It commonly stores uncompressed PCM and can preserve the same samples as a PCM WAV.

Air. A broad term for the top of the spectrum, often above about 8 kHz. Sonalix uses Air for the 10–20 kHz measurement band. The frequency spectrum

Album loudness consistency. Balancing the loudness of tracks across an EP or album so the sequence plays smoothly, whatever the streaming service does on playback. Listen to the transitions and check each track's true peak. Album loudness consistency

Apple Digital Masters. Apple's programme for delivering high-resolution masters to Apple Music. Listen to an AAC preview of your master for clipping before you deliver. See also Sound Check. Mastering for Apple Music and Apple Digital Masters

ATSC A/85. The US broadcast loudness recommendation, which normally targets −24 LKFS. LKFS and LUFS are the same unit, so the target is 1 LU quieter than EBU R128's −23 LUFS. See also EBU R128. EBU R128 vs ATSC A/85

Attack. On a compressor or gate, how quickly it responds after the signal crosses the threshold. A slower attack lets more of the initial transient through. A faster one tames it. Audition settings by ear on the part that matters, such as the drum hit or the vocal consonant. Dynamic range compression

Automatic double tracking (ADT). An effect that makes one recording sound like two takes by combining a delayed copy with small, changing delay and pitch differences. Automatic double tracking · What is automatic double tracking? · ADT plugins

Automatic EQ. Equalisation that analyses the audio and suggests or applies settings, for example matching a reference, controlling resonances or correcting a room. Treat its settings as a starting point and compare bypassed and processed at the same level. Automatic equalisation

Automation. Recording or drawing changes to a control, such as fader level or an effect's mix, over time in your DAW. Use it when one setting cannot suit the whole song. Dynamic range control

Aux send. A control on a track that sends a copy of its signal to a separate return track, usually to share one reverb, delay or parallel compressor between several tracks. See also Send, Return and Bus. Bus structure and group processing

B

Bandwidth. The width of the range an EQ band affects around its centre. For the same centre frequency, higher Q means a narrower bandwidth. Exact conventions vary by filter.

Bell filter. An EQ band shape that boosts or cuts around a centre frequency and tapers off on either side, like a bell curve. It is the default band type on most parametric EQs. See also Parametric EQ.

Bit depth. The number of bits used to represent a sample. In integer PCM, each extra bit adds roughly 6 dB of theoretical range, about 96 dB at 16-bit and 144 dB at 24-bit. Use 24-bit for most working exports. See bit depth and Dither. Supported file formats, sample rates and file sizes

Bitrate. Data per second, usually in kilobits per second (kbps). PCM bitrate follows sample rate, bit depth and channel count. With a lossy codec, bitrate is one of the encoding settings that affects size and quality. How lossy codecs affect your master

Brick-wall limiting. Limiting with a ceiling the signal is not meant to exceed, however hard it is driven. It is the standard last stage for peak control. See also Limiter. Limiting

Bus. A channel that several tracks are routed into so they can be processed or balanced together, such as a drum bus or the mix bus. Bus structure and group processing

Bus compression. Compression applied to a group of tracks or the whole mix so that the parts react to the same dynamics, which many engineers describe as "glue". Use a low ratio and a few dB of gain reduction, then check that the drums have not lost their impact. See also Glue. Bus compression

C

Ceiling. The maximum output level a limiter aims to allow. Check whether the control refers to sample peak or true peak. Spotify recommends below −1 dBTP, or below −2 dBTP for masters louder than −14 LUFS. Limiting

Clip gain. A gain change applied to a single audio clip before any plug-ins or the fader, used to even out levels or bring a source to a working level. See also Gain staging and Trim. What is gain staging?

Clipping. Flattening the tops of a waveform because the signal exceeds the maximum a stage can pass. Digital clipping is harsh and usually unwanted. Some analogue-style clipping is used on purpose for colour. See also Soft clipping and Distortion. Clipping in audio

Comb filtering. Regularly spaced peaks and dips caused by summing a signal with a delayed copy. A doubled mic signal or a short delay can sound hollow when folded to mono. Phase cancellation

Compression. A dynamics process that reduces level differences. A downward compressor reduces gain as the signal exceeds the threshold. Ratio, attack and release govern how it responds. Dynamic range compression

Correlation. See Stereo correlation.

