Video editing and creative production

Why post-production runs in stages, and what each stage is for

Post-production is a pipeline, not a single task. Each stage answers one question and hands a settled decision to the next. When teams try to do everything at once, they end up grading shots that get cut and mixing audio for a section that changes the following week.

The order exists because the cost of change rises as you move along it. Moving a scene during a rough cut takes a minute. Moving the same scene after the graphics, grade and mix are done means redoing all three. That is why we push hard for honest structural feedback early and small polish feedback late, rather than the other way round.

It also explains why a project that looks nearly finished can still be days from delivery. Titles, colour, sound and export are not decoration bolted on at the end. They are a body of work in their own right, and they only start once the cut has stopped moving.

• Ingest and organise: footage checked, backed up, labelled and synced to audio
• Selects and assembly: every usable take laid out in order, nothing judged yet
• Rough cut: structure and argument, deliberately loose and too long
• Fine cut: pacing, rhythm and the exact frame each cut lands on
• Picture lock: the edit stops changing so finishing can safely begin
• Finishing: colour, sound mix, motion graphics, captions
• Delivery: platform exports, subtitle files, clean master and archive

What separates a professional edit from an amateur one

Almost nobody can name what makes an edit look professional, but everybody feels it within ten seconds. It comes down to four things, and none of them are expensive effects.

Pacing is the first. An amateur edit holds every shot a beat too long, lets sentences run to their natural end, and keeps the pauses where the speaker breathed. A professional edit removes the dead air, cuts on the meaning rather than the silence, and varies rhythm so the piece breathes instead of plodding. Sound is the second and the one people underrate most. Clean dialogue, music that ducks under speech instead of fighting it, and room tone under the gaps so cuts do not sound hollow. Poor audio reads as poor quality even when the picture is fine.

Colour is the third. Shots that were filmed across a changing afternoon should look like one continuous scene, skin tones should look like skin, and the grade should support the mood without announcing itself. Captions are the fourth. Timed properly, sized for a phone, positioned clear of platform interface elements, and free of the errors that automatic transcription makes with product names. Get these four right and modest footage looks considered. Get them wrong and expensive footage still looks like a first draft.

• Cut on meaning, not on the pause after it
• Dialogue always sits on top of the mix and stays intelligible
• Match shots to each other before creating any look
• Caption for a muted phone viewer, because most of them are
• Every effect should have a reason, or it should not be there

Why platform specifications differ, and what that means for your delivery

Platform specs are not arbitrary. They follow from how each surface is watched and how the player crops, compresses and overlays the frame. Understanding why they differ is what stops one edit being reformatted badly for five destinations.

YouTube is watched horizontally, often on a television or a laptop, frequently with sound on and for several minutes at a time. That supports slower openings, wider shots and small on-screen text. Vertical feeds are the opposite. They are watched on a phone held in one hand, autoplayed on mute, thumb hovering over the next post. Text must be large, the first second has to carry meaning, and the safe area is much smaller than the frame because the platform lays its own buttons, captions and handles over the edges. LinkedIn and Facebook feeds sit somewhere in between, with square and four-by-five formats winning simply because they occupy more vertical space in a scrolling feed.

The practical answer is to edit once at the highest quality and reframe deliberately for each destination, rather than exporting one file and letting each platform crop it. That means shooting or sourcing footage with headroom, keeping graphics in a separate layer so they can be repositioned, and exporting a high-bitrate file so the platform only compresses once. Sending an already heavily compressed upload means it gets compressed a second time, and that is where the blocky artefacts come from.

• Horizontal 16:9 for YouTube and websites, where viewing is longer and sound is on
• Vertical 9:16 for Reels, Shorts and TikTok, with text kept inside the safe area
• Square or 4:5 for feed placements, because they take more screen space
• Reframe on purpose rather than letting the platform crop for you
• Upload high quality and let the platform compress once, not twice

How to brief an editor so the first cut is close

The quality of the first cut is decided before anyone opens the timeline. A weak brief produces two extra revision rounds, and those rounds cost more than the time the brief would have taken.

A good brief answers five things. What is this video for, in one sentence. Who is watching, and where. What must be true by the end, meaning the single thing a viewer should remember or do. How long it can be. And what it should feel like, ideally shown through two or three reference videos with a note on exactly what you like in each. References do more work than adjectives. Saying you want something to feel premium means very little. Pointing at a video and saying you like the slow openings and the restrained music is actionable.

