Notes · Music TechnologyUK · A-Levels

Audio editing

Once audio is recorded, editing shapes and repairs it: assembling the best performance, cutting and joining regions without clicks, removing clicks, hiss, hum and plosives, and setting levels by normalising and clip gain. This topic explains why discontinuities cause clicks and how fades and zero-crossing cuts avoid them, the tools for cleaning up common noises, the difference between normalising and compression, and the non-destructive workflow that makes all of it safe and reversible.

5 sections~18 min reading time3 competenciesLevel Foundation 1 · Standard 3 · Advanced 1

T·0666 / 16

Exam profile

C4 · Edit audio non-destructively: cut, fade, cross-fade and comp without clicksC4 · Remove clicks, hiss, hum and plosives and set levels by normalising and clip gainC1 · Explain why discontinuities cause clicks and choose edit points and fades to avoid them

Operators:describeexplainidentifycalculatejustify

basic level

The AS foundation is non-destructive editing, fades and cross-fades, and basic noise removal.

higher level

The full A-level adds the reason discontinuities click, targeted de-hum and de-noise, and the distinction between normalising and compression.

Depth

Text

Contents · 5 sections

Audio editing

Non-destructive editing and the timeline#

FoundationPearson 9MT0 - audio editing

Non-destructive editing operations

Fig. 1Editing is a set of reversible operations on regions over an untouched file: arranging, the aids that keep edits clean, and the repair tools of the next sections.

Editing in a DAW is non-destructive: the original recorded audio file is never altered, and every edit - a cut, a fade, a gain change, a piece of processing - is stored as an instruction that the DAW applies on playback. This means any edit can be undone, revised or removed without damaging the recording, so the editor is free to experiment. It is a fundamental break from the tape era, where cutting the tape with a razor blade was literally destructive and irreversible.

Audio is edited as regions (or clips) on a timeline: rectangular blocks representing pieces of the recording that can be moved, trimmed at either end, split, copied, duplicated and reordered freely. Trimming a region's start or end hides but does not delete the audio beyond the boundary, so a trim can always be dragged back out. This non-linear freedom - rearranging time itself - is what makes arranging, tightening and comping practical.

Several aids keep edits clean and in time. Snapping locks region edges to the bar/beat grid (or to zero-crossings) so edits land musically; ripple editing shifts everything after an edit to close a gap; grouping links regions across tracks (all the drum mics) so they are cut and moved together, preserving their phase relationships. Knowing when to snap to the grid and when to work freely is part of editing skill.

Because edits are non-destructive, they are also recallable and comparable: the DAW keeps a history, alternative versions (playlists or comp lanes) can be kept side by side, and the whole edit is saved with the project. This encourages a workflow of trying options and keeping the best, rather than committing irreversibly. The only time editing becomes destructive is when you deliberately render or bounce a processed file, which is done consciously and usually keeps the original too.

For the exam, understand and be able to describe non-destructive, non-linear editing: edits as reversible instructions over an untouched file, audio handled as trimmable, movable regions on a timeline, with snapping, ripple and grouping as aids. This model underlies every later editing operation, from a single clean cut to assembling a comp from a dozen takes.

Worked example

Why group a multitrack edit

A chorus in a multitrack drum recording needs to be moved earlier by one bar. Explain why the drum tracks must be grouped before the move.

The phase relationship
All the drum microphones captured the same kit at fixed relative timings; that alignment is what makes them sum coherently.
Move them together
Grouping the tracks means the one-bar move is applied identically to every drum track, preserving the sample-accurate phase relationship between them.
If ungrouped
Moving the tracks separately, even by a sample or two of difference, would misalign the microphones and cause comb filtering and a hollowed, phasey drum sound.
Confirm
With the tracks grouped, the moved chorus sounds identical to the original - just earlier - because their relative timing is untouched.

Result: Group the drum tracks so the move preserves their phase alignment; editing them separately would misalign the microphones and comb-filter the kit.

Typical mistakes

Believing trimming or cutting a region deletes the audio - it is hidden and fully recoverable until you render.
Editing grouped multitrack recordings without grouping them, so the tracks slip out of their phase alignment.

Active revision

Explain why editing a multitrack drum recording with the tracks grouped is essential, and what would go wrong if they were edited separately.

Active recall

Recall the key points — then reveal.

Sources: Pearson Edexcel Level 3 Advanced GCE in Music Technology (9MT0) Specification (Pearson Edexcel)

Cuts, fades and cross-fades: avoiding clicks#

StandardPearson 9MT0 - audio editing

Safe and unsafe cut points on a waveform

Fig. 2A cut at a zero-crossing leaves no step in level and no click; a cut mid-cycle, at a high value, leaves a discontinuity that clicks. Short fades achieve the same safety anywhere.

