A Practical Guide to Interview and Talking-Head Editing

Interview and talking-head editing is the craft of turning a long, recorded conversation into a short piece that sounds natural, holds attention and still says exactly what the speaker meant. It covers choosing which answers survive, trimming them to length, hiding the places where two parts of a take have been joined, and smoothing the sound so the audience never notices the surgery. It is the backbone of corporate video: leadership messages, customer testimonials, recruitment films, explainer pieces with an expert on camera and most documentary-style brand content.

It matters to two groups. Marketers and business owners commission this format more than any other, and they need to know what to ask for at the shoot so the edit is not boxed in later. Editors and producers need a dependable method, because interview editing is also where editing ethics bite hardest: the same tools that tighten a rambling answer can quietly change what it said. A cut that removes an "in most cases" or stitches the start of one answer onto the end of another can put words in someone's mouth, and that is a legal, reputational and personal problem, not just a craft one.

Rather than list techniques in the abstract, this guide follows one project from brief to delivery and explains each decision as it comes up. The project is illustrative: a composite built to show a realistic workflow, not a real client, and its numbers are plausible planning figures rather than measured results. Where a choice depends on budget, footage or approval rules, the trade-offs are spelled out so you can make the same call on your own work.

The Illustrative Brief: A Three-Minute Leadership Interview With a Sixty-Second Cutdown

The client in our example is a mid-sized engineering services firm launching a new maintenance program for industrial customers. The brief asks for a three-minute interview film featuring the firm's head of operations, explaining why the program exists, how it works and what customers should expect in the first year. Alongside the main film, the marketing team wants a sixty-second cutdown for social channels and two or three standalone clips of fifteen to thirty seconds each for paid placements.

Three constraints shape everything that follows. First, the speaker is an operations leader, not a trained presenter, so answers will be long, qualified and occasionally circular. Second, the firm's legal team must approve every claim about service levels, which means the edit cannot soften or harden a promise. Third, the head of operations has asked to review the final cut before release, a common and reasonable request that turns the meaning of every answer into a sign-off item.

What "done" meant before anyone picked up a camera

The producer wrote a one-page definition of done and circulated it before the shoot. It listed the deliverables (main film, cutdown, clips), the aspect ratios (16:9 for the main film, 9:16 and 1:1 for social), the review rounds (two for the client team, one for the speaker, one for legal), and a single editorial rule: no answer in any version may be joined to a different answer to form a sentence the speaker did not say. Writing that rule down early sounds fussy. In practice it saved an argument in week two, when a stakeholder asked for exactly that kind of splice.

If your team produces this format regularly, it is worth reading how a dedicated interview video service scopes deliverables, because most budget overruns in interview work come from cutdowns and approval rounds that nobody priced at the start.

Planning Coverage So the Edit Has Somewhere to Hide Its Joins

Most problems that surface in an interview edit were decided on the shoot day. The single most important question at the planning stage is simple: when the editor removes three seconds from the middle of an answer, what will the audience see? If the answer is "the same frame, with the speaker's head jumping slightly," you have a jump cut, a visible discontinuity created when two parts of one shot are joined. Sometimes a jump cut is a deliberate style, common in fast social content and vlogs. For a considered leadership film, it tends to read as clumsy or, worse, as evidence of heavy editing.

There are three standard ways to conceal a join: cut to a cutaway (a shot of the interviewer listening, the speaker's hands, a detail of the room), cut to B-roll (footage of the subject being discussed), or cut to a second camera angle. Each one hides the edit because the audience sees a new image at the precise moment the audio is joined, and the brain accepts the change of shot as the reason for any change in position.

Why the producer booked two cameras

For our illustrative project the producer booked two cameras. The A camera sat just off the interviewer's shoulder in a medium close-up; the B camera sat roughly thirty degrees further round on the same side of the eyeline, framed tighter. The two-camera advantage is that almost any join can be hidden by switching angles at the cut, without leaning on B-roll to cover every edit. That leaves B-roll free to do its real job, which is to show the audience what the speaker is describing.

Keeping both cameras on the same side of the eyeline matters. If the B camera crosses the line, the speaker appears to look left in one angle and right in the other, and every angle change becomes jarring. The logic is the same as in drama coverage, and our guide to continuity editing and the 180-degree rule explains why the line exists and how to stage around it.

