Reaction videos
Filming Reaction Videos on iPhone
Reaction videos fail when the reaction is too small to read. How to film the thing and your face together on one iPhone, and how big the inset needs to be.
Updated
Recommended setup
- Mode
- Picture-in-picture
- Frame
- Large PIP
- Orientation
- Portrait
Viewers came for your face, not the clip they have already seen. A large inset gives the reaction enough pixels to actually read, while the source stays visible behind it.
Most reaction videos fail for the same reason: the reaction is too small to read.
The format has an obvious shape, source material on one side and your face on the other, and it is easy to assume the source is the main event. It is not. Viewers have usually already seen the thing. What they have not seen is you watching it.
Why Large PIP, portrait
Large PIP gives your face enough of the frame to carry expression. Micro-reactions, the eyebrow, the half-second of disbelief, are the entire value of the format, and they disappear at thumbnail size.
Portrait, because reaction content lives on TikTok, Reels and Shorts, and a 9:16 frame filled edge to edge beats a letterboxed landscape clip in every one of those feeds.
If your reaction and the source genuinely carry equal weight (a first listen to a track, say, where the source is new to viewers too), switch to a 50/50 split instead. That is the one case where equal billing is honest.
Framing your own half
The inset is the shot people watch. Treat it like a shot.
Get the camera at eye level. Filming up your nose is the single most common reaction-video look, and it comes from the phone lying flat or propped low. A stack of books fixes it.
Fill the inset with your face. Head and shoulders, not head-to-waist. In a small frame, distance reads as disengagement.
Light from the front. A window, a lamp, a screen: anything in front of you. Backlight turns the inset into a silhouette, and no amount of editing recovers it.
Look where the source is, not at the lens. Reaction content is one of the few formats where avoiding eye contact is correct: you are watching something, and viewers should be able to tell what.
The sync problem, and why this removes it
The traditional workflow is to film your reaction against playback, then line the two up in an editor. It is fiddly, it drifts, and it eats more time than the recording did.
Recording both cameras simultaneously removes the step entirely. Both feeds are composited as you film, so the reaction is locked to the source frame by frame. When you stop recording, the clip is finished.
What this cannot do
If you are reacting to something on your own screen, two cameras is the wrong tool, because the rear lens sees the room, not your display. That needs a facecam screen recorder, which is a different category of app. Our screen recording guide covers where that line falls.
Two cameras is the right tool when the thing you are reacting to is in front of you: a person, a performance, a gift, a view, a prank.
Before you hit record
- Decide the corner. Whatever is behind the inset is gone, and reaction sources often have subtitles along the bottom.
- Do a ten-second test and watch it back. Check that your face is exposed correctly against the source.
- Drop to 1080p for anything over a few minutes. Two camera pipelines run warm.