Crest factor. The ratio of peak amplitude to RMS amplitude over the same window, expressed as 20 × log10(peak ÷ RMS) in dB. Smaller values mean peaks sit closer to the RMS level. The older Sonalix guide discussed below 8 dB and 8–14 dB as illustrative ranges. The current report measures PLR and LRA. Crest factor

D

DAW. Digital audio workstation: the software you record, edit and mix in, such as Ableton Live, Logic Pro, FL Studio or Reaper.

dBFS. Decibels relative to full scale. Use it for digital signal levels such as sample peak or RMS. 0 dBFS is the fixed-point sample limit. Floating-point stages can hold higher values. dBFS vs dBTP vs LUFS

dBTP. Decibels true peak, relative to full scale. It expresses the estimated maximum amplitude of the reconstructed waveform, including points between samples. True peak

DDP. Disc Description Protocol: a set of files that describes a CD master, including track starts and text, used to send a CD to a pressing plant. CD Red Book standard

De-esser. A dynamics processor that reduces sharp consonants. Some models attenuate a frequency band and others turn down the whole signal when sibilance triggers them. A lisp is a sign to ease off. Sibilance and de-essing

Delay. An effect that repeats a signal after a set time, from a few milliseconds (which can widen or thicken a sound. See Haas effect) to echoes that follow the tempo. High-pass or filter the repeats so they do not clutter the mix.

Distortion. A change to the shape of a waveform. Non-linear examples such as saturation and clipping can create harmonics or intermodulation products. What is distortion in music production?

Dither. Low-level noise added before a reduction to a lower integer bit depth. It turns signal-related quantisation distortion into a low noise floor. Apply it immediately before final quantisation. See dither. CD Red Book standard

Double tracking. Recording the same part twice and layering the takes, which thickens vocals and guitars and widens them in the stereo field. Listen in mono, where tiny timing differences can cause cancellation. Double tracking

Downward compression. The ordinary kind of compression, which turns loud signals down. It is the opposite of Upward compression.

Ducking. Lowering the level of one sound while another plays, usually with a sidechain, such as pulling a pad down under a vocal. See also Sidechain and Pumping.

Dynamic EQ. An EQ band that only cuts or boosts when the signal in that band crosses a threshold. It suits problems that come and go, such as a vocal that gets harsh only on loud notes. Dynamic EQ

Dynamic range. The difference between the loudest and quietest parts of a signal. Different meters measure different versions of it, so say which you mean. See also Crest factor, Loudness range (LRA) and PLR. Dynamic range vs loudness range

E

Ear fatigue. Reduced sensitivity and growing discomfort after long or loud listening, which makes judgements about tone and level unreliable. Work at a repeatable, moderate monitor level and take breaks. Monitoring level and ear fatigue

EBU R128. The European Broadcasting Union's loudness recommendation for broadcast, with a target of −23 LUFS integrated. See also ATSC A/85 and LKFS. EBU R128 vs ATSC A/85

Equalisation (EQ). Changing the balance of frequencies in audio by boosting or cutting ranges. Use it to reduce rumble, soften harshness or help parts sit beside each other. What is equalisation (EQ)?

Equal-loudness contours. Curves (often called Fletcher–Munson curves) showing that the ear is less sensitive to low and high frequencies than to the midrange, and more so at low listening levels. This is why a mix that sounds balanced when loud can sound thin when quiet. Monitoring level and ear fatigue

Expansion. A process that increases level differences. Downward expansion turns down quieter sounds. Upward expansion accentuates louder ones. A noise gate is a strong form of downward expansion. Upward compression and expansion

F

Fader. The control that sets a track's level in the mix. Use faders for balance and set working levels earlier, with clip gain or a trim stage. See also Gain staging. What is gain staging?

FLAC. Free Lossless Audio Codec: a compressed file format that makes files smaller without discarding any audio, so it decodes to the identical samples. See also Lossless.

Frequency. How many times per second a sound wave repeats, measured in hertz (Hz). Low frequencies sound deep. High frequencies sound bright. The frequency spectrum

Frequency masking. See Masking.

Frequency response. How a device, filter or room changes the level or phase of signals at different frequencies. A mix spectrum instead shows the distribution of energy in the material itself. Frequency response curve

G

Gain reduction. How many decibels a compressor or limiter is turning the signal down at a given moment, usually shown on its meter. Compare gain reduction with how the sound changes, since the meter alone cannot tell you whether it sounds right. See also Over-compression.