Lenses, light and sound set up for the edit

The angles only cut together well if they look different enough to justify a change and similar enough to feel like the same moment. A useful rule of thumb is to vary shot size substantially between the two cameras and to keep the lighting consistent across both, so the color grade does not have to fight skin tones that shift at every cut. The team followed the approaches in our guides to interview lighting and lens choice in video, using a longer focal length on the tighter B camera to keep perspective flattering and the background soft.

Sound was planned just as deliberately. The speaker wore a lavalier microphone and a boom microphone was positioned overhead, just out of frame. Two sources give the editor a choice at every join: if the lav rustles against clothing during an answer, the boom can cover that phrase. The sound recordist also listed the noise sources in the room (a ventilation unit, a distant forklift alarm) so they could be switched off or at least documented.

On the Shoot Day: Room Tone, Second Angles and B-Roll Captured for the Cut

The shoot ran two sessions of roughly forty minutes each, with a break between them. The interviewer asked open questions and, crucially, asked the speaker to answer in full sentences that included the question's subject. "Why did you launch the program?" answered with "Because customers kept telling us..." is hard to use without the question. "We launched the maintenance program because customers kept telling us..." stands on its own. This one habit does more for the edit than almost any technical choice.

Recording room tone before the kit was packed

Room tone is recorded silence from the location: the particular hum, air movement and distant noise of that room with everyone in position and nobody speaking. The editor uses it to fill gaps and smooth audio joins, so the background does not drop to dead silence or jump in character at each cut. The sound recordist called for sixty seconds of room tone after each session, with the speaker still seated and microphones in the same positions, because moving people or kit changes the tone.

It is tempting to skip this at the end of a long day. It takes a minute, the room is already set, and there is no second chance once the equipment is struck. Room tone recorded on a different day, or pulled from another location, rarely matches.

B-roll shot to a list, not "some nice footage"

The B-roll list was built from the question plan. Every topic the speaker was likely to mention had matching coverage: technicians inspecting equipment, a scheduling screen, a service van arriving, a handover conversation between engineers. Each shot ran long enough to be usable (at least ten seconds of steady footage, with movement starting and ending cleanly) and each was shot in both wide and detail sizes. The camera team also gathered generic but relevant cutaways: hands on a tablet, a close shot of a gauge, the building exterior. Those are the pieces an editor reaches for when a trim needs covering and nothing more specific fits.

After the interview, the operator recorded two minutes of the speaker listening and nodding, and a slow push-in on their hands. These "noddies" are old-fashioned but still useful, provided they are used honestly: a listening shot should never be used to imply a reaction to something the speaker did not hear.

From Ingest to Paper Edit: Building the Film From a Transcript

With footage in the building, the editor ingested both cameras and both audio sources, synced them by timecode with waveform matching as a fallback, and created a multicam clip for each session. The approach follows the principles in our guide to multi-camera editing: sync once, check sync at several points through each session, and then edit a single grouped clip rather than juggling separate files.

Why the transcript came before the timeline

Before any cutting, the audio was transcribed. The good-practice rule here is to build the piece from a transcript before touching the timeline, and it pays for itself many times over. Reading is far faster than watching. A transcript lets the producer, the client and the editor all mark selects in the same document, and it makes structure visible: you can see which answers repeat each other, which ones contradict, and where the strongest sentences sit.

Modern editing software makes this easier than it used to be. Premiere Pro, for example, can generate a transcript and let an editor assemble a sequence by selecting text, and the Adobe Help Center documentation on text-based editing describes the workflow in detail. Our own article on text-based editing and the decisions that matter covers when to trust the automatic transcript and when to correct it by hand. The short version: automatic transcripts are good enough for finding material, but proper nouns, technical terms and any sentence you intend to quote in captions or on-screen text must be checked against the audio.

The paper edit itself

The producer and editor then built a paper edit: a document that lists, in order, each selected passage with its timecode, the words spoken, and a note on the picture intended to cover it. The paper edit for the main film was organized under four beats: the problem customers faced, why the firm built the program, how it works in practice, and what the first year looks like. Every selected passage was marked with its source answer, so anyone reviewing could see at a glance that no sentence had been built from two different answers.

Those figures are illustrative, but the ratios are typical of the format. A relaxed speaker often lands somewhere around 130 to 160 words per minute, so eighty-odd minutes of talk produces a transcript in the ten-thousand-word range, and a three-minute film has room for only a few hundred of those words. That is why selection, not trimming, is where most of the editorial work lives.

