Module-4.2-editing

Session 2

Cutting the Voice — Cleaning a Spoken Recording & the Ethics of Editing

Session 2 · Cutting the Voice — A Pleasure to Hear, Still That Person

What You Will Learn

  • How to move a recording from the recorder into a safe, backed-up editing session
  • What to cut almost automatically: fillers, long pauses, false starts and verbal tics
  • How to make the harder cut-or-keep decisions: repetitions, lost threads, hesitations, breaths, clicks and plosives
  • The rhythm test, and why you always listen to the whole sentence after a cut
  • Where the ethical line lies when you edit an older person’s voice — and how consent works in layers
  • Why three good editors produce three different, equally defensible edits

📋 Session info

⏱ Duration: about 2 hours 45 minutes, including a 50-minute cleaning exercise

👥 Group: 10 participants

🌐 Language: English; each participant edits in their own language

🔗 Builds on: Session 1 — file hygiene, first cuts and the recorded interviews

🔧 Hands-on: everyone cleans the first 3 minutes of their own interview

We cut the noise around the message, never the message.

This is the heart of the module: learning to cut a real human voice so that it becomes a pleasure to listen to, without it ceasing to be that human. The craft comes in two passes — the easy removals and the judgement calls — and it ends with the question that matters most in a project with older people: what are we allowed to remove, and who decides?

First, Get Your Recording In Safely

Before anyone cuts a single word, the recording has to be safe. This is where the file rules from Session 1 stop being theory: each participant moves their own interview from the recorder into the folder structure, with a proper name and a backup copy, and only then opens it in the editor.

1 · Connect

The recorder over USB as a card reader, or the memory card in a reader.

2 · Find the file

Take the stereo mix; separate tracks are for later.

3 · Copy to raw

Copy, never move. The original stays on the card for now.

4 · Rename

date_person_topic, no spaces, no national characters.

5 · Back up

A second copy on a pendrive or in the cloud.

6 · Import

A new editing session; save it into the project folder.

Many portable recorders, including the ones used in this course, save a ready stereo mix plus separate tracks for each microphone. For a first edit, the stereo mix is simpler. The separate tracks are the professional option for later: they let you set the level and clean the sound of each person independently.

When the waveform appears on screen, there is a small reward for anyone who used the error-marker sound pad while recording: every mistake they flagged is now a sharp, visible spike on the timeline.

💡 Tip: One simple rule for any group: nobody edits a recording that has not been backed up. Make it a checkpoint before the editing starts.

🎓 Trainer Tip: Ten people queuing at two memory cards will eat half an hour. Copy all the recordings onto a distribution pendrive during the break before this session, and let participants take their file from there — they still perform the copy, rename, back-up and import routine on their own laptops. And when the first waveforms appear, point at the error-marker spikes: "You edited half of this episode while you were still recording it."

✏️ Exercise — click or tap to expand ▼

▶ ✏️ Try This With Your Group

From Recorder to Project

Turning the file rules into a habit with real recordings. Recorders or memory cards (or a distribution pendrive), laptops with the editor, and a backup pendrive.

Step 1 — watch (5 minutes): one or two volunteer pairs go through all six steps at the front, on the projector.

Step 2 — do it (10 minutes): everyone copies their own interview into the raw folder, renames it, makes the backup copy, imports it into a new session and saves.

Step 3 — checkpoint (3 minutes): neighbours check each other: is the file named correctly, does the backup exist, is the session saved in the project folder?

⏱ Time: about 20 minutes

📋 Alternative — Take-home task: Ask everyone to repeat the six steps at home with a phone recording, and send a screenshot of their folder structure before the next meeting.

What to Cut Without Hesitation

Some things almost always go. Naming them as an "automatic" category saves beginners from agonising over easy decisions and leaves their attention for the hard ones.

💬 Filler sounds

"Uhm", "eee" and their equivalents in every language. They carry no meaning, and the listener’s brain filters them out at a cost.

