You record a twelve-minute walkthrough. The take is good — your pacing is right, you don't stumble once. Then you play it back and notice the camera bubble sits exactly on top of the sidebar you spend the first two minutes explaining.
If your screen recorder captures the camera "into" the video as it records, the only honest fix is to record the whole thing again. This post is about why that happens, what the usual workarounds actually cost you, and what changes when the camera is kept as a separate track instead.
Why most recorders won't let you move the camera later
The short version: most tools don't record a screen and a camera. They record one already-composited video.
At capture time, the recorder takes the screen frames, takes the camera frames, composites the camera bubble on top, and encodes a single flattened stream. From that moment the two sources no longer exist independently — the overlay is baked into the pixels, the same way text burned into a screenshot can't be un-typed.
This isn't laziness. Flattening during capture is genuinely simpler and cheaper to build: one encoder, one file, no need to keep multiple streams in sync. For a quick demo it's fine. The cost only shows up the moment you want to change your mind about layout — which, for anything you plan to publish, you almost always do.
The three usual workarounds, and what they cost
Once the overlay is burned in, people typically reach for one of these:
- Re-record. The correct result, at the price of the most expensive resource you have on a recording day: a good take. Pacing and delivery are never quite the same the second time.
- Crop the frame. Zoom in so the camera bubble falls outside the crop. You keep the take, but you've thrown away part of your screen resolution and permanently changed the composition. On a 4K capture cropped back to 1080p you lose detail everywhere, not just where the camera was.
- Cover it up. A blur or a colored rectangle over the camera area. Works, looks like exactly what it is.
None of these recover the original choice. They're all ways of living with a decision you made before you knew how the take would look.
The fix is structural, not cosmetic: keep the camera as a track
The alternative is to treat the recording the way a video editor treats footage: keep the raw streams separate, and make layout a timeline decision.
That means at capture time the app stores the screen capture and the camera capture as two independent tracks in one project file. Nothing is composited until export. After recording, the camera bubble becomes a normal object on a canvas — you can drag it, resize it, round its corners, or hide it entirely, at any point, including three weeks after the take.
This is what we built Bothcast around — screen, camera, microphone and system audio all land as separate tracks, and the picture above is the post-recording canvas where the camera gets placed. The movement happens on the timeline, not at capture time.
It's worth being precise about why this is rare in screen recorders. Writing a flattened 4K stream is easy for the encoder. Writing multiple synchronized streams at full quality in real time is a harder engineering problem — it needs an efficient capture pipeline and, realistically, Apple Silicon-class hardware to do it without dropping frames. That's the trade most tools opted out of.
Doing it in Bothcast, step by step
If you're working with a track-based recorder (Bothcast or otherwise), the flow looks like this:
- Record as usual, with the camera enabled. Don't worry about where the bubble sits during capture — in track mode it isn't being written into the video anyway.
- Open the take in the editor. The camera appears as an object over the screen video.
- Drag it anywhere on the canvas. In Bothcast this is the Screen position pad — center, corners, or anywhere in between.
- Tune the shape: camera radius for the corner rounding, content scale for the zoom inside the bubble, a shadow if you want it to lift off the screen.
- Change your mind later. The layout stays editable in the project even after you've exported once. Exporting doesn't flatten the project, only the output file.
- Export at whatever resolution you need. The camera is composited at export time, at full quality — no generational loss from a pre-composited source.
The whole adjustment usually takes under a minute, and you can do it after watching the take once, instead of before pressing record and hoping.
What this doesn't fix
Track-based recording fixes layout decisions, not performance. If you flubbed a line, pulled a face at the wrong moment, or the camera focused on your forehead for a minute, no amount of repositioning helps — that's a re-record, in any tool. What it removes is the entire category of "the take was fine but the bubble was in a stupid place", which in practice is most of the reasons people re-record.
A short pre-flight checklist
A habit that helps regardless of which recorder you use:
- Frame the camera against the final layout, not a test screen — sidebar and toolbar positions change between apps.
- Record 10 seconds and watch it back before committing to the full take.
- Check what's behind the camera bubble area: notifications, tabs, anything with text.
- If your tool keeps tracks separate, leave the camera slightly smaller than feels right; it's easier to enlarge a small bubble than to shrink one that's covering something.
That last one matters most if you're still on a flattening recorder — it's the only lever you'll have.
Bothcast is a native macOS recorder that captures screen, camera, microphone and system audio as four independent tracks, then lets you reposition, reshape and rebalance everything on a timeline. The free edition covers the full workflow: bothcast.app.
Top comments (0)