Selecting the Answers: How Three Minutes Were Chosen From Eighty

Selection is a story decision before it is a technical one. The editor read the full transcript once without marking anything, then again with three highlighter colors: one for statements that carried the argument, one for vivid specifics (a named type of equipment, a concrete example of a failure the program prevents), and one for emotional moments where the speaker's conviction came through. The strongest passages usually carry two colors at once.

Criteria that decided each select

• Does it stand alone? Answers that depend on the question, or on something said five minutes earlier, need setup the film has no time for.
• Is it the clearest version? Speakers often make the same point three times in a session. The editor picked the most direct telling, even when a later version was more polished, if the earlier one sounded more natural.
• Is it accurate as spoken? Anything that sounded like a service commitment was flagged for legal review immediately, before the cut was built around it.
• Does the picture exist? A great line about a customer site the team did not film is harder to use than a good line about the operations floor that has B-roll.
• Can it be trimmed at a natural boundary? Passages with clean pauses at the start and end cut far more easily than those that run straight into the next thought.

Structure for the main film and the cutdown

The four-beat structure from the paper edit held for the main film. For the sixty-second cutdown, the team did not simply trim the main film proportionally. Instead they asked which single idea the social audience most needed, chose the "what the first year looks like" beat, and built the cutdown around two answers from that section plus a one-sentence opener about the problem. Short-form versions work better as their own story than as a compressed long one, a point developed further in our guides to clipping long video and short vertical edits.

Notice that every passage keeps its identity. The table records passages, not sentences, because a passage is a continuous stretch of one answer. Internal trims happen within a passage; passages themselves are never merged.

Cutting the Audio First: The Radio Edit

Audio-led editing means cutting the sound into shape first and then finding picture for it. For interview and talking-head editing this is the standard professional method, and the reasoning is straightforward: the audience forgives a lot of visual cutting but notices almost every audio flaw. A breath cut in half, a word clipped at its tail, a change of pitch that makes a sentence sound stitched: these break the illusion instantly. Get the sound right and the picture has a clear job to do.

In practice the editor laid all the selects onto the timeline in paper-edit order and then worked through them with only the audio in mind, often with the picture monitor switched off or ignored. This version is called a radio edit, because it should play as a coherent piece with no picture at all. The editor listened to it end to end several times before looking at a single frame of video.

1. Lay the selects in order Place each passage from the paper edit on the timeline with generous handles, so there is room to adjust every in and out point.
2. Remove false starts and repetitions Take out "so, um, what we, what we did was" style restarts, cutting on a breath or a hard consonant where the waveform shows a clean boundary.
3. Trim within answers, never across them Shorten a passage by removing a digression or a repeated clause, keeping every qualification that affects the claim.
4. Protect breaths and cadence Leave natural breaths at the start of sentences and keep the downward inflection at the end of a thought, so sentences do not sound as if they stop mid-flow.
5. Fill gaps with room tone Wherever a trim leaves silence, lay room tone underneath so the background stays continuous.
6. Play it without picture Listen end to end, ideally on headphones and then on small speakers, and mark every join you can hear.

Where to cut inside a sentence

Cutting points matter more in interviews than in almost any other format. The cleanest places to join speech are at pauses, at breaths, and on hard consonants such as "t," "k" and "p," whose sharp attack masks the edit. Vowels and soft sounds like "s" and "m" are harder, because the join often produces a click or a smear. The editor also matched inflection: if the first part of a trimmed sentence ended on a rising tone, the part it joined had to continue naturally from that tone, or the listener would hear the seam even if they could not say why.

Editing systems document these techniques in their own terms. Avid's editing techniques documentation covers trim modes and audio scrubbing that make finding the exact frame for a join much faster than dragging clips by eye, and the same ideas exist in every professional editor.

Laying Picture to the Audio: Angle Changes, Cutaways and B-Roll

With the radio edit approved by the producer, the editor turned to picture. Every audio join now needed a visual solution, and the two-camera setup meant most of them were straightforward: at each join, switch from A to B or from B to A. Because the speaker appears in a different framing at the moment the sound is joined, the audience reads the cut as a change of camera rather than a gap in time.