When two pieces of audio are joined, or a region starts or stops abruptly, the waveform can jump instantaneously from one level to another - a discontinuity. That sudden step contains a burst of high-frequency energy and is heard as a click or pop. Understanding why clicks happen, and how to prevent them, is essential to clean editing, because a single click at an edit point can ruin an otherwise perfect take.

The safest place to cut is at a zero-crossing - a point where the waveform passes through the zero (silence) line. If both sides of a join are at zero, there is no sudden step in level and no click. DAWs can snap edits to zero-crossings for exactly this reason. Cutting in the middle of a cycle, where the wave is at a high positive or negative value, leaves a step from that value to whatever the next region starts at, producing a click.

A more general solution is the fade. A fade-in ramps the level up from zero at the start of a region, and a fade-out ramps it down to zero at the end, so however the audio behaves internally, the region begins and ends at silence with no discontinuity. Even a very short fade of a few milliseconds removes a click, which is why applying tiny fades to the ends of every edited region is routine housekeeping.

To join two regions seamlessly, a cross-fade overlaps them, fading one out as the other fades in, so there is never an abrupt jump between them. Cross-fades hide edits between takes, smooth loop points and blend regions of different tone or level. The cross-fade's length and shape (linear, equal-power) are chosen to suit the material: a short cross-fade for a tight rhythmic edit, a longer equal-power one for a smooth blend of sustained sound.

For the exam, be able to explain that a click is caused by a waveform discontinuity (a sudden step in level) and that the remedies are to edit at a zero-crossing, to apply short fades so regions begin and end at zero, and to cross-fade joins so there is no abrupt jump. This is one of the most practical and frequently tested ideas in editing.

Worked example

Silencing a clicking edit

Two spoken-word regions are butted together and the join clicks audibly. Explain the cause and give a reliable fix.

Diagnose
The join clicks because the first region ends at a non-zero level and the second begins at a different non-zero level, leaving a sudden step - a discontinuity - at the boundary.
Option 1 - zero-crossings
Nudge each region edge to the nearest zero-crossing so both sides meet at the zero line, removing the step.
Option 2 - cross-fade
Overlap the two regions and apply a short cross-fade so the first fades out as the second fades in, guaranteeing no abrupt jump regardless of the levels.
Best practice
A short cross-fade is the most reliable general fix and also smooths any tonal difference between the two takes.

Result: The click is a discontinuity at the join; fix it by editing to zero-crossings or, more reliably, by applying a short cross-fade.

Typical mistakes

Cutting audio in the middle of a cycle (away from a zero-crossing) and being surprised by a click at the edit.
Confusing a fade (a level ramp on one region) with a cross-fade (an overlap of two regions fading in and out).

Active revision

Explain why joining two vocal takes with a hard cut sometimes clicks, and give two ways to make the join silent.

Active recall

Recall the key points — then reveal.

Sources: Pearson Edexcel Level 3 Advanced GCE in Music Technology (9MT0) Specification (Pearson Edexcel)

Cleaning up: clicks, hiss, hum and plosives#

StandardPearson 9MT0 - audio editing

Mains hum frequencies

Fig. 3Mains hum sits at the mains frequency and its harmonics: 50, 100, 150 Hz in the UK (60, 120, 180 Hz in North America). Narrow notches at these exact frequencies remove it.

Recordings pick up several characteristic noises, and each has a targeted remedy. Isolated clicks and pops (digital glitches, mouth clicks, edit discontinuities) are removed with a de-click tool that detects the sudden spike and interpolates across it, or by hand-drawing over the offending samples in the waveform. Because a click is a brief, broadband spike, it can be repaired without affecting the surrounding audio.

Broadband hiss - the steady 'shhh' of self-noise, tape or a hot pre-amp - sits across the whole spectrum and is reduced with a de-noise (noise-reduction) tool. Such tools take a 'noise print' from a silent section, learn the hiss's spectral shape, and subtract it from the whole recording. The trade-off is that aggressive noise reduction introduces artefacts (a watery, phasey quality), so it is applied only as much as needed.

Mains hum is a tonal, not broadband, noise: a steady low tone at the mains frequency - 50 Hz in the UK and much of the world, 60 Hz in North America - together with its harmonics (100/150 Hz or 120/180 Hz). Because it lives at precise, known frequencies, it is removed with narrow notch (band-reject) filters tuned to the mains frequency and its harmonics, or a dedicated de-hum tool. Targeting the exact frequencies removes the hum while leaving the music almost untouched.

Plosives - the thumps of 'p' and 'b' sounds hitting a microphone - and low-frequency rumble (traffic, footsteps, stand vibration) are low-frequency problems removed with a high-pass filter that cuts everything below the voice's useful range (often around 80-120 Hz for a voice). A pop shield prevents plosives at source; a high-pass filter cleans up any that got through and any rumble, without thinning the wanted sound if the cutoff is chosen sensibly.

