Should you look at the camera or the screen when recording dual camera video?
Look at the lens for anything addressed to the viewer, and at the screen when you are genuinely checking something. That sounds settled until you are recording with both cameras at once, because a dual recording gives you a real reason to keep watching the screen — the back camera’s view is live in front of you, and it is usually the thing you are talking about.
The distance between those two places is a couple of centimetres on the phone and clearly visible in the finished video. It is worth understanding where that comes from, because then the choice stops being a rule to obey and becomes something you decide per shot, rather than eye contact you keep meaning to hold and keep breaking.
Why looking at the screen reads as looking away
The front camera lens sits above the display, not behind it. When your eyes are on the middle of the screen, they are aimed at a point some distance below the lens, and the recording captures exactly that — a face looking slightly downward, past the viewer rather than at them. Nothing is wrong with the camera; it is faithfully recording where you were looking.
What decides how obvious this is, is the angle rather than the physical gap, and the angle depends on how far away the phone is. At arm’s length, the drop from the lens to the middle of the screen is a wide angle and the downward gaze is unmistakable. With the phone a few steps back on a stand, the same physical gap becomes a narrow angle, and the difference between looking at the lens and looking at the screen becomes hard to see at all.
None of this is specific to dual-camera recording — it is true of every selfie video anyone has ever shot. What is specific is that a dual recording puts something genuinely worth watching on that screen.
A dual recording gives you a reason to watch the screen
With one camera pointed at your face, the screen shows only you, and after a few seconds there is not much to check. With both cameras running, the screen also carries what the back camera sees: the object you are pointing at, the street you are walking down, the second person sitting opposite. That is live information about half your frame, and it is genuinely useful.
It matters more here than it would in a preview that only approximates the result. In a dual-camera app that composites in real time, the frame in the viewfinder is the frame being written to the file — checking the screen really is checking the finished shot, not a rough guide to it. That is exactly why the pull is stronger than in an ordinary selfie video, and why people who hold eye contact comfortably in a single-camera take find themselves drifting in a dual one.
So the honest framing is not “stop looking at the screen”. It is that each glance costs you a moment of eye contact, and some glances are worth that and some are not.
When the screen is the right place to look
A reaction shot is the clearest case. You are reacting to something the back camera is pointed at, and looking at it is the content — holding eye contact with the lens while reacting to something off to the side would read as strange rather than engaged. The same applies to a tutorial where your hands are doing the work: your eyes following your own hands is what a person doing that task actually looks like.
A two-person interview is another. When someone is sitting opposite you, they are what you look at, and a phone standing between you catches both faces without either person having to perform eye contact at a lens. Looking at the phone in that situation would be the odd choice, not the polished one.
The moments that genuinely want the lens are the ones addressed to the viewer directly: an intro, an outro, a question you are asking them, the line you want to land. Those are usually short, which is what makes this workable — you are not holding eye contact for ten minutes, you are finding it for the handful of seconds that need it.
- Talking to the viewer — intro, outro, a direct question, the punchline: lens.
- Reacting to something, demonstrating with your hands, talking to a second person: whatever you would look at if the phone were not there.
- Checking framing, exposure or what the back camera has drifted onto: screen, briefly, then back.
Setup choices that shrink the gap
Distance and height do most of the work
Because what shows is the angle rather than the gap, putting the phone further away and closer to your own eye level is the single most effective change available. A phone on a stand a few steps back, roughly at eye height, makes the lens-to-screen difference small enough that a glance at the screen barely registers. A phone held at arm’s length and slightly below your face is the opposite: both the downward angle and the screen-versus-lens gap are at their most visible.
Put the self-view inset in a top corner
In a picture-in-picture layout, the small window can be dragged to any of the four corners, and it can be moved while the recording is already running. When the small window is your own face, a top corner puts it near the lens rather than at the bottom of the phone, so the glance you take to check yourself travels much less far from where the camera is. It is a small effect, and it is free — the corner has to be somewhere regardless.
Use the countdown to be settled before the first frame
A countdown timer — off, 3, 5 or 10 seconds — exists so the take does not begin with you reaching for the phone. That reach is also the moment your eyes are furthest from where you want them, so a few seconds of countdown buys you a first frame where you are already in position and already looking at the right place.
Do not trust the mirror to tell you where you are looking
The front camera preview is mirrored, the way a bathroom mirror is, because that is the only way aiming your own face at a screen works. That means it is not a reliable instrument for judging gaze direction while you shoot — glance left on screen and the recording has you glancing right. If the direction of a look matters in a shot, the way to check it is to watch the take back afterwards, not to read it off the live preview.
The practical version of all this
Most of the fix is not discipline, it is arrangement. Put the phone where the two places you might look are nearly the same place, decide before the take which few seconds actually want the lens, and stop treating a glance at the screen as a mistake — half the time it is exactly what the shot is about.
DualCam makes the checking part cheap: the picture-in-picture window drags to any of the four corners mid-recording, layout and size go with it, and every take stays in the app’s own library with a thumbnail, so you can watch one back before deciding whether to keep it. “Was I looking in the right place” is a question you answer by watching, rather than by guessing while you shoot.
Common follow-up questions
Will viewers actually notice if I look at the screen instead of the lens?
It depends almost entirely on distance. At arm’s length the downward gaze is obvious. With the phone several steps back at eye level, the angle between the lens and the screen is small enough that most viewers will not register a difference.
Can I check the screen mid-take without ruining the shot?
Yes. Brief glances read as normal — people look away while they think and while they gesture. What reads as disengaged is watching the screen continuously through a passage that is meant to be addressed to the viewer.
Does the mirrored self-view help me tell which way I am looking?
No, it works against you, because left and right are swapped in the preview. Mirroring is there so you can aim your face, not so you can judge gaze direction. Watch the recording back if the direction of a look matters.
Want to just do this?
DualCam records the iPhone front and back cameras at the same time and writes one finished MP4 while you shoot. Free, no account, no ads, nothing leaves the phone.