Why the sixth take is rarely better than the first

Here is the short version. You are not getting worse at talking between take one and take six. You are doing four separate jobs at the same time, and the first one to give way is the one that was quietly holding your place in the script. Everything after that feels like a delivery problem because that is where it shows up, on your face and in your voice, but it started somewhere else.

This is worth knowing because the usual response makes it worse. You decide you need to know the script better, so you read it through twice more. Then you hit record with even more to hold in your head, and you get four sentences in instead of three.

What is happening in those three sentences

Say something out loud to a camera and you are running four jobs at once:

• Recalling what comes next, which is the job nobody counts because it feels like it should be free.
• Sounding like a person who means it, rather than someone reciting.
• Driving whatever else is in the room. A slide that has to change, a diagram you promised to draw, a window you need to share.
• Marking yourself as you go, which is the one that ends the take.

The fourth job is the killer. You cannot judge a sentence and say it well at the same time, and the moment you notice a stumble you have moved from performing into reviewing. Recovering means finding your place again, and by then the sentence is gone.

A stumble on its own costs about two seconds. Stopping costs the whole take.

What a restart takes off you

Restarting is not free, and the cost is not the ninety seconds you lose.

The first time you say a sentence out loud you are explaining something. By the sixth you are reciting a sentence you have heard five times, and people can hear the difference even when they cannot name it. The words arrive a fraction early. The emphasis lands where it landed last time rather than where the meaning is. You stop discovering the thought in front of the camera, and discovering the thought is most of what makes someone watchable.

There is a physical cost too. Six run-ups at the same opening paragraph flattens your voice, and it moves you from interested in the subject to wanting to be finished. Some of that reaches the recording.

Four things that help, in the order they help most

Stop trying to memorise it

Memorising converts a reading job into a recall job, and recall is the thing that fails under a red light. You do not need to know the script. You need to be able to see the next line without looking away from the lens, which is a different problem with a much better solution.

Let the small stumble stand

Watch anyone you like on camera and you will catch them re-starting a word, losing a thread and picking it back up. It reads as thinking. A take with two of those in it is almost always better than the fifth clean one, and if a stumble genuinely has to go, it takes seconds to remove afterwards.

Give yourself the rule out loud before you begin: keep going unless the sentence was wrong, and be specific about what wrong means. A word said twice is not wrong.

Hand off the driving

If a deck has to advance, or a diagram has to appear at the right moment, that is a second person's job and you do not have a second person. Every time you reach for a key you leave the sentence you were saying. Either arrange things so nothing needs driving mid-take, or use something that holds those jobs while you talk.

Record it in one sitting, then stop

Set a limit before you start. Three takes, then whichever was best goes out. Almost nobody regrets this a week later, and the alternative is an afternoon that ends with eleven files and no video.

Where this leaves the script

The unhelpful part of all this advice is that it asks you to be more relaxed on demand. The useful part is that most of the load is not really yours. It belongs to the setup. A script you can read at the lens removes the recall job. Something that keeps pace with you removes the finding-your-place job. Slides and a board that sit in the same place as the camera remove the driving job. What is left is saying the thing, which is the part you were good at before you turned the camera on.

That is what we built LensFloat to do. You paste in what you want to say, it sits at the lens and moves when you move, and the deck and the board are on the same stage rather than in another window. Take one is usually the one you keep.