Watch someone record themselves and you can see the exact moment it happens. They are two sentences in, the delivery is warming up — and then they catch their own eye in the preview. A hand goes to the hair. The shoulders square themselves into the posture of someone being observed. The sentence they were building loses its footing, because the part of the mind that was carrying it has been reassigned to a different job: monitoring. The preview window is a mirror with a record light on it, and everybody knows what mirrors do to behaviour. Nobody walks past one without a glance; nobody talks in front of one the way they talk across a table. The camera was supposed to capture you speaking to a person. What it captures instead is you supervising a broadcast of yourself, and viewers can tell the difference even when they cannot name it.
This is not vanity, and it does not spare the experienced. Psychologists have a name for the state — objective self-awareness: the moment you become an audience to yourself, attention splits, and self-correction crowds out expression. You hear it as much as you see it. Sentences flatten into recitation; the natural gesture freezes halfway; the eyes flick, every few seconds, from the lens to the little picture and back, a movement small enough to be unnameable on camera and large enough to read as distraction. Television learned this decades ago. A news studio is full of monitors, and they face everyone except the anchor; the anchor gets words. And it is worth asking what the preview is actually for. It exists so you can frame the shot — a job that is finished before the take begins. Every second it stays on screen after that, it is doing its other job: turning the speaker back into a spectator.
The fix is almost embarrassingly direct: put the words where the mirror is. In Talk2Camera the script panel is not fixed to a strip of the screen — you drag it, and it grows as you pull; drag it past about nine-tenths of the way and it fills the whole frame, covering the camera preview at an opacity you choose. The picture is gone, the words are everywhere, and the camera keeps recording you behind them exactly as before — same framing, same focus, same file as if the preview were still glowing away. Nothing about the capture changes; the only thing removed is the mirror. What replaces it is the thing you actually needed in front of you all along: the next sentence, large, at eye height, roughly where the lens is.
The opacity slider decides how much truth you want underneath. Some people want a faint ghost of the frame behind the text — enough to sense they have not slid out of shot, too little to make eye contact with. Others want none: solid words, total commitment, the closest a solo recording setup gets to just talking. Neither is wrong, and the choice is not permanent — the panel comes back down with the same gesture that raised it, any time. Check your framing between takes, fix the stray collar, and cover the picture again before you speak. The result on screen is a strange sight the first time — you are recording a video you cannot see — and the recording itself is the argument: shoulders that stay loose, sentences that land where they were aimed, eyes that hold the lens because there is nowhere else left to look.
The gesture is the same wherever the phone happens to be — held in your hand for a walking clip, or standing on a desk mount for the sit-down explainer. That matters more than it sounds, because the handheld case is where self-monitoring does its worst work: your own face is inches away, filling a third of the screen, impossible to politely ignore. Covered by the script, the phone in your hand stops being a compact mirror and becomes a page. And because the app follows your voice through the script — the text advancing as you speak, at your pace rather than a motor’s — there is nothing on that page to manage either: no scroll to chase, no speed dial to nudge. What is left is the arrangement every other kind of speaking already uses: the words in front of you, the audience past them, and your whole attention on the line between the two. The camera does the watching. That was always its job. There is a fear the first time, and it deserves naming: if I cannot see the picture, how do I know it is still there? You know it the way every presenter before the preview era knew it — you set the shot, you covered it, and the take that comes back is the proof. After two or three sessions the trust is boring, the way trusting a tripod is boring. What felt like flying blind turns out to be the opposite. It was the preview that was flying you.