The other half is logistics. Send the original camera files rather than a phone or app export, name them so takes are identifiable, flag the parts you already know you want kept or cut, and supply brand assets, fonts and any music you have already licensed. If there is a hard deadline tied to a launch or a campaign, say so at the start rather than at review, because it changes how the schedule is built.

• One sentence on the purpose, before anything else
• Two or three reference videos plus what specifically you like in them
• The single takeaway the viewer must leave with
• Original files, sensibly named, not compressed re-exports
• Brand assets, fonts, licensed music and any legal wording up front
• The real deadline, stated at brief rather than at review

What your footage does and does not allow

Editing shapes what was captured. It does not create what was not. Being clear about that early prevents a lot of disappointment, and it changes how you shoot next time.

Recoverable in post: uneven exposure, wrong white balance, mild handheld shake, background hum and air conditioning noise, inconsistent levels between two speakers, dead air, rambling that needs tightening, and shots that need reframing if there is resolution to spare. Not recoverable: focus that missed, highlights blown to pure white with no detail left, audio that distorted at the source, and wind noise on an unprotected microphone. Detail that was never recorded cannot be restored, and tools that claim otherwise are guessing at the missing information.

If you only improve one thing, improve sound. A modest camera with a proper lapel or shotgun microphone will beat an expensive camera relying on its built-in mic every time, because viewers forgive a soft image far more readily than they forgive muddy audio. If you improve a second thing, shoot slightly wider than your final frame. That headroom is what makes vertical versions possible without cutting off the top of a head.

Turnaround realities: what is fast, what is not, and why

Turnaround depends far more on decision speed than on editing speed. The editing hours for a given piece are fairly predictable. The waiting is not.

A short-form vertical cut from existing footage is quick, often a day or two. A long-form piece with multiple cameras, graphics, a full mix and a grade takes considerably longer, because each of those stages is real work and several of them cannot start until picture is locked. Anything involving voiceover recording, music licensing, legal sign-off or footage that still needs to be shot adds calendar time that has nothing to do with the edit itself.

The two things that most often stretch a schedule are late structural feedback and scattered review. If a comment arrives after finishing has started asking to reorder sections, the colour, graphics and mix for those sections have to be redone. If feedback arrives from four people in four emails with contradictory notes, somebody has to reconcile them before work can continue. We ask for consolidated feedback on a timestamped review link, from one named decision maker, within an agreed window. That single arrangement saves more days than any faster software.

• Short vertical cuts from supplied footage: fastest turnaround
• Long-form with grade, mix and graphics: multi-stage, plan for it
• Voiceover, licensing and legal sign-off add calendar time, not edit time
• One consolidated set of notes from one decision maker per round
• Structural changes belong at rough cut, polish notes at fine cut

Video as a search asset, not just a social post

A video is one of the few assets that works in three places at once: on the platform it was made for, in search results, and as text. Most businesses only use the first.

YouTube functions as a search engine in its own right, with its own query behaviour and its own ranking signals. Titles, descriptions, chapters and a clean uploaded transcript all give it something to index. That transcript is the part most people skip, and it is the most valuable, because it turns a video into crawlable text. Automatic speech recognition will mangle brand names, product names and technical terms, so uploading a corrected subtitle file is worth the small effort. Embedding the same video on a relevant page of your own site, with a written summary or a full transcript alongside it, gives that page substance it would not otherwise have and serves people who would rather read than watch.

On AI answer engines, keep expectations honest. Video hosted on large platforms is visible to those systems and video results do appear in AI-generated answers, but no agency can promise a citation or a stable position. What we can say with confidence is the mechanism: text is what these systems read, so a video with an accurate transcript, a clear written summary and a page that provides context is far more legible to them than an unlabelled upload with auto-captions. Treat the transcript as content, not as an accessibility afterthought, and the same asset earns its place twice.

• Upload a corrected transcript rather than relying on automatic captions
• Write real titles, descriptions and chapters, not filename placeholders
• Embed video on the page it supports, with a written summary beside it
• Repurpose one long-form piece into vertical cuts, stills and quotes
• Judge video on watch time and enquiries, not view counts alone