Gain staging. Setting the level of a signal at each point in the chain, deliberately, so nothing clips, nothing sits in the noise and every plug-in receives a sensible input. Many analogue-modelled plug-ins expect around −18 dBFS to behave like 0 VU, but check the plug-in's manual. What is gain staging?

Gating. A loudness measurement rule which excludes quiet blocks. BS.1770 uses 400 ms blocks, a −70 LUFS absolute gate and a relative gate 10 LU below the loudness calculated after the absolute gate. See Noise gate for the audio processor. K-weighting and gating

Glue. A description of the way gentle bus compression makes separate parts feel like one performance. It is a perceived effect, not a measurement. See also Bus compression. Bus compression

Goniometer. A display that plots the relationship between the left and right channels. On the usual rotated display, a narrow vertical trace suggests mono and a horizontal trace suggests strong opposite-polarity content. Check the display orientation. Goniometers and vectorscopes

Graphic EQ. An EQ with fixed frequency bands, each controlled by a slider, so you move gain only. Handy for broad shaping. Less precise than a parametric EQ. Types of EQ

Group bus. A bus that collects related tracks, such as all drums or all backing vocals, so you can balance and process them together. See also Bus. Bus structure and group processing

H

Haas effect. The way a short delay of a few milliseconds up to roughly 30 ms between two copies of a sound is heard as one sound shifted towards the earlier one, not as an echo. Used to widen or position sounds. The delay can cause comb filtering in mono. Haas effect

Harmonics. Frequencies at whole-number multiples of a note's fundamental. Their balance gives an instrument its tone, and distortion and saturation add more of them.

Harshness. An unpleasant, piercing quality, usually from a build-up of energy in the upper midrange and presence region, roughly 2 to 8 kHz. Find the range with a narrow sweep before cutting. See also Resonance and Sibilance. The frequency spectrum

Headphones. A useful second view of a mix, especially for stereo detail and low-level noise, though they can exaggerate width and change how bass feels. Check your mix on speakers as well. How to mix on headphones

Headroom. The space between your signal's peak level and the maximum the system can pass. A signal peaking at −6 dBFS has 6 dB of sample-peak headroom. Headroom in mixing

Hertz (Hz). The unit of frequency, in cycles per second. 1 kHz is 1,000 Hz. See also Frequency.

High-pass filter (HPF). A filter that attenuates frequencies below its cutoff. A 12 dB/octave slope is gentler than 24 dB/octave. Raise the cutoff in context and listen for the point where the part begins to lose body. High-pass filters

I

Imaging. How clearly sounds are placed across the stereo field, and how stable and distinct those positions are. See also Stereo width. Stereo width explained

Insert. A plug-in placed directly in a track's signal path, so all the audio passes through it. The alternative is a Send.

Integrated loudness. The average loudness of a whole programme or track, measured with BS.1770 K-weighting and gating, in LUFS. It is the figure platforms use for normalisation. See also Short-term loudness and Momentary loudness. Integrated vs short-term vs momentary loudness

Inter-sample peak. A peak that exists in the reconstructed analogue waveform between two stored samples and can exceed the highest sample value, so a track can clip in conversion even when its DAW meter stays below 0 dBFS. See also True peak. Inter-sample peaks

ISRC. International Standard Recording Code: a unique identifier assigned to a specific recording so that its plays and sales can be tracked across services.

Issue categorisation. An older Sonalix report term for grouping findings under clipping, low end, dynamics, high end, midrange, stereo, loudness or a general category. In the current report, start with the ranked Mix improvements. Issue categorisation

ITU-R BS.1770. The international standard for measuring programme loudness. It defines K-weighting, the block-based averaging and the gating that LUFS meters use. K-weighting and gating

K

Knee. How gradually a compressor starts acting as the signal approaches the threshold. A hard knee switches on abruptly. A soft knee eases in, which tends to sound smoother on gentle compression.

