How to make a picture-in-picture video on a Mac
Picture-in-picture is one small clip laid over a larger one: a webcam over a screen recording, a reaction over gameplay, a detail shot over a wide. The technique takes a minute. The things that make it look amateur are the inset being too small to read, both audio tracks playing at once, and the inset parked on top of something that mattered.
Size and position
Aim for the inset at roughly a fifth to a quarter of the frame width. That sounds small until you remember most viewers are on a phone: a webcam at a tenth of the width is a thumbnail of a face nobody can read expressions on, and expressions are the entire reason the webcam is there.
Bottom-right is the default for landscape, because Western viewers finish scanning there and most existing on-screen furniture sits there. Two reasons to move it: the main footage has something important in that corner, or the platform puts its own interface there — vertical video on any social platform has controls and captions crowding the bottom, so a bottom inset is a bad idea and top-left is usually safer.
Whatever you pick, keep it in the same place for the whole video. An inset that hops around draws attention to itself and away from the content.
The audio decision, made once
Two audio tracks playing simultaneously is unlistenable, and it is the single most common mistake in PIP video. Decide which clip carries the sound and mute the other outright.
Usually that is the person speaking. If the main footage has meaningful audio too — gameplay, a demo with sound — you have a mix, not a PIP problem, and the fix is to duck the background under the voice rather than to let both run flat. If you cannot duck, mute the background.
The mistake that looks amateur
Stretching. A 16:9 webcam squeezed into a square inset, or a vertical phone clip stretched to fill a landscape box, distorts faces in a way viewers notice instantly even when they cannot say what is wrong. Crop the inset to the shape of the box instead of scaling it — you lose some edges and keep correct proportions, which is always the better trade.
A close second: no separation. An inset that shares a colour with the region behind it dissolves into the background. A thin border or a subtle drop shadow fixes that in one step, and is the difference between "layered" and "pasted on".
Syncing two clips without timecode
If both cameras recorded the same moment, clap once at the start of the take. In the edit, line the two audio waveforms up on the clap spike and the picture follows. It is a fifty-year-old trick because it works, and it takes ten seconds at the shoot versus a frustrating hour afterwards.
Without a clap, find any shared transient: a door, a laugh, a hand hitting a desk. Failing that, align on a visible action and accept a frame or two of drift, which nobody notices unless the subject is speaking on camera in both.
In Crisp
Add both clips, choose the picture-in-picture layout, pick which one keeps its audio, and export. The inset is cropped to fit rather than stretched, so faces keep their proportions. It runs on your Mac, so a screen recording that contains anything private never leaves it.