⏸️ Very long pauses

Trim them down, but do not delete them entirely. A pause is part of speech; dead air is not.

🔄 Obvious false starts

The sentence that collapses after four words and starts again. Keep the restart, cut the wreck.

🔁 Verbal tics

Stray, repeated interjections that add nothing — "you know, you know, you know".

Two rules while cutting

🌬️ Leave the breath

Cut the filler, but keep the breath before the sentence. A voice with every breath removed sounds like a robot.

🎧 Listen to the whole sentence

After every cut, play the whole sentence again, not just the join. The splice can be clean and the sentence still broken.

With the editor’s keyboard shortcuts, the rhythm becomes quick: place the playhead, split, delete, then preroll and listen. If the recording has error-marker spikes, start with those — jump to each spike and cut the mistake just before it.

🎓 Trainer Tip: Demonstrate on the projector and think out loud with every single cut: "This is a filler, it goes. This pause is too long — I shorten it, but I keep half a second." Name the keyboard shortcut every time you use it. Repetition installs the habit, and hearing the reasoning teaches more than watching the mouse.

✏️ Exercise — click or tap to expand ▼

▶ ✏️ Try This With Your Group

Count the Fillers

Training the ear before the hands. A one-minute fragment of a spontaneous interview, paper and pens.

Step 1 — listen and count (3 minutes): play the fragment once. Everyone makes a tally mark for every filler sound and every false start they hear.

Step 2 — compare (4 minutes): compare the counts. Why did people hear different numbers? Which fillers were easy to miss?

Step 3 — listen again (3 minutes): play it once more and notice the long pauses: which would you shorten, and which carry meaning?

⏱ Time: about 10 minutes

📋 Alternative — Take-home task: Ask everyone to listen to five minutes of a live radio interview and note the fillers and false starts they hear. Which of them would a podcast editor cut?

The Harder Cases: Cut or Keep

Mechanical cleaning only goes so far. The next category needs editorial judgement, where rhythm and meaning matter more than surgical cleanliness.

🔁 Repeated words

Cut it when it is a stumble. Keep it when it is emphasis: "it was very, very hard". The difference is audible, not grammatical.

🧶 Lost, tangled threads

Cut down to the sense — or keep the thinking-out-loud because it is honest and human. Ask: does the wandering add character or cost attention?

🗣️ Stuttering, hesitation

Trim with tact. Removing every hesitation from a hesitant speaker erases the person.

🌬️ Breaths

Thin them out, never remove them all.

👄 Mouth clicks

Cut them, or lower their volume.

💨 Plosives (p, b)

For now, a small fade or a dip in the clip’s volume. Equalisation, covered in Session 3, is the better tool.

The rhythm test

When unsure, close your eyes and listen to 20 seconds around the cut. If the speech still walks naturally, the cut is good.

🎓 Trainer Tip: After going through the harder cases, ask the group one quick question: "Which of these will be hardest with your own older speakers?" Take two or three short answers. The responses usually point straight at hesitation and slow pace — which is the perfect bridge into the ethics discussion that follows.

🎙️ Audio note — to be recorded and embedded

Topic: Cut or keep — the judgement calls, heard rather than described

Suggested script outline:

  • A repeated word as a stumble, then as emphasis ("very, very hard") — the audible difference
  • A short answer with all breaths removed, then with breaths thinned out — the robot effect
  • A hesitant speaker edited lightly, then over-cleaned — when editing starts to erase the person
  • A closing reminder of the rhythm test: 20 seconds, eyes closed, does it walk?

[Replace this block with a WordPress audio block after recording]

✏️ Exercise — click or tap to expand ▼

▶ ✏️ Try This With Your Group

Stumble or Emphasis?

Practising the judgement calls as a group. Five or six short moments from a spontaneous recording, each containing a repetition, a hesitation or a wandering thought; a laptop and a speaker.

Step 1 — play (5 minutes): play each moment twice.