K-weighting. The filter BS.1770 applies before measuring loudness, so that the meter responds more like human hearing: it reduces low-frequency energy and slightly lifts the upper range. It is why bass-heavy tracks read lower in LUFS than their peak levels suggest. See also LUFS and Gating. K-weighting and gating

L

Limiter. A compressor with a very high ratio and a ceiling, used to stop peaks exceeding a set level. It is usually the last processor in a mastering chain. Use gain reduction in moderation and listen for lost punch. See also Ceiling, Lookahead and Release. Limiting

Linear-phase EQ. An EQ that changes tone without shifting the timing relationships of frequencies, at the cost of added latency and possible pre-ringing. It is mostly useful in mastering or on parallel paths. See also Minimum-phase EQ and Pre-ringing. Linear-phase EQ

LKFS. Loudness, K-weighted, relative to full scale: the name used by ATSC and ITU documents. LKFS and LUFS are the same unit, and 1 LU equals 1 dB. ATSC A/85 normally targets −24 LKFS. EBU R128 vs ATSC A/85

Lookahead. A limiter setting that lets the limiter see the signal slightly ahead of time so it can turn down smoothly before a peak arrives. More lookahead can reduce distortion but may soften the attack of drums. Limiting

Lossless. A format that keeps every sample of the audio, such as WAV, AIFF or FLAC. Lossless files can be converted between lossless formats without changing the sound.

Lossy codec. A compression scheme that permanently discards audio that is judged least audible, such as MP3, AAC or Ogg Vorbis. Peaks can rise and artefacts can appear, especially on loud, dense masters. How lossy codecs affect your master

Loudness matching. Setting two versions or tracks to the same perceived loudness before comparing them, usually with a loudness meter, so level does not bias your choice. Loudness targets by genre

Loudness normalisation. Adjusting a track's playback level, up or down, so that tracks play at a similar loudness. Streaming services do this on playback. The file itself is not changed. Spotify's Normal setting uses a −14 LUFS reference, not a rule for how you must master. See also Spotify normalisation and Sound Check. Loudness normalisation

Loudness range (LRA). A measure, in LU, of how much loudness varies across a programme, from quiet sections to loud sections. It describes the contrast between sections, not the peak-to-average ratio inside a section. See also Crest factor. Loudness range (LRA)

Loudness war. The long-running trend of mastering music ever louder to stand out, at the expense of dynamics. Streaming normalisation reduces the advantage of a louder master. The loudness war

Low-mid. A region commonly understood as roughly 100–500 Hz. Sonalix’s earlier five-band display used 100–500 Hz. Its current measurement band is 120–350 Hz, and its tonal comparison uses 120–400 Hz. 5-band energy

Low-pass filter (LPF). A filter that removes frequencies above a cutoff and lets lower ones through. It is useful for taming hiss or making a sound darker and more distant.

LU. Loudness units: the unit for differences in loudness. 1 LU equals 1 dB. LUFS values are absolute. LU describes relative differences, as in loudness range.

LUFS. Loudness units relative to full scale. BS.1770 uses K-weighting to calculate loudness. Integrated LUFS also uses gating. More negative values mean lower measured loudness. What LUFS actually means

M

Makeup gain. Gain added after a compressor to restore the level it turned down. Match the processed and bypassed levels, or the louder version will seem better whether or not it is. See also Compression.

Masking. When one sound hides another because they occupy similar frequencies at similar times. It is why a part can seem to vanish even though it is clearly audible alone. Solve it with level, arrangement, EQ or sidechain compression, in that order. Frequency masking

Mastering. Preparing a finished stereo mix for release: final tonal balance, loudness, peak control, and the files each format needs. What is mastering?

Mastering chain. The ordered set of processors used in mastering, commonly some corrective EQ, gentle compression and a limiter, with checks of loudness and true peak at the end. A typical mastering chain · Master equalisation

Matching EQ. EQ that compares the spectrum of your audio with a reference and builds a correction curve. Apply a fraction of the suggested curve and judge by ear. See also Automatic EQ. Automatic equalisation

Mid. The sum of left and right in mid/side encoding, commonly scaled as (L + R) / 2. Centred sounds contribute strongly, and panned material can contribute too. Mid/side processing

Mid/side (M/S). An encoding of stereo as a sum (mid) and a difference (side). It lets you process the two components separately. Check the stereo version and the mono sum after a change. Mid/side processing

Minimum-phase EQ. The usual EQ type, whose filters shift the phase of the signal around the boosted or cut region without adding pre-ringing or large latency. See also Linear-phase EQ. Types of EQ

Minor adjustments. An older Sonalix report label for lower-priority advice. In the current report, try the ranked Mix improvements in order and audition each separately. Minor adjustments

Mix bus. The channel that all tracks and buses feed into before the master output. Processing here affects the whole mix. Keep its peaks well below 0 dBFS. The mix bus is not a limiter. See also Premaster. Bus structure and group processing

Mix review. A check of a mix before mastering or delivery. A Sonalix mix review compares the uploaded file with stated measurements and offers changes to audition. See your mix report. What is a mix review?