The disciplined approach is to identify the noise type and apply the matching tool: de-click for spikes, de-noise for broadband hiss (sparingly), notch filters at 50/60 Hz and harmonics for mains hum, and a high-pass filter for plosives and rumble. Fixing noise at source (good gain structure, balanced cables, pop shields, quiet rooms) is always better, but targeted cleanup is a core exam and Component 4 skill.

Worked example

Targeted cleanup of a vocal

A UK vocal recording has a constant low hum and a few plosive thumps. Describe the tools and settings to clean it.

Identify the hum
A constant low tone in the UK is 50 Hz mains hum, with harmonics at 100 and 150 Hz.
Notch the hum
Apply narrow notch filters at 50, 100 and 150 Hz (or a de-hum tool set to a 50 Hz base), removing the tonal hum while leaving the voice untouched.
Tackle the plosives
Apply a high-pass filter around 80-100 Hz to remove the low-frequency plosive thumps and any rumble, keeping the cutoff below the voice's body so it is not thinned.
Check
Solo the vocal: the hum and thumps are gone and the voice is unchanged, because each tool targeted only the offending frequencies.

Result: Notch 50/100/150 Hz for the mains hum and high-pass around 80-100 Hz for the plosives - each noise met with its matching, frequency-targeted tool.

Typical mistakes

Using broadband noise reduction to remove mains hum - hum is tonal and better removed with narrow notches at 50/60 Hz and its harmonics.
Setting a high-pass filter cutoff too high, thinning the wanted low frequencies along with the rumble.

Active revision

A vocal recording has a steady low hum and occasional 'p' pops. Identify the two noises and state the specific tool and settings for each.

Active recall

Recall the key points — then reveal.

Sources: Pearson Edexcel Level 3 Advanced GCE in Music Technology (9MT0) Specification (Pearson Edexcel)

Level editing: normalising and clip gain#

StandardPearson 9MT0 - audio editing

Peak normalising to a target

Fig. 4Peak normalising applies one fixed gain to bring the loudest peak to a target (here -1 dBFS). The whole signal moves up together, so the dynamics are unchanged.

Normalising raises (or lowers) the level of a whole file or region by a single fixed amount so that a chosen measure hits a target. Peak normalising finds the loudest peak and applies the one gain change that brings that peak to a set ceiling - commonly around -1 dBFS - so the file is as loud as possible without clipping. Because it applies the same gain everywhere, normalising changes the overall level but not the internal dynamics: the quiet parts and loud parts keep their relationship exactly.

This is the crucial distinction from compression. Compression reduces the difference between loud and quiet parts, changing the dynamics; normalising simply scales the whole signal by one number, leaving the dynamics untouched. Normalising a quiet recording makes it louder overall but does not make quiet details easier to hear relative to loud ones - only compression does that. Confusing the two is a common and heavily penalised error.

There are two kinds of normalising. Peak normalising targets the peak level and guarantees no clipping, but two files peak-normalised to the same level can still sound very different in loudness because of their different crest factors. Loudness normalising targets a perceptual loudness measure (LUFS) instead, so files match in perceived loudness - the method streaming services use to make tracks play back at a consistent volume. Choosing the right target depends on whether you care about peak level or perceived loudness.

Clip gain (or region gain) is a level change applied to an individual region before it reaches the channel fader, and it is the editor's tool for evening out a performance. Riding clip gain to lift a too-quiet word or pull down an over-loud phrase tidies the dynamics manually before any compressor, so the compressor then works on a more consistent signal. Because it is non-destructive region gain, it can be drawn in freely and revised.

For the exam, be able to state that normalising applies one fixed gain to hit a peak or loudness target and does not change dynamics (unlike compression), to distinguish peak from loudness normalising, and to explain clip gain as per-region level editing. A typical calculation is to find the gain a normalise applies: it is simply the target level minus the current peak level, in decibels.

Worked example

Gain applied by a normalise

A file peaks at -9 dBFS and is peak-normalised to -1 dBFS. Find the gain applied and state the effect on the recording's dynamics.

Gain applied
Normalising moves the peak from -9 to -1 dBFS, a change of target minus current peak = (-1) - (-9) = +8 dB, applied to the whole file.
Effect on dynamics
Because the same +8 dB is applied everywhere, every part rises by 8 dB and the quiet-to-loud relationships are unchanged - the dynamic range is identical.
$G = (-1) - (-9) = +8\,\text{dB}$
Contrast
To actually reduce the gap between quiet and loud you would need compression; normalising only makes the whole thing 8 dB louder.

Result: The normalise applies +8 dB to the whole file; its dynamics are unchanged, because normalising scales, it does not compress.

