← VideoBGNinja

Green screen vs AI matting — which should you use?

For decades there was exactly one way to replace the background of a video: film in front of a green (or blue) screen and key the colour out. Today, AI matting can cut a person out of ordinary footage with no backdrop at all. Both approaches work — but they fail in different ways, cost different amounts, and suit different jobs. Here's an honest comparison.

一How a physical green screen works

Chroma keying is colour-based: the software makes every pixel close to a chosen colour transparent. Green is used because it's far from human skin tones and because camera sensors capture the most detail in the green channel. That simplicity is the strength — and the source of all the classic problems:

• Lighting is everything. The screen must be lit evenly; wrinkles, shadows and hot spots become different shades of green that the keyer can't remove cleanly. In practice you need two lights on the screen and separate lights on yourself.
• Spill. Light bounces off the backdrop and tints your hair, shoulders and skin green. Keyers have spill-suppression controls, but heavy spill (standing too close to the screen) never keys perfectly.
• You can't wear green. Green clothing, green props, sometimes even green-ish eyes get keyed into holes.
• Space and cost. A usable setup is a backdrop (fabric or paper), a frame or wall mount, and at minimum two softbox lights — realistically €80–300 and a couple of square metres you have to keep clear. Setup and teardown take time every session.

In exchange, you get something valuable: a key that works on anything in front of the screen — people, pets, products, smoke, glassware — and that runs in real time, which is why broadcast weather studios and streamers with a permanent setup still use it.

二How AI matting works

AI matting doesn't look at colour — it looks at content. A neural network trained on huge amounts of footage of people predicts, for every pixel of every frame, how much of it belongs to the person (a soft alpha value between 0 and 1, not a hard yes/no). The model behind VideoBGNinja, Robust Video Matting (RVM), adds a crucial ingredient for video: temporal tracking. It carries a recurrent memory from frame to frame, so it remembers where you were a moment ago. That's what keeps edges stable while you move, instead of flickering like per-frame photo cutouts do.

The practical consequences:

• Zero setup. Film in your living room, office or outdoors. No backdrop, no lighting rig, no spill, and you can wear green if you like.
• Genuinely good hair edges. Because the model outputs soft alpha, fine hair strands get partial transparency — comparable to a decent chroma key, without the green fringe.
• Zero marginal cost. No hardware to buy. Tools like VideoBGNinja run the model on a GPU server and give you a transparent WebM, a green-screen MP4 and an alpha matte for free.

The honest limitations

• People only. RVM is a human matting model. It has learned what people look like — not cats, cars or coffee mugs. A clip without a clear person in frame comes out nearly empty. This isn't a bug to be patched next week; it's what the model is.
• Busy backgrounds are harder. A moving crowd behind you, strong patterns, or background objects that look person-like can confuse the model and cause edge wobble.
• Low contrast hurts. Dark clothing against a dark room, or hair against a similarly-toned wall, gives the model less to work with — just as it would a human rotoscoper.
• Not real-time for you. A hosted tool processes your clip in seconds-to-minutes, fine for recorded content but not for keying a live camera feed (that's still green screen territory, or GPU-hungry local tools).

Want the best possible AI result? See our guide on how to film yourself for a clean cutout.

三Head to head

	Physical green screen	AI matting (RVM)
Cost	€80–300 + space + setup time	Free (hosted tools)
Works on	Anything in front of the screen	People only
Lighting demands	High — even screen lighting required	Low — normal decent lighting is enough
Spill / colour restrictions	Green spill; can't wear green	None
Hair & soft edges	Excellent with a good key, poor with a bad one	Good, driven by subject/background contrast
Live use	Yes — real-time keying	Recorded clips (or heavy local GPU)
Filming location	Wherever the screen is	Anywhere

四So which should you choose?

Choose a physical green screen if you key a live camera (streaming with real-time background removal in the studio sense), if you shoot non-human subjects like products or animals, or if you produce daily in a fixed spot where a permanent, well-lit screen amortises the effort.

Choose AI matting if you record clips of yourself — talking-head videos, TikToks/Reels, course material, demo overlays — and don't want to buy, light and store a backdrop. Upload your clip, get a transparent WebM, a green-screen MP4 (yes, AI can hand you green-screen footage without owning a green screen) and an alpha matte for your editor.

And the hybrid answer many creators land on: AI matting for everything recorded, and only invest in a physical screen if a real-time or non-human use case actually shows up.

✂ Try AI matting on your own clip — free

五Questions

Is AI matting better than a green screen?

Neither is universally better. A well-lit green screen is still the most controllable option and works on any subject. AI matting wins on convenience and cost — but only for videos of people.

Does AI background removal work on any video?

No — and be wary of tools that claim it does. Person-matting models are trained on humans. Pets and objects are hit-and-miss; clips without a person come out nearly empty.

Why do green screens produce green edges?

That's spill: light reflecting off the backdrop onto your subject. Keying removes the backdrop, not the reflected green on skin and hair — spill suppression tools reduce it but distance from the screen prevents it.

More guides: use your cutout as an OBS overlay · film yourself for a clean cutout · remove photo backgrounds