Ten mistakes people make filming work in a workshop

The build is the easy part. You have done it for years. The camera is the thing you keep remembering to reposition between operations, usually a moment after you have made the cut you meant to film.

Then a fortnight of footage goes into an edit and what comes out is either too quick to follow or too tidy to believe.

A workshop earns a list this long because it is one particular room with its own ways of defeating a camera: noise, dust, a workpiece that changes shape daily, a job running across days, and hands doing the interesting part at the one angle nothing is pointing at. Most cost nothing to put right; a couple assume a clip microphone or a clamp for the phone, which is cheap beside a reshoot. The tenth is the odd one, about leaving something in rather than taking it out, and it is worth more than the rest together.

Ten ways a workshop defeats a camera

1. Lighting the bench instead of the workpiece

The strip light overhead lights the room well and the work badly, because you are standing between the two. Everybody watching came to see the cut, the joint or the glue line, and it is sitting in your own shadow. Put one lamp low and to the side of the workpiece, aimed at the work rather than the room; "Where to put your light when you only have one" is written for faces and the geometry is identical.

2. The machine and your voice on one microphone

When the planer and the explanation are written into one file as a single signal, nothing afterwards separates them, so pulling the machine down takes your voice with it. That is why build videos are so often either deafening or inaudible, which is a recording failure rather than a mixing one. Put a microphone on you and leave the camera's own to the machine, so the two arrive as two files; "Recording an instrument and a voice together" is the same problem in a different room.

3. No overhead angle on hand work

Filmed from the front, your arms and shoulders sit between the camera and the work for most of the interesting part. The overhead angle, a phone clamped above the bench looking straight down, is the only one showing what your hands are actually doing, which is the reason somebody clicked. Set it up once, leave it running through the operation, and cut to it; "Giving every camera angle a job" is about settling that before the day rather than during the edit.

4. Filming only the interesting operations

The bandsaw gets filmed. The reaching, the wiping, the clamping and the standing back to look at it do not, so the edit becomes a string of loud events with nothing between them, and every join announces itself. Film the small actions too, in the same light on the same day; "A b-roll list you can use again" gives you a list of your own actions rather than somebody else's.

5. A build that runs across days

Daylight, what you are wearing, the state of the bench and how far the piece has progressed all move overnight, and day two will not cut against day one. This is the fault behind videos where shavings appear and vanish behind you and nobody can say why. Photograph the bench and yourself before you stop for the day, and look at it again before you start; "Resuming a shoot after a break" is the checklist version.

6. Filming at the size you deliver at

Frame every shot exactly as you want it seen and you have left yourself nothing. The close-up you decide in the edit that you need cannot be lifted out of that wide, because enlarging throws away resolution and the softness reads as an accident. Film the wide a little wider than you need and film the detail separately at the size it matters at; "Zooming in without going soft" explains where the limit sits.

7. A dust mask or ear defenders over a clip microphone

The strap goes over the cable, the mask sits over the capsule, and every word through it arrives muffled and rubbing. You discover this at the edit, by which point the operation is finished and the piece has moved on. Run the cable outside the strap, clip the microphone high and clear of anything that moves, and say the sentences that matter with the mask off, but only with the machine stopped and the dust settled. Nothing on this list is worth running a tool without the protection it needs, and a guard never comes off for a tidier shot; record the talking and the cutting as separate moments instead. "Lavalier microphone mistakes" covers the rest.

8. No spoken markers across a fortnight

A build filmed over two weeks produces hours of footage across dozens of files, and the only way to find the mortise you cut on Tuesday is to scrub the lot. Say the stage out loud before each operation, into whatever is recording: the name of the step and what you are about to do. A partner who films cooking was asked to do this, and the pile stopped needing to be searched by hand; "What to say out loud while you film" is the full version.

9. Never filming the finished piece properly

A fortnight goes on the process and the finished piece gets a phone snap on the bench, under the strip light, vice still in shot. That frame is the one people share and the one somebody looks at while deciding whether to commission something, so it does more work than the entire build. Clear the bench, put the piece against something plain, use the lamp you moved for item one, and shoot it from the angles somebody would want to buy it from.

10. Cutting out every mistake

The tear-out, the mark measured from the wrong end, the joint that closed with a gap you had to fill: the instinct is to remove all of it, the same reflex that strips every breath out of a voice track, and it produces a build that appears to have gone perfectly. What that costs you is the only thing on your channel a better-equipped workshop cannot copy, which is that you have made this mistake before and know what to do about it. Keep the mistake and keep the recovery; "Should you cut out every breath" is the same reflex applied to sound.

Why does the mistake belong in the video?

Because of who is watching. A workshop video is watched by people who make things, people about to try, and people working out whether to pay somebody instead. All three want the same thing, and it is not a flawless build.

A build with no problems in it teaches nothing, because nothing was decided. It also invites the least useful reading available, which is that the maker simply has better equipment. A tear-out, followed by what you did about it, is the only footage you own that somebody with a bigger workshop cannot reproduce.

Cut the part where nothing is happening. Keep the part where you work out what to do. That distinction is most of the edit, and also the difference between showing your work and advertising it.

What would you change on the next build?

Two things, not ten, chosen because they apply to every build you film afterwards rather than this one. One for the camera and one for the sound.

For most workshops those two are the overhead angle and getting your voice onto its own microphone, because between them they repair the angle that shows the work and the recording that ruins it. Set both up before the first cut rather than partway through, and mark where the clamp and the stand go. None of it changes the one rule that outranks the shot: the guard, the mask and the defenders stay on while the machine runs, and the filming is arranged around that rather than the other way about.

Then film the tear-out when it comes. It will come, it is the part you will want to delete, and it is why somebody watches you rather than somebody else.