Step 2 — vote (5 minutes): for each moment, everyone shows a card: cut or keep.

Step 3 — argue (5 minutes): where the votes split, ask one person from each side to explain. Was the argument about rhythm, about meaning, or about the person?

⏱ Time: about 15 minutes

📋 Alternative — Take-home task: Ask everyone to find three repeated words in their own recording and decide for each one: stumble or emphasis? Bring the decisions, with a one-sentence reason each, to the next meeting.

The Ethics of Editing an Older Person’s Voice

Technically, you now know how to remove almost anything from a recording. The question of this part is different: what are you allowed to remove, and who decides? It is the part that makes this module distinctly about the voices of older people.

The purpose of editing is clarity, not creation

The meaning of what a person said is untouchable. We cut the noise around the message, never the message.

Principle 1

Over-smoothing is erasure

The way a person speaks — their pace, their hesitations, their repeated phrases — is part of who they are. An 80-year-old voice edited into a newsreader is no longer a testimony; it is a costume.

Principle 2

Consent has layers

Consent to be recorded, consent to be edited, consent to be published. Each is a separate question.

Principle 3

The stakes are real

Older people are often asked to trust a microphone for the first time. One manipulative edit, even a well-meant one, can end that trust for good.

The best practice is simple: offer the speaker the edited version to hear before it is published. It costs one e-mail or one visit, and it builds the trust that a community podcast depends on.

What if the speaker said something factually wrong? You may cut the error with the speaker’s agreement, or add a correction around it — a comment from the host or a note in the episode description. What you may not do is silently re-edit their words into a different claim.

💡 Tip: Ask for each layer of consent separately, and write down the answers. A short note — "agreed to recording, editing and publication on this date" — saves misunderstandings later.

The Voice of Seniors rule

If a cut changes what the person would say they said, the cut is off limits.

🎓 Trainer Tip: The most effective way to teach this is a blind comparison. Record a 60–90 second statement with a qualifier in it — for example "I enjoyed it, although at first I was sceptical, because…" — and prepare two edits: one honest, one where the qualifier is removed or two sentences are swapped so the emphasis flips. Keep a written note of exactly which cuts differ. Play both without saying which is which, and let the group find the manipulation themselves.

✏️ Exercise — click or tap to expand ▼

▶ ✏️ Try This With Your Group

Two Edits, One Statement

Finding the line between editing for clarity and editing that changes meaning. Two prepared edits of the same short statement — one honest, one subtly manipulative — and a speaker.

Step 1 — listen blind (4 minutes): play both versions without saying which is which.

Step 2 — find the change (8 minutes): ask: what changed between the two? Which version would the speaker sign their name under? Let the group identify the exact cuts that shifted the meaning.

Step 3 — state the rule (3 minutes): together, put the line into one sentence. Compare it with the Voice of Seniors rule above.

⏱ Time: about 15 minutes

📋 Alternative — Take-home task: Ask everyone to write down a sentence they once said that could be cut to mean the opposite — and how. Share the examples at the next meeting; they make the risk very concrete.

Cleaning a Real Voice

The long practice block brings everything together. It starts with a shared warm-up on one common clip, then moves to each participant’s own interview, in their own language. The target for a first session is realistic: the first 3 minutes cleaned, fillers and false starts out, pauses trimmed, rhythm intact, meaning untouched.

There is no single correct edit

Three good editors, three different edits, all defensible. There are decisions and their consequences, not one truth.

Four editing types — and what helps each one

🖱️ The mouse-editor

Does everything with the pointer. Prescription: park one hand on the trim and split keys, and watch the speed double.

✂️ The over-cutter

Everything sterile, rhythm destroyed. Prescription: the 20-second, eyes-closed listen.

😶 The under-cutter

Afraid to touch anything. Prescription: show again that dragging the edge brings everything back — courage is free.

📍 The marker-user

Recorded with the error-marker pad. Prescription: chase the spikes first — it is the fastest route through your own mistakes.