Momentary loudness. Loudness measured over the last 400 ms, for spotting sudden level changes. See also Short-term loudness and Integrated loudness. Integrated vs short-term vs momentary loudness

Monitoring level. The volume at which you listen while mixing. A repeatable level helps you judge balance consistently, and moderate levels reduce ear fatigue. Monitoring level and ear fatigue

Mono. A single channel of audio, or a stereo file in which left and right are summed. Many phones, smart speakers and PA systems play in mono or near-mono. Collapse your mix to mono and listen for lost bass, thin vocals or missing parts. See also Mono compatibility. Mono compatibility

Mono bass. Keeping the low end identical in the left and right channels so it stays strong in mono and on speakers. Try a crossover around 100 Hz and adjust by ear. Mono bass

Mono compatibility. How well a stereo mix still sounds when its channels are summed to mono. Poor compatibility shows as weaker vocals, bass or drums. Mono compatibility

MP3. A widely used lossy audio format. Encoding at too low a bitrate causes audible artefacts, and loud masters may need extra true-peak headroom to avoid clipping after encoding. How lossy codecs affect your master

Mud. A loss of definition, often when several sounds overlap around 150–400 Hz. Mute groups or effects returns in turn to find the source. See why a mix sounds muddy. Frequency masking

Multiband compression. Compression that splits the signal into frequency bands and compresses each separately, for when dynamics problems affect one range, such as a boomy bass. See also Dynamic EQ. Multiband compression

N

Nearfield monitors. Studio speakers designed to be listened to from a short distance, about an arm's length or two, so the room has less influence on what you hear. Home studio acoustics

Next step. An older Sonalix report label for the first action to try. In the current report, start with the first Mix improvements card. Primary Issue and Next Step

Noise gate. A processor that turns down or mutes a signal when it falls below a threshold, used to remove noise or bleed between notes. Set the release so note tails are not chopped.

Null test. Combining two versions of audio with one polarity-inverted. If they are identical, the result is silence. Whatever remains is the difference. It is the standard way to check that exported stems sum to your mix. How to export stems

Nyquist frequency. Half the sample rate: the highest frequency a digital system can represent. At 44.1 kHz it is 22.05 kHz. See also Sample rate.

O

Octave. A doubling of frequency. 100 Hz to 200 Hz is one octave, as is 1 kHz to 2 kHz. EQ widths and filter slopes are often described in octaves.

Ogg Vorbis. An open lossy audio codec that some streaming services have used. See also Lossy codec. How lossy codecs affect your master

Overall verdict. An older Sonalix label for the overview above the measurements. The current report opens with a summary and lists what it measured and what it could assess. Overall verdict

Over-compression. Compression or limiting so heavy that transients flatten, the mix loses punch and movement, and it becomes tiring to hear. Compare with the processing bypassed at matched loudness. See also Crest factor. Signs of over-compression

P

Pan law. The rule your DAW uses to adjust a track's level as you pan it, so it does not sound louder or quieter as it moves. Common options are −3 dB (constant power), −4.5 dB and −6 dB (linear). Leave the setting alone unless you have a reason. Pan laws explained

Panning. Placing a sound between the left and right speakers. Hard-left and hard-right are the extremes. Centre puts it equally in both. See also Pan law. Pan laws explained

Parallel compression. Blending a heavily compressed copy of a signal with the uncompressed original, to add density while keeping transients. Set up with an aux send or a mix knob. Parallel compression

Parametric EQ. An EQ whose bands have adjustable frequency, gain and Q (or bandwidth). It is the standard mixing EQ. See also Bell filter. Types of EQ

Peak. The highest instantaneous level of a signal. See also Sample peak, True peak and RMS. Peak vs RMS

Per-metric severity. An earlier Sonalix rating for a reading, from Good through Caution and Warning to Critical. The current report gives findings an importance of high, medium, low or info. Per-metric severity

Phase. A position within the cycle of a waveform. For equal-frequency signals, the phase difference determines the degree of reinforcement or cancellation when they are added. Cancellation is greatest at 180°. Polarity vs phase

