Picture-in-picture for a product demo
A screen recording with your face in the corner is the most common product demo format, and the most commonly botched. Here is how to record the two parts, keep them in sync, and place the overlay so it helps rather than hides the product.
The format is everywhere: a screen recording of the product, and a small video of the person explaining it in a corner. It works because a face holds attention far better than a cursor does, and it makes a demo feel like someone showing you something rather than a tutorial.
It is also easy to get subtly wrong — the face covers the button being clicked, the lips drift out of sync with the voice, the overlay is too large on a phone. None of it is hard to fix once you know where the problems come from.
Record the two parts separately
You can record a screen and a camera at the same time on a Mac with no extra software, as two separate files:
- Screen: press Shift-Command-5, choose Record Entire Screen or Record Selected Portion, and pick your microphone under Options.
- Camera: open QuickTime Player and choose File → New Movie Recording. Photo Booth also works.
Recording them separately is a feature, not a limitation. You can re-record the camera part without redoing the screen, crop the camera shot freely, and choose later how large the overlay should be.
Record the screen at the size you will publish. If the final video is 1920 × 1080, recording a small window on a 5K display and scaling it up will look soft; recording a portion of the screen that matches the output shape looks sharp.
Keep them in sync
Two recordings started by hand will not start at the same instant, and you will have to line them up.
The old trick is still the best one: clap once, visibly, at the start, with both recordings running. The clap is a sharp spike on the audio waveform of each file and a single frame on camera. Line the spikes up in the editor and the two tracks are in sync to within a frame.
Then choose one audio source — usually the camera's microphone if it is closer to your mouth, or a separate microphone if you have one — and mute the other. Two microphones picking up the same voice a few milliseconds apart sounds hollow.
Placing the overlay
A few rules that hold for almost every demo:
Keep it small. Roughly a fifth to a quarter of the frame's width is enough to read expression. Larger than that and it competes with the product.
Put it where nothing happens. Most apps keep their busiest controls along the top and the left. The bottom-right corner is usually the safest default — but watch the whole recording, because the one moment you click a button in that corner is the moment the overlay covers it. Moving the overlay for that section is better than hiding the click.
Leave a margin. A consistent gap from the edge of the frame, the same on both sides, looks deliberate. An overlay pushed flush to the corner looks accidental.
Crop the camera shot. Nobody needs to see your bookshelf in a thumbnail. Crop close to head and shoulders so the face fills the overlay.
Think about where it will be watched. On a phone, a 1080p frame is shrunk to a few inches wide, and a small overlay becomes a postage stamp. If the demo is for social feeds, consider a portrait or 4:5 frame with the screen recording on top and the face below, rather than a corner overlay at all.
Captions are not optional
Most people watching a demo on social media have the sound off. Captions turn a silent screen recording with a talking thumbnail back into something that makes sense, and they help anyone watching in a second language. Keep them clear of the overlay.
Where it gets awkward without an editor
QuickTime can trim and join clips, but it cannot put one video on top of another. The screenshot toolbar cannot record a camera overlay into the screen recording. And a live overlay recorded straight into the screen capture, which some apps offer, fixes the size and position before you have seen the result — if it covers something, you record again.
What Garfi does
Garfi Video Editor is a native Mac video editor, and picture-in-picture is one of its core tools: put the camera recording over the screen recording, crop it to the face, position and size it, and change it section by section. It supports one picture-in-picture layer at a time, which is exactly what this format needs, and worth knowing if you had planned several.
Waveforms make the clap easy to find. You can detach audio, mute the track you do not want, add a voice-over, and fade in and out. Editable automatic captions use Apple speech recognition, and you can correct them before export. Canvas formats cover landscape, portrait, square and 4:5, so one project can become both a website demo and a feed version.
Export is MOV or MP4 in H.264 or HEVC, up to 4K where supported, with no watermark. The download is free with an export trial you start yourself; a one-time purchase unlocks export permanently, and editing and saving .gvideo projects stays available either way. There is no account, no developer cloud and no generative AI.