Rules the editor used for angle changes

• Change angle at the join, not a few frames either side, unless a small offset lands the cut on a natural head movement or blink, which hides it even better.
• Avoid switching angles too often when there is no join to hide. Alternating every few seconds makes the film feel restless and can itself signal heavy editing.
• Use the tighter angle for emotional or emphatic moments and the wider one for explanation. That gives the switches a logic beyond concealment.
• Where two joins fall close together, cover the second with B-roll rather than switching back to the first angle after a very short shot.

Using B-roll honestly

B-roll was reserved mainly for the "how it works" beat, where the speaker described the inspection cycle. There, pictures of technicians and schedules told the audience something the words alone could not. The editor was careful to match B-roll to what was being said. Footage of a specific machine under a sentence about a different type of equipment would be misleading even if no one intended it, and a technical audience notices. When nothing relevant fit a join, the editor used a neutral cutaway such as hands on a tablet rather than stretching a specific shot beyond its meaning.

Consistency matters here too. The client had a brand style for lower thirds, captions and pacing, and the editor applied it from the first assembly rather than retrofitting it, the approach recommended in our guide to editing for brand consistency.

Smoothing the Sound: Room Tone, Dialogue Editing and the Mix

Audible jumps in background noise at every cut are one of the most common giveaways of an amateur interview edit. The dialogue can be joined perfectly and the angle changes can be seamless, yet a small shift in air-conditioning hum at each join tells the ear that something happened. The dialogue editing pass exists to remove those tells.

The dialogue pass, step by step

The editor split the lav and boom onto separate tracks and chose, phrase by phrase, which source to use. Where the lav rustled, the boom covered; where the boom picked up a distant alarm, the lav took over. At each join the editor added short crossfades, usually a few frames long, to avoid clicks, and ran room tone on a continuous track underneath the whole interview so that the background never dropped out. Where one session's room tone differed from the other's, because the ventilation had cycled, the editor used each session's own tone under its own passages and crossfaded between them at beat changes, where a small change in atmosphere is less noticeable.

Noise reduction was used sparingly. Aggressive processing makes voices sound thin and metallic, and it can make the joins more obvious rather than less, because the processing responds differently to each segment. The goal was a consistent, natural background rather than silence.

Music and loudness

Music was chosen after the edit was locked, kept low under speech and dipped further during the most important statements. The final mix was delivered to the loudness specification each platform expected, and the music licensing details were logged at the same time, which is much easier than reconstructing them later; our guide to cue sheets and music reporting explains what to record.

The Meaning Check: Ethics in Interview and Talking-Head Editing

This is the step that separates responsible interview editing from merely competent editing. Once the cut was close to final, the producer read every edited answer back as text, alongside the original transcript, and asked one question of each: does this still say what the speaker meant? Not just what they said, but what they meant, including their hedges, conditions and emphasis.

Broadcasters have long treated this as a matter of editorial standards. The European Broadcasting Union's editorial guidance is a useful reference point for why fair representation of contributors matters, and many broadcasters' own guidelines state explicitly that edits must not distort a contributor's meaning. Corporate video is not bound by broadcast rules, but the reasoning applies with even more force when the speaker is an employee whose words may be quoted back to their employer.

What counts as acceptable tightening

Removing fillers, false starts, repetition and digressions is normal and expected; almost every interviewee would be dismayed to hear their raw answer played in full. The line is crossed when an edit changes the substance: what is claimed, who it applies to, how certain the speaker was, or what they were responding to. A helpful test is to imagine the speaker hearing the cut with their full original answer in front of them. If they would say "yes, that's what I meant, just shorter," the edit is fair. If they would say "I didn't say that," or even "that's not quite what I meant," it needs to change.

In our illustrative project, the meaning check caught two issues. One trim had removed "once the initial survey is complete" from a sentence about response times, which made the commitment sound immediate. The other placed an answer about pricing flexibility directly after an on-screen caption about small customers, implying a link the speaker had not made. Both were fixed before the speaker saw the cut.

Speaker Review, Legal Approval and Delivery

When a subject requires approval of the final edit, the review process needs structure, or it can drift into re-shooting the interview through comments. The producer sent the speaker a review link with the full edit, the transcript of the edited version and a note explaining that the purpose of this round was to confirm accuracy and fair representation, not to rewrite answers. Comments were collected in one place with timecodes. The speaker asked for one change: a phrase they felt sounded too casual was swapped for a slightly longer version of the same point from the second session.

Legal review followed, using the same edited transcript with every claim highlighted. Because the meaning check had already caught the two risky trims, legal approved with a single caption amendment. Running review this way, with one shared set of notes, is covered in more depth in our article on collaborative editing.