Typical mistakes

Saying normalising reduces dynamic range - it applies one gain to the whole signal and leaves the dynamics unchanged; compression changes dynamics.
Assuming two peak-normalised files sound equally loud - crest factor differs, so use loudness (LUFS) normalising for matched loudness.

Active revision

A recording peaks at -9 dBFS and you peak-normalise it to -1 dBFS. Calculate the gain applied, and state whether the quiet-to-loud balance of the recording changes.

Active recall

Recall the key points — then reveal.

Sources: Pearson Edexcel Level 3 Advanced GCE in Music Technology (9MT0) Specification (Pearson Edexcel)

Comping, timing edits and consolidation#

AdvancedPearson 9MT0 - audio editing

The comping workflow

Fig. 5Comping: the best sections of several takes are chosen, cross-faded into one comp track, then consolidated to a single clean file.

Comping (compositing) assembles one ideal performance from several recorded takes by choosing the best phrase, line or even word from each. The performer records the part several times into take lanes; the editor auditions the takes, selects the strongest section of each, and swaps them into a single 'comp' track, cross-fading the joins so they are inaudible. Comping is standard practice for lead vocals and solos, where a flawless whole is built from imperfect parts.

Successful comping depends on the editing discipline of the earlier sections: joins are made at sensible points (breaths, gaps, zero-crossings) and cross-faded so no click or level jump betrays the edit, and the tone and timing of adjacent takes are matched so the comp sounds like one continuous performance. A comp that jumps in tone or clicks at every join is worse than a single honest take, so the craft is in making the seams disappear.

Timing edits tighten a performance to the groove. A late or early note can be cut out and nudged onto the beat, or a whole phrase slipped into time; drum hits can be moved to the grid; and, with more advanced tools, a performance can be sliced at its transients and quantised (as in the pitch-and-rhythm topic). The goal is a performance that feels tight without becoming mechanical, so timing edits are used with judgement, preserving natural feel where it serves the music.

When the editing is complete, regions are often consolidated (or bounced/rendered): a series of edited fragments and their fades are combined into a single continuous audio file. Consolidation tidies the session, guarantees the edits and fades are baked in, makes the track portable to another system, and frees the DAW from constantly recalculating many tiny edits. It is done consciously, usually keeping the original takes archived in case a revision is needed.

For the exam, be able to describe comping (building a best performance from take lanes with inaudible cross-faded joins), timing editing (nudging or slicing to tighten the groove with judgement), and consolidation (rendering edited fragments to one clean file). These finishing operations turn raw multi-take recordings into the polished performances that a mix is built on, and they draw on every editing idea in the topic.

Worked example

Comping a lead vocal

Four takes of a lead vocal are recorded. Outline how to build one seamless comp and keep the joins inaudible.

Audition and mark
Play the takes in their lanes and mark the strongest phrase, line or word in each - for example the best first verse from take 2 and the best chorus from take 4.
Swap in the best
Copy each chosen section into the comp track, keeping the timing aligned across takes so nothing lands off the beat.
Hide the joins
Place edits in breaths or gaps, snap to zero-crossings, and cross-fade each join so there is no click or level jump; match the tone of adjacent takes if needed.
Consolidate
When the comp is complete and sounds like one performance, consolidate it to a single file, keeping the original takes archived in case of revision.

Result: Select the best sections from the take lanes, cross-fade the joins in breaths at zero-crossings, and consolidate - a single seamless vocal from four imperfect takes.

Typical mistakes

Comping with hard, unmatched joins that click or jump in tone, drawing attention to the edits.
Over-quantising timing until the performance sounds rigid and lifeless.

Active revision

Outline the steps to comp a lead vocal from four takes into one seamless performance, naming the techniques that keep the joins inaudible.

Active recall

Recall the key points — then reveal.

Sources: Pearson Edexcel Level 3 Advanced GCE in Music Technology (9MT0) Specification (Pearson Edexcel) · Ofqual - GCE AS and A level qualifications (subject-level conditions and requirements) (Ofqual)

References & sources

Sources

Pearson Edexcel

Pearson Edexcel Level 3 Advanced GCE in Music Technology (9MT0) Specification

Ofqual

Ofqual - GCE AS and A level qualifications (subject-level conditions and requirements)

Audio editing

5 sections~18 min reading time3 competenciesLevel Foundation 1 · Standard 3 · Advanced 1

Audio editing

Key points

Why group a multitrack edit

Key points

Silencing a clicking edit

Key points

Targeted cleanup of a vocal

Key points

Gain applied by a normalise

Key points

Comping a lead vocal

Sources

Audio editing

Key points

Why group a multitrack edit

Key points

Silencing a clicking edit

Key points

Targeted cleanup of a vocal

Key points

Gain applied by a normalise

Key points

Comping a lead vocal

Sources