Reading to camera is a mechanical problem
The reason a piece to camera is hard has almost nothing to do with the writing. It is that looking at a script and looking at a lens are two different directions, and every glance between them is visible to whoever watches it later. A teleprompter exists to collapse those two directions into one.
A dedicated hardware prompter does this with a beam splitter and costs real money. A software prompter on the screen you already have gets you most of the way, provided the text is large enough to read from where you are sitting and moves at a pace you do not have to think about.
Both of those are settings on this page, and both are the things people get wrong on their first attempt. The font is usually too small, because it looks fine while you are typing at desk distance and is unreadable from where the camera wants you. The speed is usually too fast, because reading silently is quicker than reading aloud.
Setting up a take
- Paste or type your script into the editor. It is kept in your browser between visits.
- Set Font size first. Sit where you will actually sit and check you can read it comfortably.
- Set Speed second, since the two interact.
- Enable Mirror mode only if you are using a beam splitter rig.
- Press start. A three second countdown runs, then the text begins to move.
- Use space to pause, the arrow keys to adjust speed and escape to get back to the editor.
Why speed is ten steps and not a slider
Scroll speed is quantised into ten levels, each twenty pixels per second faster than the last. That is a design decision rather than a limitation.
The moment you need to change speed is the moment you have started reading and discovered the pace is wrong. At that moment you have a fraction of a second, your eyes are on the text and your hands should not be visible. A discrete step you can nudge with an arrow key is usable in that situation. A continuous slider is not, because it requires you to look at it.
The same reasoning drives the keyboard shortcuts. Space, arrows and escape are all findable without looking, and adjusting speed mid take is normal rather than a failure.
Set the font before the speed
Because the scroll is measured in pixels rather than in words, the two settings are not independent. Doubling the font size at a fixed speed halves the rate at which words pass, since each line is now twice as tall.
That is why the order matters. Choose the font size from where you will be sitting, which is nearly always larger than feels right while you are setting it up, and only then find the speed that matches your speaking pace. Doing it the other way round means re-finding the speed every time you touch the font.
If you want to know in advance roughly how long the script will run, Speech Time Calculator estimates it from the word count at three delivery rates, and that estimate is a better starting point than guessing.
Recording alongside it
Nothing here records anything, which is deliberate: a prompter that also captured video would need the camera and the screen at once and would compete with whatever you are actually recording on. Use it next to your real recording setup.
If the take is a screen capture rather than a piece to camera, Screen Recorder captures the screen in the browser. For audio only, Voice Recorder writes an MP3 or WAV. If you would rather have the words spoken than read them yourself, Text to Speech generates the audio from the same script. The rest of the tools for checking a setup are on the testers hub.
