Picture-in-picture is one video playing on top of another. It is how a reaction sits over a clip, how a face sits over slides, and how a phone recording sits over a screen capture.

The distinction that matters: clip or overlay

A clip goes in the sequence. It plays after the thing before it and before the thing after it, and adding one makes the video longer.

An overlay goes on top. It plays at the same time as whatever is underneath, and adding one does not change the length of anything.

An audio waveform shown before and after silences are closed up, the take shorter.

Picture-in-picture is always the second. If you drop a second video in and your timeline gets longer, you added a clip, and no amount of resizing will turn it into a panel.

Adding one

  1. Open your main recording in Zella.
  2. Drag the second video onto the canvas, or add it as a video overlay from the timeline.
  3. It lands at 45 percent of full size, which is a deliberate starting point rather than a default nobody chose: big enough to read a face, small enough not to cover the thing you are talking about.
  4. Drag it where you want it. Drag a corner handle to resize.

The overlay gets its own lane on the timeline, so you can trim it, move it in time, and have it appear only for the part of the video where it belongs. A panel that is on screen for the whole video is usually a panel nobody notices by the end.

Sizing and placing it

Two rules cover almost every case.

Put it where nothing is happening. On a screen recording that is normally a bottom corner, because applications put their content top and left. On a talking head it is the side you are not looking towards, because a face pointing at a panel reads as intentional and a face pointing away from it reads as a mistake.

Keep it out of the platform's furniture. Vertical feeds put usernames, captions and buttons over the bottom and right of the frame. A panel in the bottom right of a Reel is a panel behind a Follow button. The platform sizes and safe zones are worth checking before you place anything, not after.

Cropping, rounding and the bubble

An overlay can be cropped from each edge independently, which is the fix for the most common picture-in-picture problem: a 16:9 camera panel showing a metre of empty room on either side of your head. Crop the sides, and the panel is suddenly a portrait of a person rather than a portrait of a room.

Masking the overlay to a circle gives the webcam-bubble look, and if your recording came from a screen-and-camera take, that bubble is loaded for you automatically as its own overlay - see adding a webcam bubble to a screen recording.

Background removal works on an overlay too, so a camera panel can lose its room entirely and sit on the screen recording as a cutout instead of a rectangle. Click the panel first: with nothing selected the background fills the whole canvas instead of cutting out the panel, which looks like the feature doing nothing.

Getting the audio right

This is the step that gets skipped, and it is the one people hear.

Each overlay carries its own volume, separate from the main clip. That matters because two videos playing at once means two sets of audio playing at once, and the default of both at full is almost never what you want.

  • Reaction over a clip: your panel at full, the clip underneath dipped enough that its speech does not compete.
  • Face over slides: the panel full, the screen recording's system audio near silent unless a sound is the point.
  • B-roll over narration: the overlay muted entirely. B-roll is picture, not sound.

If a panel needs to be muted for the whole video, mute it rather than pulling it to zero and hoping. Muted is a state you can see; a fader at zero is a state you will re-discover during export.

Layer order and locking

Overlays stack, and the stacking is explicit: higher layer draws on top, the same way Final Cut Pro and Premiere order theirs. Two panels in the same corner is a layer-order question, not a positioning one.

Two switches save real time on a busy canvas. Hide takes an overlay out of both the preview and the export without deleting it, which is how you compare two versions. Lock makes it unselectable, which is how you stop dragging the panel every time you meant to drag the thing behind it.

Frequently asked questions

How big should a picture-in-picture panel be? Big enough to read the face at the size people actually watch, which on a phone means larger than it looks correct on your monitor. Around a third of the frame width is a safe start; the 45 percent default is deliberately generous so you shrink rather than hunt for a handle.

Can I make the panel a circle? Yes. Masking an overlay to a circle gives the webcam-bubble treatment, and a screen-and-camera recording loads its camera as a circular overlay automatically.

Why can I hear two videos at once? Both are unmuted. An overlay has its own volume control, separate from the clip underneath, so set the panel and the main clip deliberately rather than leaving both at full.

Can the panel move during the video? Yes, with keyframes on its position, behind the Advanced Tools switch. For most videos a fixed panel that appears and disappears at the right times is better than one that travels.

Does an overlay make my video longer? No. Overlays play on top of the timeline. Only clips extend it, which is the difference to watch for if your video grew when you did not expect it to.

Can I do this on iPhone? Yes, and the panel resizes free-form with independent width and height - see resizing a picture-in-picture video on iPhone.