Phase cancellation. Loss of level, usually in certain frequencies, when two similar signals are combined out of phase. It makes tracks sound thin or hollow and is a common cause of mono problems. Flip polarity or nudge timing, and listen. Phase cancellation

PLR. Peak-to-loudness ratio, calculated as true peak minus integrated loudness and expressed in dB. A small PLR describes a dense peak-to-loudness relationship. The reading alone cannot identify the processing. See PSR. PLR and PSR

Polarity. Whether a waveform is "right way up" or inverted. Flipping polarity reverses the sign of every sample without delaying it. The button labelled "phase" in most DAWs is actually a polarity switch. Polarity vs phase

Premaster. The finished stereo mix handed over for mastering, with headroom left for the mastering engineer and exported at the right sample rate and bit depth. Decide deliberately which mix-bus processing stays on. How to prepare a mix for mastering

Pre-ringing. A faint "ring" that precedes a sound's attack, caused by linear-phase filtering, most audible on sharp transients with steep or large EQ moves. See also Linear-phase EQ. Linear-phase EQ

Presence. The part of the upper spectrum associated with clarity and proximity. Sonalix’s earlier five-band display used 4–8 kHz. Its current measurement band is 5–10 kHz and its tonal comparison region is 3.5–8 kHz. 5-band energy

Primary issue. An older Sonalix label for the first finding to investigate. The current report presents its suggested changes in priority order. Primary Issue and Next Step

Priority fixes. Ranked recommendations with a setting to try and a description of what you should hear. The current Sonalix report calls its main section Mix improvements and shows up to four cards. Priority fixes

PSR. Peak-to-short-term loudness ratio. It compares a peak with short-term loudness rather than the whole-track value used by PLR. Check the meter’s peak and time-window conventions when comparing values. PLR and PSR

Pumping. An audible, rhythmic rise and fall in level caused by a compressor or limiter reacting to loud events, such as a kick ducking the rest of the mix. Sometimes wanted, often a sign of over-compression. Lengthen release or reduce gain reduction. See also Sidechain.

Q

Q. The measure of how narrow an EQ band is. A high Q is a narrow, surgical band. A low Q is wide and gentle. Q and bandwidth are inversely related. What is equalisation (EQ)?

Quality tier. A score out of 100 in an earlier Sonalix report. It combined distance from −14 LUFS, true peak, crest factor and stereo correlation. The bands were 75–100, 55–74, 35–54 and 0–34. The current report gives ranked findings without a numeric score. Quality tier

R

Ratio. How strongly a compressor reduces the signal above its threshold. At 4:1, a signal 4 dB over the threshold comes out 1 dB over. Low ratios for gentle control, high ratios for limiting-style behaviour.

Red Book. The standard for audio CDs: 16-bit, 44.1 kHz stereo audio, up to 99 tracks. Do sample-rate conversion before dither, and set track starts and gaps. CD Red Book standard

Reference track. A finished, professionally released track that you compare your mix against for tone, level and width. Route it around your mix-bus processing and match loudness before comparing. How to use reference tracks

Release. On a compressor, limiter or gate, how quickly it lets go after the signal falls back below the threshold. A short release brings the level back fast. A long one is smoother but can hold down the next hit. Dynamic range compression

Resonance. A narrow frequency range that rings or booms, from an instrument, microphone, room or plug-in. Sweep a narrow boost to find it, then cut a little. If it appears in your speakers only, suspect the room. How to find and reduce resonances

Return. A channel that receives the output of an aux send and carries a shared effect, such as reverb. See also Aux send and Send. Bus structure and group processing

Reverb. The persistence of sound after the source stops, from the reflections of a space. In a mix it adds depth and places parts in a shared room. Too much washes out definition and widens the image unpredictably. Use a shared return rather than one reverb per track. See also Aux send.

RMS. Root mean square, a measure of a signal’s average amplitude over a stated interval. RMS gives context that a peak alone cannot provide. A LUFS reading also accounts for frequency weighting and, for integrated loudness, gating. Peak vs RMS

Roll-off. A gradual reduction in level beyond a point on the spectrum, in either direction. Often described by slope, such as 12 dB per octave. See also High-pass filter.

