How to match a YouTube channel's editing style

How do I get an editor to match another channel's editing style?

Measure the reference instead of describing it. Count picture changes per minute, how shots join, how long text stays on screen, how much of the picture moves and how loud the music sits under the voice. Two channels in the same genre can differ about sixfold, so "like X" needs numbers.

Why does "make it like X" fail as a brief?

Because "like X" is a feeling, and feelings don't transfer between two people who haven't watched the same clip fifty times. On two narrated spy-documentary channels measured frame by frame, Philip Thompson changes the picture about 12.8 times a minute detected (about 14.7 corrected), and Librarian changes it about 2.16 times a minute. Same genre, same kind of true story, and one edits roughly six times faster than the other. Neither is wrong; they are different products. An editor told "cut it like a spy documentary" has to guess which of those two you mean, and a name-only brief survives being restated a dozen times while quietly meaning something different to each person reading it. A number does not have that problem.

What are the eight things to measure?

These describe an editing style well enough for another person to reproduce it, taken from measuring three narrated documentary channels frame by frame.

1. Picture changes per minute. Count every cut, dissolve or other change. Philip Thompson runs about 13 to 15; Librarian about 2; The Interpreter, a faceless disaster-documentary channel, 12.9 to 13.6.
2. Median shot length, not just the average. Philip Thompson's shots run 0.6 s to 22.6 s, median 4.1 s. Librarian's median is 29.4 s. A single average hides that range, and the range is part of the style.
3. How shots join. Hard cut, cross-dissolve, dip to black or flash, and how long each lasts. Librarian joins 95.5% of its changes with a roughly 2-second cross-dissolve; Philip Thompson splits about a third each between hard cuts, dips to black and dissolves; The Interpreter uses hard cuts on all 26 changes checked.
4. Text on screen. What share of the runtime has a caption, label or card, and how long each holds. Philip Thompson: 19.6% of the time, about 6 seconds median. Librarian: none added. The Interpreter: 34% to 42% of the time, callouts averaging 2.5 seconds.
5. How much of the picture moves, and how. Zoom in, zoom out, or a still frame held with no motion. Philip Thompson zooms in about 68% of the time; Librarian zooms out about 69% of the time on stills; The Interpreter holds a frame with no movement 18.1% of the time.
6. Whether the picture runs in color or black and white, and where that changes. A story choice more than a fixed rule, so measure it on your own reference rather than assume a default.
7. How loud the music sits under the voice. Measured in loudness (LUFS), not by ear. The Interpreter runs a constant bed about 24 dB under the narration; Librarian's bed sits about 31 dB under, changing character as the story moves; Philip Thompson has no bed in the body at all, music only in the cold open and close.
8. The opening, on its own. Pace is a curve, not a flat number. Philip Thompson's cold open runs about 19 changes a minute, easing to 14.5 in the first 10 minutes, 12.5 from minute 10 to 25, 10.8 from 25 to 30, and 8.2 at the close. The fastest cutting sits where a viewer decides whether to stay.

Full method and every number are on the pacing data page and the music-under-voiceover page; this page covers the method for any reference, not a repeat of those tables.

How do you measure these without expensive tools?

None of the eight categories need anything beyond what most editors already have open.

• Picture changes and shot length: most editing software and several free desktop tools flag scene changes automatically by comparing consecutive frames for a sharp jump. Scrub through the flagged moments yourself and log each timestamp in a spreadsheet, which gives you shot length as well.
• How shots join: step through a handful of the flagged changes frame by frame. If the picture switches in a single frame, it's a hard cut; if it takes several frames, note whether it goes through black, through white or through a blend.
• Text on screen: a stopwatch and the same spreadsheet. Start the clock when a caption appears, stop it when it's gone, and add up the seconds against the runtime.
• Loudness under the voice: Premiere Pro and DaVinci Resolve both have a built-in LUFS meter. Solo the voice, read the loudness; solo the music, read it again; the difference in LUFS is how far under the voice the bed sits. Full steps for both are on the music-level page.

How did a real brief use two reference channels?

The Spy Show is a narrated spy-history documentary series I edit for Vincent Meijer, still in production. The brief named two reference channels and asked the series to match their level of quality, rhythm and narrative, not to copy either one. That instruction only becomes usable once someone measures what "level" and "rhythm" mean on each channel, so both were analyzed before the first cut.

The two references sit at opposite ends of the same genre: Philip Thompson changes the picture about 13 to 15 times a minute with a 4.1-second median shot, roughly a third each of hard cuts, dips to black and dissolves; Librarian changes it about 2 times a minute with a 29.4-second median shot, almost all 2-second cross-dissolves, and no added text at all. The series took its cutting rhythm from Philip Thompson, frequent changes and short shots that ease off after the cold open the way his do. On music, the brief asked for a bed running through the whole video rather than only the open and close, closer to how Librarian handles sound (about 31 dB under the voice, changing with the story), even though the picture pacing did not come from Librarian. Full detail is in the Spy Show case study.

What should you copy, and what should you leave alone?

Copy the structural choices: cut rate, shot length, how shots join, how much text stays on screen, where the music sits under the voice. Leave the rest of the channel alone.

• Copy: pacing (cuts per minute and the curve across the runtime), transition mix, text timing, motion on stills, music level under the voice.
• Don't copy: the other channel's specific music tracks, its graphics and fonts, its narrator's voice or on-screen identity, and its topic or footage. A pacing choice is a technique; a music track belongs to that channel.

Pick and mix is normal, and it's how the Spy Show brief actually worked: rhythm from one reference, music approach from a different one. A brief that says "match this channel's rhythm, but run music through like this other one" beats "make it feel like either," because it names the choice instead of the channel.

How do you turn the numbers into a brief an editor can use?

Write the eight numbers down once, on one page, before the first cut, and the brief stops depending on memory or taste: cuts per minute (opening separately), median shot length, transition mix, text-on-screen percentage and timing, motion style, whether color changes, and the music level under the voice. A series style guide template lays those fields out ready to fill in, and the editor brief template covers the rest of a project around it. For how the rest of a brief should read, see how to brief your video editor.

FAQ

How do I get an editor to match another channel's editing style?

Measure the reference instead of describing it. Count picture changes per minute, how shots join, how long text stays on screen, how much of the picture moves and how loud the music sits under the voice. Two channels in the same genre can differ about sixfold, so "like X" needs numbers.

Why doesn't "edit it like [channel]" work as a brief?

Because two channels in the same genre can differ sixfold on how often the picture changes: Philip Thompson about 13 to 15 times a minute, Librarian about 2. Without numbers, "like X" could mean either one.

Can I count cuts automatically on any type of footage?

Not reliably on gameplay. A scene-change detector built for narrated documentary footage counted camera movement as cuts on a gameplay video, nine "cuts" in 30 seconds of one continuous shot. Count a sample by eye instead of trusting the automated number.

What should I copy from a reference channel, and what should I leave alone?

Copy structural choices: cut rate, shot length, how shots join, how much text stays on screen. Leave alone the other channel's music tracks, graphics and on-screen identity. A series can take its rhythm from one reference and its music approach from a different one, the way The Spy Show brief did.