Finishing and versions

With picture locked, the project moved to finishing: color correction to match the two cameras precisely, a final sound mix, burned-in or sidecar captions for each platform, and exports at the required aspect ratios. The vertical versions were reframed rather than simply cropped, keeping the speaker's eyes in the upper third and leaving room for captions. If an offline-to-online workflow is involved, which is common when the edit is done on proxies, the conform needs to be checked shot by shot; our guide to conforming and online editing covers the details.

Captions deserve a separate check against the meaning review. A caption that paraphrases a spoken line can shift its meaning just as a bad trim can, so the editor captioned verbatim, trimming only fillers, and had the producer proof them against the approved transcript.

When You Only Have One Camera, or the Interview Was Remote

Our illustrative project had the advantage of two cameras and planned B-roll. Plenty of real interviews do not. The most common difficult case is a single-camera interview with little or no cutaway material, and the options there are narrower, so the decisions need to be made consciously.

Options for a single-camera interview

• Punch in. If the footage was shot at a higher resolution than the delivery format, such as 4K for an HD master, a digital crop to a tighter framing at each join creates a pseudo-second angle. It works best when the size change is significant, so it looks like a deliberate shot change rather than a wobble.
• Accept the jump cut as style. For some audiences and platforms, visible jump cuts are acceptable and even expected. If you choose this, commit to it consistently rather than hiding some joins and not others.
• Use graphics and text. On-screen text, simple animated diagrams or quote cards can cover joins while adding information, provided they accurately represent what is being said.
• Gather B-roll after the fact. A short pickup shoot of the workplace or product can supply cover, as long as it is matched honestly to the words.
• Cut less. Sometimes the right answer is to keep longer passages intact and choose fewer of them.

Remote interviews recorded over video calls add their own issues: variable frame rates, compressed audio, and room tone that changes whenever the speaker's laptop fan starts. Recording locally on the speaker's side where possible helps greatly, and our guide to remote video interviews covers the setup in detail.

The trade-offs of full coverage

Weigh the approach used in the walkthrough against what your budget and schedule allow. Its advantages:

• A second camera hides joins without implying anything about content.
• Planned B-roll adds information and pace while covering trims.
• A transcript-first workflow makes selection faster and review clearer.
• Recorded room tone makes dialogue joins close to invisible.

And its costs:

• A second camera and operator add cost and setup time on the shoot day.
• B-roll that is not matched carefully to the words can mislead viewers.
• Structured approval rounds lengthen the schedule, especially with legal involved.
• Single-camera workarounds such as punch-ins cost resolution and flexibility.

What the Walkthrough Teaches: Decisions to Copy on Your Next Interview

Stripped of its specifics, the illustrative project shows a pattern that holds for most interview and talking-head editing. The edit starts before the shoot, with a decision about how joins will be hidden and a written rule about what the edit may never do. The shoot delivers the raw materials the edit needs: complete-sentence answers, a second angle, targeted B-roll and room tone at every location. The edit itself runs in a specific order: transcript, paper edit, radio edit, picture, dialogue smoothing, meaning check, review. Skipping a step usually costs more time later than it saves.

It also shows where to bring in help. If an interview must be cut hard for time without altering meaning, if only one camera was available and the edit needs to look polished, or if the speaker or a legal team must approve the final cut, an experienced editor earns their fee by knowing which trims are safe and which quietly change a claim. The same holds when the footage is destined for many versions, because each cutdown needs its own meaning check. You can see how we structure this kind of work on our video editing and production services page, and the broader process decisions are covered in our article on the video editing workflow.

For in-house teams doing this themselves, the most valuable habit to adopt is the read-back. Before any interview edit goes out, read every edited answer as plain text next to the original and ask whether the speaker would recognize it as what they meant. That one check prevents the problems that no amount of technical polish can fix.

Where this comes from

• Avid — Editing techniques documentation
• Adobe Help Center — Premiere Pro text-based editing
• European Broadcasting Union — Editorial guidelines

The figures and practices above come from the sources listed.

Working on something like this?

We take on Video Editing & Production work for teams who want it done once, properly. Tell us what you are building and we will tell you honestly whether we are the right studio for it. Start a project.

Where to go next

• B-Roll and Coverage, Done Properly
• Video for Landing Pages: What Actually Works