On this page
What the effect actually is
You've seen it even if you've never named it: a title or a word that sits behind the person in the shot, with their head and shoulders passing in front of the letters as if the text were a real object in the room. That's the text-behind-subject effect, and the reason it reads as premium is that it gives the frame genuine depth. The text isn't pasted on top of the picture; it lives inside the scene at a particular distance from the camera, and anything closer than that distance occludes it naturally. It's become a staple of both social content and polished title design precisely because that small cue — something passing in front of the words — sells the whole composition.
Why it used to be a chore
The traditional way to achieve this, in After Effects or a similar compositor, is rotoscoping: hand-tracing the outline of your subject so you can mask them out and slot the text behind. On a ten-second clip at 30fps that's potentially three hundred individual frames to trace and refine, and the refinement is the painful part — hair, fingers, and soft edges never sit still and never trace cleanly. Semi-automated helpers like Roto Brush take some of the grind out of it, but they still need real cleanup whenever the background is busy or the subject's silhouette is irregular. It's achievable, but it's an afternoon, not a minute.
How AI does it differently
The modern approach skips masking entirely and works from depth instead. A depth-estimation model looks at each frame and assigns every pixel a value representing how far it is from the camera — a depth map rather than an edge outline. Once you have that, putting text behind a subject becomes almost trivial: you give the text its own depth value, and any pixel nearer to the camera than that value simply renders in front of it. The model recomputes this for every frame automatically, so a moving subject keeps occluding the text correctly without anyone tracing anything. In v8eo this all runs in the browser on WebGL, so your footage never leaves your machine.
The workflow reflects that simplicity. You upload your clip (MP4, MOV, or WebM), click to place your text — bold, high-contrast, and concise reads best behind a subject — and click the person or object you want it to sit behind so the AI can segment and track them. A threshold slider then controls how aggressively the text gets hidden: lower values tuck it behind more of the scene, higher values only let the very closest objects pass in front. When it looks right, you export in full HD — no account needed.
Getting the cleanest result
Some footage simply depth-maps better than others, and a few habits noticeably improve the outcome. Clear separation between subject and background helps most — a person against a plain wall reads far more cleanly than the same person in a cluttered scene. Even, consistent lighting helps too, since hard directional shadows can confuse the model about what's near and what's far. The effect handles natural movement comfortably, but heavy motion blur on fast action can soften the result on those frames. And placement matters creatively: the effect is most striking when the text passes behind a face or the upper body, where the eye expects depth, so position it to interact with your subject rather than floating in empty background.
Where the pro tools still win
To be fair about the trade-off: After Effects and dedicated compositors give you finer, frame-by-frame control, and for broadcast work or anything that demands pixel-perfect edges they remain the standard. The browser approach is trading a sliver of that control for an enormous amount of time. For social content, fast iteration, and the large category of projects where a convincing result in under a minute beats a perfect result in an afternoon, that's a trade worth making — which is why plenty of editors use it to prototype a look quickly before deciding whether a shot is worth a full compositing pass.
The fastest way to judge it is on your own footage: open the editor, drop a clip in, and try the effect — it takes under a minute and needs no account. Once the depth text is sitting where you want it, auto-generated captions and a film color grade are the natural next steps to finish the shot.
Related: Cinematic video editing for beginners | Best caption styles for social media
Tagged