How to film a dog training video with both iPhone cameras
A training clip has one job a lot of other videos do not: it has to show that the dog is actually responding to the cue, not just performing near it. Cut between a separately-filmed hand signal and a separately-filmed dog, and the timing between them is whatever the edit says it is — the viewer has no way to tell a real response from a lucky one stitched in afterwards. Record both at once instead, and the timing is not a claim you are making, it is what happened. DualCam composites the front and back camera into one frame on the GPU as you shoot, so the cue and the response are not two clips brought together later — they were always the same frame.
The general mechanics of dual-camera recording do not change because the subject has four legs. What is specific to training is below: where a handler and a dog actually fit when they are two different heights, which layout survives a trick that travels versus one that stays put, why a training session is one of the few subjects that is genuinely fine to shoot as several short takes, and what the one microphone does with a clicker next to it and a dog a few feet away.
Why filming it live beats filming it in two pieces
The whole value of a dual-camera training clip is that the cue and the response are not just close together, they are the same instant, and that comes directly from how the recording is built. The two camera images are drawn into one frame in real time, and the exact frame in the viewfinder is the frame written to the file — nothing is re-timed or re-rendered afterwards. Filming your hand signal separately and cutting to the dog's reaction is a different, older way to build the same kind of video, and it always carries a small, honest question mark: how many takes did that cut hide?
That is worth having in mind before you frame anything, because it changes what "good footage" means here. A clean single take where the cue and the sit happen together in one frame is more convincing than a technically sharper video built from two.
Picture-in-picture for a trick that stays put; split for one that travels
A "sit," "stay," "down" or "shake" happens in roughly one spot, the same way a wrapped box on a table does, so a picture-in-picture inset framed once holds up for the whole take: the dog full-frame, your hand and the treat in the small window, or the reverse. A "spin," a recall across the room, "roll over," or fetch covers real ground, and a fixed corner inset was never going to follow it — the dog will simply leave the part of the canvas that camera owns.
For anything that travels, a top-and-bottom split keeps the full horizontal range of both cameras in the shot, so a dog that moves sideways stays inside the frame instead of running off the edge of an inset. Layout is one of the controls you can change mid-recording, so it is fine to start in picture-in-picture for the stationary part of a sequence and switch to split the moment the dog is asked to move.
Both lenses are fixed to the same body, and a dog and a handler are two different heights
The front and back camera sit on the same piece of phone, so they move together — you cannot put one at floor level and the other at your own eye height. Lowering the phone to a dog's height for a clean, level shot of it also lowers the front camera to that height, which means your own face reads as a lit chin and the underside of your jaw unless you crouch down to it rather than standing and shooting down and over the phone.
The practical fix for training specifically is to let go of holding the phone by hand for this one. Set it on a low stand or a stack of books at the dog's height, aimed so there is enough floor in front of it for the trick to actually happen in, and then stand or crouch wherever the cue is easiest to give. A wider back-camera lens, where your iPhone offers one, buys back some of the floor space a low, close shot otherwise loses.
One microphone, and the clicker is closer to it than the dog is
A dual-camera recording writes a single audio track from the phone's own microphone, not a separate feed tied to whichever camera is on the dog. A clicker held near the phone, or a cue spoken from where you are standing, will always come through louder than a dog's paw tap, tail thump or panting a few feet further out — that is distance to the one microphone doing what distance always does, not a setting to fix.
If the sound of the response matters as much as the cue does, it is worth closing that gap for the moment that matters rather than trusting the framing to carry it. A phone propped closer to the dog, even a little, does more for the audio than any layout choice will.
Treat a session as several short takes, not one long recording
A dual-camera recording runs straight through from record to stop, with no pause button that lets you freeze it and carry on later into the same file — that is true of any subject, but it matters less here than it usually does, because a trained dog is one of the few subjects you can actually ask to do the thing again. A reaction you only get once needs one long recording running in the background just in case; a rep of "sit" does not.
Stop after each successful rep and start a fresh take for the next one. Resolution, mute and the countdown timer unlock again between takes, so a setting you got wrong on the first attempt is easy to fix before the next; a missed cue or a dog that wanders off mid-take costs you nothing more than that one clip. Every take lands separately in the app's own recording library with a thumbnail, so a session of ten attempts gives you ten clips to scrub through afterwards, not one long file with nine mistakes buried inside it.
Expose for whichever side is harder to see
The two cameras set exposure independently, and a training session often puts the dog and the handler in genuinely different light — a dog on bright grass in full sun while you stand in the shade of a porch, or the reverse near a window indoors. Tap the frame that needs help before you start the rep, since tap-to-focus carries exposure with it on that camera. There is no shared correction across the two feeds and no way to fix a dark or blown-out side afterwards, because the composited frame you recorded is the only version of that moment there is.
Common follow-up questions
Should the dog or my hand cue be the main frame?
Either works, and you can swap which camera fills the frame between takes, or mid-recording if you want to try it within one. Dog full-frame with your hand and the treat as the inset reads as a demo of the trick; the reverse reads more like your own tutorial. Neither is wrong — pick based on what the clip is meant to teach.
What if the dog does not perform the trick on the first try?
Stop and start a new take. Unlike a reaction you only get once, a trained cue is something you can reasonably ask for again, and a dual-camera recording cannot be paused mid-file anyway — starting a fresh take costs you nothing but that one clip, and it lands in the app's library alongside the others so you can pick the best rep afterwards.
Why can I hear the clicker clearly but barely hear my dog?
The recording carries one audio track from the phone's own microphone, not a channel tied to whichever camera the dog is on. A clicker or a cue spoken near the phone is simply closer to that microphone than a dog a few feet away, so it comes through louder regardless of which camera is the main frame. Moving the phone a little closer to the dog helps the audio more than any setting will.
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.