Room modes. Resonances of a room's dimensions that boost or cancel certain bass frequencies at different positions, so notes boom in some places and vanish in others. Move your monitors and listening position before buying treatment. Home studio acoustics

S

Sample peak. The largest absolute value among the stored samples, expressed in dBFS. The reconstructed waveform can peak higher. See True peak and Inter-sample peak. dBFS vs dBTP vs LUFS

Sample rate. The number of samples per second in digital audio. 44.1 kHz and 48 kHz are common. The rate sets the upper frequency limit. See sample rate and Nyquist frequency. Supported file formats, sample rates and file sizes

Sample-rate conversion. Changing a file from one sample rate to another, such as 48 kHz to 44.1 kHz. Do it before dither and bit-depth reduction, and check the converted file for peaks. CD Red Book standard

Saturation. Gentle, harmonically rich distortion, from tape, tubes or plug-ins, that adds warmth or density. See also Distortion and Soft clipping. What is distortion in music production?

Send. A control that routes a copy of a track's signal to another destination, such as a reverb return. The original signal stays in place, unlike an Insert. See also Aux send.

Shelving filter. An EQ band that raises or lowers everything above (high shelf) or below (low shelf) a chosen frequency. Good for broad tonal tilt, such as adding brightness. Types of EQ

Short-term loudness. Loudness measured over the last three seconds, useful for judging a chorus or a section. See also Integrated loudness and Momentary loudness. Integrated vs short-term vs momentary loudness

Sibilance. The sharp "s" and "sh" sounds in speech and singing, which can become piercing on bright monitoring or after heavy compression. See also De-esser. Sibilance and de-essing

Side. The difference between left and right in mid/side encoding, commonly scaled as (L − R) / 2. It is zero for identical channels and can carry panned sounds, stereo reverbs or differences in polarity. Mid/side processing

Sidechain. Using one signal to control a processor acting on another, such as a kick drum triggering compression on the bass so the bass ducks out of its way. How to mix kick and bass

Soft clipping. Rounding off peaks gradually instead of cutting them flat, producing a gentler distortion than hard clipping. It can shave peaks before a limiter and add density. See also Clipping. Clipping in audio

Sound Check. Apple's playback normalisation setting, which adjusts the level of tracks to make them play at a similar loudness. It applies the same gain to a whole album, so level differences between its tracks are kept. See also Album loudness consistency. Mastering for Apple Music and Apple Digital Masters

Spectral centroid. The weighted mean frequency of a spectrum, expressed in Hz. A higher centroid usually indicates more energy towards the top of the spectrum. The frequency that divides spectral energy into equal halves is a different statistic. Spectral centroid

Spectrum. The distribution of a sound's energy across frequency. See also Spectrum analyser. The frequency spectrum

Spectrum analyser. A display of level against frequency, used to see tonal balance, spot resonances and compare mixes. Its slope and resolution settings change how it looks. Read it alongside what you hear. How to read a spectrum analyser

Spotify normalisation. Spotify's playback loudness adjustment. Its Normal setting uses a −14 LUFS integrated reference and recommends maximum true peaks below −1 dBTP (below −2 dBTP if the master is louder than −14 LUFS). It is a playback reference, not a mastering rule. Mastering for Spotify · Streaming loudness targets · Mastering for YouTube for YouTube

Stem. A stereo or mono export of a group of tracks, such as all drums or all vocals, with effects and automation included. Stems should sum to your mix. See also Null test. How to export stems

Stereo. Two channels, left and right, which give the impression of width and placement.

Stereo balance. Whether the left and right sides of a mix have similar level and weight. If the mix leans to one side, rule out monitoring problems before changing tracks. Stereo balance

Stereo correlation. A measure of how closely the left and right waveforms vary together. +1 means proportional in the same polarity, 0 means no measured correlation and −1 means proportional with opposite polarity. Negative values suggest checking the mono sum. Stereo correlation

Stereo widener. A tool that increases perceived width, using mid/side gain, delay or phase processing. Compare in mono for lost level or tone. Stereo wideners

Stereo width. How far apart sounds appear in the left-right field. Wider is not better. Check bass, vocals and mono playback. Stereo width explained

Sub bass. The lowest audible region. Sonalix’s earlier five-band display used 20–100 Hz. The current Sub measurement band is 20–60 Hz, and the tonal Sub-bass region is 20–45 Hz. The frequency spectrum

Sub stereo width. An earlier Sonalix metric describing left-right differences below 100 Hz. The current report checks correlation, balance and the mono sum in the 20–120 Hz low region. See measurement methodology. Sub stereo width