In an international group, the trainers will not understand most of the languages being edited — and that is fine. Ask the participant to explain a cut-or-keep decision. The reasoning about rhythm, meaning and the person transfers across every language.

💡 Tip: Note how many seconds you removed from the first 3 minutes. Comparing that number with others is a quick way to spot over- and under-cutting.

🎓 Trainer Tip: In the warm-up, put two or three results of the same fragment side by side on the projector and the speaker. Ask how many seconds each person removed, which cuts the versions share and where people decided differently. Then say the punchline out loud: there is no single correct version. Learners who were anxious about "doing it wrong" relax visibly — and edit better for it.

✏️ Exercise — click or tap to expand ▼

▶ ✏️ Try This With Your Group

Cleaning a Real Voice

Applying both the automatic removals and the judgement calls. A shared practice clip, each participant’s own interview in an editing session, headphones, the Cleaning Checklist and the Shortcut Card.

Step 1 — shared warm-up (15 minutes): everyone cleans the same 1–2 minute fragment. Compare two or three results: seconds removed, shared cuts, different decisions.

Step 2 — your own voice (30 minutes): clean the first 3 minutes of your own interview. Chase the error-marker spikes first, then go through the checklist: automatic removals, then judgement calls.

Step 3 — the rhythm test (5 minutes): close your eyes and listen to the whole 3 minutes. Put back anything that made the speech stop walking.

⏱ Time: about 50 minutes

📋 Alternative — Take-home task: Ask everyone to clean the next 3 minutes of their interview at home with the Cleaning Checklist, then let the interviewee (or a friend) listen and say whether it still sounds like them.

Closing Round: One Cut-or-Keep Decision

A good way to end the session is a full round in which each participant describes one decision they made in their own material, and why — 30 to 45 seconds per person. It turns individual work into shared learning and shows how different the reasons can be.

🎵 Rhythm-based

"It sounded broken after the cut, so I put it back."

💡 Meaning-based

"Cutting that clause changed what she meant."

🤝 Person-based

"Without his pauses, it no longer sounded like him."

🎓 Trainer Tip: Listen for these three kinds of reasoning and highlight them out loud when they appear — especially echoes of the ethics discussion. Then close with a preview of the next session: why one podcast is a pleasure to listen to and another a chore, and how to hear the difference, name it and fix it.

✏️ Exercise — click or tap to expand ▼

▶ ✏️ Try This With Your Group

One Decision Each

Sharing editorial reasoning across the group. Everyone’s cleaned recording and a timer.

Step 1 — choose (2 minutes): everyone picks one moment where they hesitated between cutting and keeping.

Step 2 — the round (30–45 seconds each): describe the moment, what you decided and why. Volunteers may play the moment.

Step 3 — sort (3 minutes): as a group, sort the reasons heard into rhythm, meaning and person.

⏱ Time: about 15 minutes for 10 participants

📋 Alternative — Take-home task: Ask everyone to write a short note on one difficult cut-or-keep decision from their recording and send it to the group before the next meeting.

Key Takeaways

  • Safe first, then edit. Copy, rename, back up, import, save — nobody edits a recording without a backup.
  • Two kinds of cutting. Automatic removals — fillers, long pauses, false starts, tics — and judgement calls that need your ear.
  • Keep the life. Trim pauses rather than deleting them, thin out breaths rather than removing them all, and keep repetitions that are emphasis.
  • Listen to the whole sentence. A join can be clean and the sentence still broken.
  • The rhythm test. Close your eyes, listen to 20 seconds, and decide by how the speech walks.
  • The ethics line. If a cut changes what the person would say they said, it is off limits. Consent comes in layers, and the speaker should hear the edit before publication.
  • No single correct edit. Good editors differ — defensibly.

Resources & Downloads

A printable one-page guide to keep next to the laptop while cleaning any spoken recording. The File Hygiene sheet and the Shortcut Card from Session 1 are useful here too.