T

Target loudness. The loudness you aim for when finishing a master, usually in LUFS integrated. It depends on where the file will be heard (streaming, broadcast, CD), and on your genre and taste. Measure three reference tracks and compare their LUFS and true peaks. Loudness targets by genre

Threshold. The level at which a compressor, limiter or gate starts to act. On a compressor, lowering the threshold means more of the signal is processed. Dynamic range compression

Tonal balance. The relative level of bass, mids and treble across a mix. Compare with reference tracks at matched loudness, then choose between a broad EQ move and a correction in one range. Tonal balance

Tonal balance summary. A headline and brief explanation of the balance across frequency. The current report also shows how the main sections compare with its selected reference balance. Tonal balance summary

Transient. The brief, sharp start of a sound, such as a drum hit or a plucked string, which gives it definition and punch. Over-compression and limiting blunt transients. Transient shaping

Transient shaper. A processor that raises or lowers the attack or sustain of a sound independently of its overall level, to sharpen drum hits or soften harsh attacks. Transient shaping

Translation. How well a mix preserves its important parts and balance on different systems. A phone can miss the deep sub while retaining the pitch of the bass line. See why mixes fail to translate.

Trim. A gain control at the start of a channel, used to set a track's working level before inserts or faders. See also Clip gain and Gain staging. What is gain staging?

True peak. An estimate of the maximum amplitude of the reconstructed waveform, expressed in dBTP. It can exceed sample peak. Spotify recommends keeping it below −1 dBTP, or below −2 dBTP for masters louder than −14 LUFS. True peak (dBTP)

U

Unity gain. A setting at which a channel or plug-in leaves the level unchanged, input equal to output. A fader at unity (0 dB) adds nothing and removes nothing.

Upward compression. Compression that raises quiet parts toward the threshold, rather than turning loud parts down. It adds density but can raise noise. See also Expansion and Downward compression. Upward compression and expansion

V

Vectorscope. A stereo display that plots the left and right channels, similar to a goniometer, used to check width, balance and phase. See also Goniometer. Goniometers and vectorscopes

Vinyl cutting. Transferring a master to a lacquer for pressing. Side length, bass and sibilance all affect how well a record cuts, so vinyl masters plan for them. Mastering for vinyl

VU meter. Volume unit meter: an average-reading level meter with slow response, which gives a better sense of perceived level than a peak meter does. Many plug-ins are calibrated so that 0 VU corresponds to a level well below 0 dBFS. Check the plug-in's manual. What is gain staging?

W

WAV. A common audio container for mixes and masters. A PCM WAV stores uncompressed samples and can use integer or floating-point encoding. See AIFF and Lossless. Supported file formats, sample rates and file sizes

Z

Zero crossing. A point where a waveform passes through zero. Editing at zero crossings helps avoid clicks when cutting audio.

Terms that are easy to mix up

Pair The difference Read more
Sample peak vs true peak Sample peak is the highest stored sample (dBFS). True peak estimates the highest level of the reconstructed waveform (dBTP) and can be higher. dBFS vs dBTP vs LUFS
dBFS vs LUFS dBFS expresses digital signal level relative to full scale. LUFS expresses programme loudness. Integrated LUFS uses K-weighting and gating. LUFS
Crest factor vs LRA vs PLR Crest factor is peak to RMS. LRA is how much loudness varies between sections. PLR is true peak to integrated loudness. Dynamic range vs loudness range
Polarity vs phase Polarity flip inverts a waveform without delaying it. A time delay produces a phase shift that varies with frequency. Polarity vs phase
Mixing vs mastering Mixing balances the tracks into one stereo file. Mastering prepares that file for release. What is mastering?
Normalisation vs limiting Normalisation changes playback level without altering the file's dynamics. Limiting changes the audio itself by capping peaks. Loudness normalisation
Static EQ vs dynamic EQ Static EQ applies the same boost or cut all the time. Dynamic EQ acts only when the band crosses a threshold. Dynamic EQ
Gating (loudness) vs noise gate Loudness gating excludes quiet blocks from a measurement. A noise gate mutes quiet audio. K-weighting and gating
Mid/side vs left/right Left/right are the two speaker channels. Mid is their sum and side is their difference, with scaling set by the encoder. Mid/side processing

For terms used in the current Sonalix workflow, see your mix report, measurement methodology and metric thresholds.

Was this page helpful?

    Type to search titles, headings and summaries.